I will post the link to my working version of TalkingPixie2 V1.0 shortwave transmitter which was discussed at great length in a previous thread, during the time I got the audio to modulate.
I will post the link to my working version of TalkingPixie2 V1.0 shortwave transmitter which was discussed at great length in a previous thread, during the time I got the audio to modulate.
http://www.kdxradio.com/pixie2.html
I wish to ask opinions about one very specific portion of this circuit, namely R3 and R5.
In my first working transmitter, Qr (2N222A) was inadvertently damaged, and I replaced it with a part from the junk box, GE-20 (Phillips ECG 123A). This change of transistor caused the modulation to drop to almost nothing, but I found that making R3 = 6.8k and R5 = 10k the full modulation returned.
The question I am asking now, is, what are R3 and R5 doing which so critically matters with regard to the modulation working correctly?
|
|
Get your Quartz Crystals, Including 13.56 MHz for Short Wave and 1.0 MHz for Medium Wave – click here for a full selection! |
Is Something Missing…
R3 and R5 are a voltage divider to set the bias voltage at the base of the transistor. The bias voltage sets the operating point of the transistor.
Since there is no Emitter resistor (R4 is not shown) there is no negative feedback to swamp out differences between substitute transistors. As such, a different transistor with only slight differences could cause a large difference in how the stage works.
Was R4 mistakenly not shown or is there no R4? Usually this would be a small value, less than 1000 ohms.
Wherefore R4
Hello MRAM:
Thank you for commenting.
I eliminated R4 thinking it would maximize the power output of Q2, but your observation makes me want to restore R4 to its suggested value which was stated as 10 or 15 – ohms.
The original circuit of “Talking” Pixie 2 AM Transmitter
http://www.indianapolis.net/QRPp-I/talking_pixie2.html
was an untested modification of an earlier CW keyable QRP circuit.
If R3, R4 and R5 can be correctly chosen the transmitter should modulate for all the listed Q2 transistor choices, and get Bruce MICRO1700 and others set with their versions of the circuit.
KDX Now Switches To Our International Broadcast Frequency…
Let me know what your schedule will be. I listen to that frequency sometimes hoping to hear someone. Knowing when to listen would be great.
Notification Will Be Sent
YES I will send out word when 13.560 is on the air again. It will be a few days, since the work area needs to be cleared of whatever the last ten projects were.
Thank you Carl for bringing this back up!
I have been primarily working on my 1962
Gates mixing board. However, the 13.560
project is still of great interest to me!
Anything you or MRAM or anybody else
come up with will be gladly received here.
Best Wishes,
Bruce, MICRO1689/1700
Planning Next Version
Yes, Bruce MICRO1700, and all interested parties. The project is back to life, and this time we will produce a version that works with all appropriate output transistors, which will solve your situation, Bruce, and at least one other person who is having modulation problems.
This Node, “13.560 Back In the News,” is the Official point of communication about this subject. The link to the schematic for V1.0, given in the opening panel, will serve as the official reporting point of results.
I have posted the request that all communications about this project be directed here, and I will not answer e-mail sent to my website, simply because it makes too many lines of contact to keep track of.
Please accept a slow start, as I have misplaced some of my earlier notes and also need to order a few new parts. But for start, the focus is the voltage divider consisting of R3 and R5, and the question of a resistor R4 for the base of Q2. I think getting those parts correct is the main task at this time.
When we have the improved circuit it will be named Version 1.1
To Carl And The Rest Of The Troops
Thank you for this. I have
a lot going on today, but I
will have some comments very
soon. I have to think of a
list and write them down.
One thing – 13.560 is a really
good yardcasting channel if you
have a portable shortwave radio.
My 13.560 MHz modulated crystal
oscillator went 1000 feet down
my street. The power was a few
miliwatts. The antenna was a
22 meter dipole in the attic.
The transistor was not tuned to
the antenna. The output impedance
was unknown. The modulation was
not perfect either – but the AM
signal was good, unless I tried to
tune the stage. Then it FMed like
crazy. 2F, 3F, etc. harmonics were
removed by having a ham radio antenna
tuner in the line. The tuner was
peaked for max signal. This took the
harmonics away. That’s as far as I
got. Oh – one more thing – I had to
put two modulation transformers together
to get modulation. That was just the
nature of the thing.
Best Wishes,
Bruce, MICRO1690/1700
Good Input
Very good input, Bruce MICRO1700.
Now I see that the transmitter schematic is not complete without a section on modulation connection, and frankly, I spent a lot of time matching and refining the modulation amplifier. I was too lazy to document it and rather dismissively said something like, “You are on your own.”
This revision and re-design will have a modulation page added.
A Place to Start
Right at the beginning of this thread on finalizing the Talking Pixie2 MRAM saw right away something not right in the way R4 was bypassed to ground and the voltage divider R3 & R5 was therefore only suitable with the old replacement transistor I used for Q2. To fix it, all we need to do is select proper values for R3, R4 and R5, then the transmitter should work with the specified Q2 (2N2222A) or any replacement transistor.
To retrace what I did so far, here is the original Talking Pixie2 QRPp diagram which got this whole project started.
http://www.indianapolis.net/QRPp-I/talking_pixie2.html
At the start I built this circuit exactly as shown, but it would not produce audio, which I solved by replacing L1 with a 10k resistor which I called R5. Unfortunately a mistake ruined my 2N2222A (Q2) and the substitute I used was different enough that I was only able to modulate by changing R3 from 33k to 6.8k.
It may interest you to go back farther in time to the very earliest circuit from which the “Talking” version was built
http://www.qsl.net/we6w/projects/pixie2.gif
Good Stuff!
I’ll be pulling out all the pieces of my
13.560 kHz project from storage.
Here we go again!
Bruce,
MICRO1690/1700
Maybe This
Just thinking about this and having looked at many different transistor circuits, perhaps the voltage divider in my V1.0 circuit does NOT need the part I called R5 (formerly L1), in which case the emitter of Q2 could be left to ground as I have it drawn….
That would leave the question of R3, which was originally 33k …
Maybe restoring R3 to 33k and removing R5 would fix the problem?
Can’t try it yet, don’t have Q2
A couple of odds and ends…
Carl:
Sorry – I can’t answer your question, but
I have a few additional comments.
I found a whole other unbuilt Pixie kit here
from HSC electronics. I’m still looking for
the original project because it has the 13.560 MHz
crystal.
The original Pixie transmitter (which was made for
morse code) will work off of a single 9 volt
battery. It may be easier to modulate if the supply
voltage is lowered. (Although I think you already
tried that.)
The original morse code Pixie transmitter has been
used on the 80 and 40 meter ham bands by many
people – so the transmitter does work – even though
the circuit maybe hasn’t been optimized. Contacts
of several hundred miles have been made. But this is
a few hundred milliwatts of just carrier that can be
combined with a product detector in the receiver
at the other ham station.
Under the limitations of Part 15 – and with
a few milliwatts into a 22 meter dipole for 13 MHz, I don’t
expect that the experimenter will achieve useful
skywave skip. The AM signal, in my opinion, is just
too weak. Although the ionosphere can be a real clean
“RF mirror” just below the maximum usable frequency –
I think it is still very unlikely for that kind of unusual condition to
occur. The best example of this that I can think of is
hearing the second harmonic of the BBC Europe transmitter
(15.070) on 30.140 MHz with an indoor antenna in my
receiving shack in 1990. In that case, the MUF must have
been just above 30 MHz, and the power of that second
harmonic was probably A LOT more than a few milliwatts.
However, I still think that this is a great Part 15 experiment
for many reasons. First, we all have learned a lot from
this. Even my little one transistor circuit was a great yardcaster,
as I mentioned before. And my original idea was just
to have a signal that my ham radio friend could hear at
his house a mile away. And, the unit could turn out to be a
great link transmitter.
So Carl, keep the ideas coming. I’m going back to work on
the Gates board tonight, but I will be digging up the other
parts for the 13 MHz project.
Thanks and Best Wishes,
Bruce, MICRO1690/1700
Just Found More Stuff
I just found the original one
transistor 13 MHz AM transmitter,
the crystal, some modulation
transformers, and a bunch of
other parts. I have almost
enough parts to build a whole
other Pixie, as well.
And I’ve got an audio amp here
that should drive the thing well,
however we eventually decide how to
build it.
Woo Hoo!
Bruce, MICRO1690/1700
Hey Carl…
I know I posted that a
13.560 MHz Part 15
transmission would probably
just be local.
I’m just wondering if I
discouraged you from the
idea of transmitting in that
band, because you haven’t
commented again. (Of course
there could be many other reasons
why you haven’t made another
post.)
Best Wishes,
Bruce, MICRO1690/1700
On Mission
Hello Bruce MICRO1700 and all others
No discouragement here. I have been spending large amounts of time reviewing what got done the first time we tackled this shortwave design for 13.560 mHz, all starting with the Talking Pixe2 diagram.
It seemed to me confusing that the Talking Pixie2, when built exactly according to the schematic, didn’t modulate audio. Now that work has been proven by another participant who built Talking Pixie2 also with no audio success, that being Radio Bon Espirit in France, who wishes to have a functioning transmitter for summer use.
Today I am ordering the parts I need for not one, but two SW transmitters. The first one will be the hopefully final version of Talking Pixie2 with proper transformer modulation using Q2 2n2222A; and the 2nd unit will allow transformerless modulation of audio, something that already is known to work from a previous experiment.
Another report later today.
Research Committee
Dear Research Committee:
Our member Radio Bon Espirit has submitted an interesting AM SW Transmitter diagram which may hold lessons to help us solve the Talking Pixie2
http://www.kdxradio.com/am_files/Low_Budget_1_Watt_AM_shortwave_transmitter.pdf
Study this circuit and submits your observations about it.
Yikes!
Carl, I looked at that link of the
transmitter from the person in
Europe. You can’t get much
simpler than that.
If he can fully modulate a 1 -2
watt 2 stage transistorized
transmitter (I’m assuming this)
then we must not have that far
to go.
Best Wishes,
Bruce, MICRO1690/1700
Look Again
Take a second look at the 1-Watt Transmitter schematic and notice several capacitors listed as “nF”
nF means nano-Farad, a very tiny tiny capacitance value. My electronics catalog doesn’t even list nF capacitors.
I think the author meant to say uF, as in micro-Farad.
Anyone agree or disagree?
milli, micro, nano, pico
I don’t think the author meant uF. I’ve seen many European schematics that use nano Farad, where in the USA it would be expressed in uF or pF.
For those who wonder what the heck we’re talking about:
Milli, Micro, Nano, Pico – Scientific Notation for Electronics
Capacitor Conversion
Thank you SCWIS. We did not realize the European practice of notating in nano-Farads. Here is a conversion for the two nano-Farad capacitor values which appear in the 1-Watt SW circuit:
10 nF = .01 uF
100 nF = .1 uF
Capacitor Question
There is a.01uF cap, shown on the 1-Watt Circuit between the junction of the audio transformer and the inductor to ground.
There is no corresponding capacitor in the Talking Pixie2 between T1 and L2.
What is the function of having a capacitor in that location?
Capacitor Question
That cap is pretty much standard procedure to bypass any RF at the top of the inductor to ground. This isolates RF from the modulation transformer and prevents effects on the circuit from any RF that gets into the transformer.
Don’t use the output filter circuit as shown if you want to build this for 13.56MHz. The filter is likely optimized for the ~6MHz operation stated for this transmitter.
A little Caveat Emptor here concerning published circuits. Over the years I have built several circuits published in magazines and on the web. My success factor has been about zero. They may work, but typically not as described. There is too much opportunity for typos and stupid mistakes in the publishing process, not to mention the question of how competent the publisher was. Real commercial devices go through design iterations and refinement before they go on sale, then they get supported for fixes and improvements over time.
My most successful magazine build was an electronic keyer published in QST years ago. It actually worked as described, but was pretty much unusable because the specified relay was too loose ans sloppy to produce good CW without terrible contact bounce. I sort of made it work by adjusting the spring and bending the contacts, but that was sure not mentioned in the article.
I think a factor is that many people who may respond with their build experience are reluctant to expose themselves in public as being somehow incompetent, so they just keep quiet or say something like “I don’t quite have it working, but I think it’s a great design”. I
Exactly
Thank you PhilB for commenting on this SW project…
Indeed the Talking Pixie2 as first printed does not produce modulated audio when built exactly according to the diagram. Through blind chance I got a version to work, but not with all mentioned output transistors, which is why this thread came to life, trying to build a universal version which works reliably with all types of Q2.
No commercial part 15 transmitter exists for 13.560 shortwave, and I’m sure you have considered developing such a product, but obviously there is little interest in shortwave in the U. S.
There are several of us who would invest in a SS Tran SW AM device.
HINT. HINT.
SS Tran 13.560 kHz
The problem is, in France, that 13560 is an utility like or alarm device channel … I use 13870 kHz that is free channel.
I wait for the US ECG123AP to make a test with Pixie 2.
73s
Bertrand
Welcome Here
Welcome here to the conversation, Radio Bon Espirit. Also, as you perfect your transmitter, note what PhilB told us about the capacitor seen on the 1-Watt Schematic (which you contributed), between the audio transformer and the inductor which heads to output transistor. He says that a capacitor at that location is a normal and expected part, even though the Talking Pixie2 did not have such a capacitor. So that makes it official, there needs to be a .01uF (Europe 10nF) capacitor added to Pixie2 at the junction between T1 and L2. The other end of the capacitor to ground.
capacitor
Carl,
As I build the “Talking Pixie” by modifying Pixie V1.0, I put the capacitor and let it in his place.
Thank you. I let you know when the new transistors arrive from USA.
Bertrand
Useful To Know
In several experiments tried by Radio Bob Espirit, one was to try a resistor for R4, based on MRAM’s comment that the existing circuit for Talking Pixie V1.0 should have some value of R4, whereas I had the emitter grounded. Here is what Radio Bon Espirit wrote:
“With 680 ohms as R4, no modulation and the TX is less powerfull.
I put the strap back.”
Yes, a resistor alone at R4 would reduce signal gain. But there is a solution, and it is one I did not think of at that time.
By adding an electrolytic in parallel with R4, perhaps 1uF, with (+) toward Q2 emitter and (-) to ground, the full gain of Q2 is preserved even with R4 at 680-ohms.
I am not suggesting anyone try this, unless for experimentation.
Hi Carl!
I’m glad you are working on this stuff.
I got all of my 13.560 odds and ends
back together in one place.
I plan to start experimenting maybe as
soon as tomorrow night.
Best Wishes,
Bruce, MICRO1690/1700
I forgot one thing.
Aside from experimenting with
more complicated circuits, I
am going to take the one transistor
13.560 MHz transmitter and put it
in a nice little box.
Along with a whip antenna, a volume
control, an on/off switch, and an LED
power indicator.
It’s an extremely practical yardcaster
if you have a shortwave portable.
Bruce, MICRO1690/1700
Getting to Work
Good, Bruce, you are ahead of me because I’m waiting for parts that have been ordered, and right now there is a power line down in the street, so I want to shut the computer off for safety as soon as the repair truck arrives. We had a furious 20-minute storm that ripped things all apart. One neighbor has a tree on their house. KDX is safe, but I’m being very careful.
Probably what I will do very soon is post the audio circuit and show the way it is connected to the modulation transformer. This could be a point of some misunderstanding for some builders.
ECG123AP
Hello all,
Since the ECG123AP are just mailed today from USA, there will be no news from the french Pixie on 13870 kHz for weeks.
It takes a long long time to get amplitude modulated.
Cheers
Bertrand
ECG123AP
Hello Carl,
Did you find info on the ECG123AP transistor ? I can’t find anything on the web.
Thank you
Bertrand Bon Esprit
Class C Amplifier
Hello France! Two days ago I put a request here for help finding transistor data sheets, but so far no one has responded. I have no info on the ECG123A. Probably the source where you purchased yours can tell you where to look, or you can contact jameco.com
Found a better description of Class C Amplifier than my books contained, simply from Wikipedia:
Class C amplifier
Class C amplifiers conduct less than 50% of the input signal and the distortion at the output is high, but high efficiencies (up to 90%) are possible. Some applications (for example, megaphones) can tolerate the distortion. A much more common application for class C amplifiers is in RF transmitters, where the distortion can be vastly reduced by using tuned loads on the amplifier stage. The input signal is used to roughly switch the amplifying device on and off, which causes pulses of current to flow through a tuned circuit.
The class C amplifier has two modes of operation: tuned and untuned.[11] The diagram shows a waveform from a simple class C circuit without the tuned load. This is called untuned operation, and the analysis of the waveforms shows the massive distortion that appears in the signal. When the proper load (e.g., a pure inductive-capacitive filter) is used, two things happen. The first is that the output’s bias level is clamped, so that the output variation is centered at one-half of the supply voltage. This is why tuned operation is sometimes called a clamper. This action of elevating bias level allows the waveform to be restored to its proper shape, allowing a complete waveform to be re-established despite having only a one-polarity supply. This is directly related to the second phenomenon: the waveform on the center frequency becomes much less distorted. The distortion that is present is dependent upon the bandwidth of the tuned load, with the center frequency seeing very little distortion, but greater attenuation the farther from the tuned frequency that the signal gets.
The tuned circuit will only resonate at particular frequencies, and so the unwanted frequencies are dramatically suppressed, and the wanted full signal (sine wave) will be extracted by the tuned load (e.g., a high-quality bell will ring at a particular frequency when it is hit periodically with a hammer). Provided the transmitter is not required to operate over a very wide band of frequencies, this arrangement works extremely well. Other residual harmonics can be removed using a filter.
Data on ECG123a
http://alltransistors.com/transistor.php?transistor=33366
ALSO
The transistor might arrive with a data sheet included.
B E C
Carl,
Thank you but my point is to know where is collector base emitter on the ECG123A …
To avoid burning the 5 Usa sent to me waiting for weeks and weeks …
Bertrand
HEATSINK
Carl, I found a good heatsink for TO 92. I’m waiting for it. The only thing is to know where are base emitter and collector.Sorry to repeat but I CAN’T find it anywhere . Does anybody know ?
Bertrand
Velleman K4001
Hello Carl,
I made a Velleman K4001 7 watts amplifier kit. I want it to use as a modulator for the Tx, but gain is very high. The potentiometer 47 K inout is un efficient (doesn’t change anything) What can I do to make the sound lower ? Put a 1 Mohms variable resistor ? Thanks
Version 1.0.6 Features Class C Output
Based on the valuable input from Radio8Z, we have updated our link to reflect the suggested Class C mode for Q2.
http://kdxradio.com/am_files/pixie2.html
As of the moment, this new circuit has not been tested.
Feedback
I am mulling over your current thoughts ….. I presume your recent note addressed the failure I had at making the Class C circuit work for Q2…
Indeed in comparing to the version that is working for Radio Bon Espirit, he used a different audio transformer, which you said may possibly not be performing correctly in my case under Class C, although it works perfectly using the collector feedback bias with R3=22k..
Previously you thought my circuit might be Class A, although tonight you referred to Class E and I don’t recall any previous mention of Class E so maybe you can clarify what you mean…
With Main Power Buss 9-VDC (VCC) the voltage right now at the collector of Q2 is 7-VDC.
I have an oscilloscope but my work area is so confined there is no place to put it. I need to re-arrange something….
Starting to Build 13 MHz transmitter
OK. I’ll be starting to build it
tomorrow night. This may take me
awhile. Weeks? Months? Due to my
vision situation which you guys
are familiar with now.
Anyway, I’ll keep you posted. Also, when
the unit is done – I do have a low power
ham radio watt meter that goes down to
about 5 miliwatts. When this thing is
done I can measure the output power.
Another note for Bertrand. I checked my
records and I have received Radio France
on 162 kHz here in Connecticut 4 times.
I do not DX the longwave broadcast band very
much because I don’t have a lot of time.
The 162 kHz signal is one of the very strongest
here in the east coast of the U.S.
Also before 1980, I know that Radio France was
on 164 kHz. I have a very close friend who did
hear the 164 kHz signal in Connecticut back in
the 1970s. I did not have a good longwave receiver
then, so I did not hear it myself.
I do find the longwave broadcast band to be a
very interesting and fun part of the spectrum.
Every station I hear is like a gift because they
are all so far away.
Best Wishes,
Bruce, MICRO1690/1700
Confusion
Sorry, Carl, that I wasn’t more clear about the class E tx. I mentioned. It was in this thread: http://www.part15.us/node/2791
I didn’t mean to pull you off topic rather to point out how readings can be misleading.
Keep trying, you will get yours to work.
Neil
Hello France
Hello Radio Bob Espirit, I am also waiting for parts to arrive, but things are usually pretty quick over on this continent, so they could arrive anytime.
Then I will instantly go into high gear and work all day for as long as it takes to solve that tricky mystery….. amplitude modulation with The Talking Pixie2 transmitter.
Bruce MICRO1700 is also setup to work at his location along the eastern seaboard of the U.S.
Many others follow these shortwave subjects and contribute engineering guidance, even having no shortwave themselves.
Transistor App Sheets
Something I’m sure has been posted before on the site, where do we find application sheets for transistors?
Radio Bon Espirit tells us he order the Phillips ECG 123A, also known as GE-20, the Q2 in my Version 1.0 schematic, which is a TO-92 can, quite large compared to 2n2222. Although I didn’t put a heat sink on my GE-20, we probably should find a heat sink if one is available, and need the pin arrangement data to be sure of what’s going on.
One transistor 13.560 MHz Transmitter Running
Hi Carl and everyone!
I fired up the one transistor transmitter for
13.560 MHz and it is working OK.
You can barely call it a transmitter, but it
is transmitting music here in my office, and
it sounds OK through a Tecsun PL-380
portable.
The modulation may not be 100%, but it
isn’t bad.
The RF oscillator is running at 1.5 volts,
and the audio amplifier modulating it
comes from one of those little Radio Shack
grey pocket amps that runs on a 9 volt
battery.
That is being fed by an MP3 player.
There are two modulation transformers –
one feeding the other.
What I enjoyed about this circuit is that
I was able to take some knowledge that
I had and make something work. I mentioned
it before, but this little rig went about 1000
feet down my street with a 13 MHz dipole
in my attic.
However, I look forward to us experimenting
with the more sophisticated circuits again.
Best Wishes,
Bruce, MICRO1690/1700
Buffer Stage
It occurred to me that when we start working
with the more complicated circuits again –
maybe it would be useful to put a buffer stage
between the oscillator and the RF amp.
I heard a friend say that it a bit difficult to
modulate a 2 stage transmitter. He said
something about leakage from the oscillator
going through the final stage – therefore it
interferes with the modulation process. So
he said a buffer would help. I have no idea
if this is true in this application, but my
friend has a great deal of knowledge – so maybe
he has a point. The old Medium Wave Alliance
web pages have some buffer circuits.
Best Wishes,
Bruce, MICRO1690/1700
Q2 May Become Q3
The focus is on Q2, and then a word to Bruce MICRO1700 about the buffer idea.
Bon Espirit has ordered the ECG 123A, equivalent to GE-20, the Q2 that got me modulating with success. First, pin arrangement. Here is the package from GE-20
http://kdxradio.com/am_files/ge-20.JPG
According to this pack, the can size of GE-20 is TO-18, the same as 2N2222A, but Radio Bon Espirit reports his will be a TO-92, a different size.
To find a heat sink try
http://jameco.com
they will find it for you and are very helpful.
Bruce,
Just based on your suggestion, the plans are expanded now. There will be a total of 3-projects all at once:
1.) make the 2-transistor Talking Pizie2 Amplitude Modulate with 2N222A as Q2;
2.) make a 3-stage version with the same type of transformer modulation;
3.) make a 3-stage version that can be modulated without a transformer.
Those are the objectives of this project.
Carl…
That sounds great!
I’ll be standing by, but will
also be doing some research.
Best Wishes,
Bruce, MICRO1690/1700
The Audio Part
Here is a drawing of my audio amplifier showing how it connects to the Talking Pixie2 AM transmitter
http://kdxradio.com/am_files/pixie_audio.JPG
You only need about 1-Watt into the Talking Pixie if it’s to be Part 15 compliant.
Radio Bon Espirit said he has a 50-Watt amp and that worries me a lot.
What should he do to make that safe?
Diode
Here’s a note from 11-19-09, from Radio8Z who recommended trying a IN914 or IN4148 diode in place of L1, a part that didn’t seem correct for a modulated transmitter.
Calling Radio8Z.
Are you there, Radio8Z?
Let’s talk some more about adding a diode. The question is, what will it do?
Diode
Hi Folks,
I’ve been viewing this thread from afar but have been called out of the woodwork about using a diode in the circuit. My suggestion was based on a Class C amplifier but it appears that this amplifier is biased Class A so the diode would not function as planned. In a Class C amplifier the oscillator signal needs to “punch” the base of the final pretty hard to drive it into saturation. To enable this, the diode keeps the base from going below -0.7 volts by developing a DC charge on the coupling capacitor which produces a higher positive current on the positive cycle of the oscillator signal into the base. You do not need nor want to do this if running Class A.
Time is short for me at the moment but I wanted to respond to Carl’s query and can give more detail later about this. There is also a comment to make about Carl’s reported overheating observation which may be due to operating Class A which produces more final transistor heating than would Class C. It looks to me that the original and modified amp. is Class A which is hard to modulate properly compared to a Class C design.
The class of an amplifier is determined by the biasing and I haven’t worked this out for the given circuits.
Keep researching and I can add more later.
Neil
Amplifier Class
There is a double meaning to the expression “amplifier class.” Thank you Radio 8Z for pointing out that we are, at the present time, dealing with a Q2 being a biased Class A type amplifier.
We have been building upon a circuit from the QRPp International Radio Club, based on what was originally a simple QRP CW (code keying) transmitter, the actual author of the AM version being unsigned and thus unknown, and I suspect he drafted the AM circuit without actually building and testing it. Three of us have built it and it doesn’t modulate audio.
I just reviewed The ARRL Radio Handbook Section 11-17 on the several Amplifier Classes, AB, B and C, and what jumps out is the description of Class C, used for LARGE AMOUNTS of RF power…..stop right there. We are dealing with small amounts of RF power. What I am thinking is, perhaps the author of the Talking Pixie2 AM version was thinking “low power” when he chose the class of amplification.
Comments are welcome.
Amplifier class is dismissed.
Never Easy
As mentioned my audio modulated version of Talking Pixie2 used a GE-20 in place of my damaged 2N2222A. The GE-20 is obsolete, but is equivalent to ECG 123A, and Radio Bon Espirit has ordered one of them hoping it will bring him audio.
But here is a passage from the book Basic Electronics Technology, p. 50 –
“Two transistors, though they are the same part number, may have slightly different operating characteristics.”
This tells us that using the same part may or may not solve the issue. Be sure to let us know.
Class C is back in Session
Back from my road trip…some comments follow.
A good circuit design is one in which the passive components set the circuit operating characteristics rather than the particular transistor used. Gain, for example in Class A, is set by the ratio of the collector AC resistance over the emitter AC resistance. Biasing is set by the base voltage divider resistors and the emitter resistor. Then, the selection of the transistor is essentially a matter of choosing one which will handle the power dissipation, voltage required, and the operating frequency. For part 15 powers the 2N2222 or 2N3904 should work up to about half a watt. The circuit Carl posted lacks an emitter resistor which makes the biasing highly dependent on the transistor hFE (DC gain) and that is why the difficulty in substitution. The hFE can vary three to one for the same part number so there is no assurance that substitution will work. The hFE also varies with temperature and collector current which causes distortion in amps where the transistor sets the bias.
Since a Class C amp is biased into cutoff the transistor parameters are of lesser importance and substitution is easy.
Power can be controlled a few different ways with the most reliable being to select the supply voltage which gives the power input needed so there is not reason to be concerned about over power operation due to the class of amplifier used.
Keep reading references, but I suggest using a Class C output stage with another reason being they are easy to modulate for AM.
Neil
Progress Today
The package of parts has arrived, especially a 2N2222A transistor, so that work can begin on finding a stable AM circuit for Pixie2.
During a several hour tornado warning today I sat in the basement reading about classes of amplifier circuits, based on Radio8Z’s assistance, and I spot three types of circuits to consider. They are “Fixed Bias”, “Collector Feedback Bias”, and the one we have right now, “Voltage Divider Bias,” that 3rd type said to be the most common circuit used for small amplifiers.
Opinions or experience with these type circuits would be appreciated.
To Carl: and Neil
This is all good info..
I was talking again with my
friend about this circuit.
By the way – this guy is amazing.
More than 30 years ago he built
a 10 watt crystal controlled FM
broadcast band tube transmitter from
scratch! It was based on a Gates
tube FM exciter schematic that we
saw on the back of the First Class
Radio Telephone License Study Guide. And
most of the circuit values were not there.
So he had to calculate them himself!
And the transmitter did work, very very well!
He told me a few general facts that I will just
pass on. He told me that it is easier to modulate
a transmitter that is running a couple of hundred
milliwatts (such as the Pixie 2 circuit), than trying
to say, modulate a ten milliwatt AM transmitter with
a very small audio amp and modulator. Again, I’m
not sure what this means relative to our application,
but I thought I would pass it on.
Also, as you know, an AM transmitter running on the
test bench into a 50 ohm load might sound great –
but when you try to hook it to a real antenna – the
modulation may become very poor. If the Pixie 2
(with 200 or 300 milliwatts out) runs into an RF
attenuator to achieve the few milliwatts we need
(what is it, around 2 – 5 milliwatts into a full size
dipole) then we can expect a benefit from this.
If the arrangement that I described above is used,
the circuit will see an RF load of around 50 ohms
most of the time – even if the antenna is mistuned.
This is because most of the RF energy will be going
into the RF attenuation network. So then the modulation
characteristics will be stable even with a varying antenna
load.
If we don’t have luck with this arrangement, there are still
other circuits around.
Also to Carl: I thought you achieved good audio modulation
with the original
Pixie 2 circuit. I think you did this with the 7 watt Veleman
audio amp kit that you had.
It seems to me you originally achieved good modulation
right before your dummy load became disconnected
and your final transistor blew. After that, I know you started
changing some of the circuit values. Am I right about this?
I’m not sure. I guess it was more than a year ago when you
and I were working on this.
Anyway, this is fun!
Bruce, MICRO1690/1700
Made For TV Mystery
Radio8Z thank you for the detailed technical explanation about the choices in amplifier class. Class C is the way to go.
Bruce, my memory of exactly what happened and in what order is a blank, and some of my detailed notes are for now misplaced. I think it modulated before the 2n2222A went bad, but I’m not sure. What I will do out of curiosity is spin through that other thread we had, where we reported a lot of what happened.
Meet back here tomorrow.
Book Education
I have two text books that discuss Amplifier Classes, BASIC ELECTRONICS TECHNOLOGY and RADIO HANDBOOK, and both books contain what I feel is the same editing error. They describe A, AB, B and C, the amplifier classes, only in a verbal way with no example illustrations.
Next, there’s a section with illustrations showing three methods of biasing transistors, but fail to tell what Class each method belongs with.
Book learning alone does not work. One also needs a teacher.
Any teachers……. we need an example illustration of a Class C transistor circuit.
Difficult
While trying to find the information about the pin configuration of ECG-123A transistor for RTBE I read this interesting page on Wikipedia
http://en.wikipedia.org/wiki/Transistors
The part that impressed me was the description of naming problems as different manufacturers apply their own names and numbers to transistors. Also there is ambiguity regarding the shapes of transistors, which vary also.
This isn’t an answer to the problem, but a confused person likes to find that he is justified in being confused.
Surely there’s a source for transistor data sheets. Some websites claim to have data sheets, but often fail to provide.
ECG123AP Data
It is often difficult to find the exact datasheet via a search, especially for obsolete parts. The best I could do is cross reference the part with an old RCA SK databook where this crosses to SK3854 in a TO-92 package. The basing is the same as the 2N3904.
For the SK3854, viewing the part from the bottom with the flat facing away the emitter is on the left, the base is in the middle, and the collector is on the right.
You can tell which lead the base is by using an ohmmeter. With the positive lead (usually the red) on the base you should measure about 600 ohms to each of the other two leads.
Neil
thanks
Thanks Neil, I’ll check as you do.
Have you got an idea to low the audio down (my post before ) Thank you very much.
Audio Gain Question
Have you got an idea to low the audio down (my post before )
I searched the circuit diagram for this amp assembly and also for the TDA2003 it uses. The gain is set by the resistors R1 and R2. For this op amp circuit the gain is K = 1 + R1/R2. The data manual shows R1 = 4.7k and R2 = 470 which gives a gain of 11. I suggest changing R2 to 4.7k which will drop the gain to two. You could even remove R2 go get a gain of 1 but it may be unstable.
You can try other values for R2 or even use a 10k potentiometer (or close to this) in place of R2 to get adjustable gain.
The high frequency roll off is set by R3 and C1 and you may have to change R3 if the frequency response is not right.
Neil
Remove R2
Thank you Neil
By removing R2, gain is correct now, when I change it from the Ipod, it ‘s loud enough, or low as I want. I think I have a correct modulator now.
Thank for your help. I let you know about the ECG123AP, and you’re right, it may be like all the TO-92.
Thanks again
Bertrand
More Audio Info
RTBE I’m glad you found a solution to fixing the audio level. For your information here is how I wired my Velleman amp.
http://kdxradio.com/am_files/pixie_audio.JPG
Thank you Carl
I’ll do the same !
ALSO
Please note that the 8-ohm resistor shown in the Velleman Amplifier link should be 10-Watts.
Check the Drawing
Carl,
In your link http://kdxradio.com/am_files/pixie_audio.JPG it appears that you have reversed the amplifier input and output. The output on the triangle symbol used for amplifiers is from the pointed end. You labeled it correctly but you should reverse the triangle.
Not trying to be picky but rather trying to avoid confusion.
Neil
Loose Ends
You are correct, Radio8Z, I drew the amp symbol backwards. Within an hour from this post I’ll replace it with a corrected picture, and some additional notes.
I just cleared off my disaster area work bench and pulled my Talking Pixie boards out of their big box, and see a few loose ends to also mention.
– The Velleman K1823 Power Supply kit is labeled 1-Amp, but Jameco has it in their catalog as 1.5-Amp. I’ll send Jameco an e-mail to get their comment. It does run very hot with only the two transistors of my presemt SW transmitter;
– In my version, the Velleman 7W Amplifier is being powered by a separate power supply, a wall wort rated at 12-VDC 1500mW
– The audio modulation transformer has a center tap on the 1K side, which should not be connected.
Audio Pic is Repaired
As promised, the picture showing the 7-Watt Audio Modulation Amplifier for Talking Pixie2 has been repaired
http://kdxradio.com/am_files/pixie_audio.JPG
Hi Guys
Hi Carl and Neil. I am not feeling
well. Just some kind of bug – not
serious. I just seem to need a lot
of rest.
Anyway, when I get back up to speed
I will have some more comments. And
I am looking for a very good electronics
theory book I had around here somewhere.
Carl, there is one thing that I sort of
remember in your original tests of the
first Pixie 2 with the original power
transistor and the Veleman amp.
I seem to remember that you had more
modulation with less supply voltage
to the Pixie 2. If you lowered the
Pixie 2 supply voltage, the modulation
increased. If you raised the voltage
the modulation became less.
Actually, I could be wrong about this.
It may have not happened that way, but
the idea is stuck in my brain as a fuzzy
and distant memory.
Also, I could be wrong about this next idea.
But in the discussions about amplifier class,
I’m wondering something.
Wouldn’t the designers of the original Pixie want to
use a
class C amp configuration in the second stage?
After all, the original power amp was just for morse
code. Oh well, just a thought. If the amp really
isn’t class C, I guess that’s the way it goes.
I seems to me that the power input of the oscillator
stage is just ten or twenty miliwatts. So the
second stage would have a gain of more than ten.
I’ll have to read up on some things. I’ve got
to find this book.
Best Wishes,
Bruce, MICRO1690/1700
On the Air Now
I spent most of the evening stringing up an indoor dipole, not based on a calculated length, but on the convenience of available space going in two directions from where the workbench is located.
The carrier has high hum in some receiver locations, but as I move around with a portable there are spots where the hum is gone.
At 9-Volts the modulation is somewhat low, but at 5-Volts it becomes strong.
I have yet to start trying changes.
Bruce, I believe Neil suggested Class C for the final stage because he said it will modulate well, so when we go to a 3-stage version, the 2nd stage will probably be some other class, not C.
In a minute I’ll see what happens with a dummy load.
Take good care of your health, Bruce.
Reception Report
Still testing by antenna I took a 120-foot walk into the night and the signal remained constant the whole distance, so it is penetrating the walls and getting out.
I think it’s time to crank up some coffee and post pone bedtime.
THIS MAY BE IT!
Here is what I just tried and it is working better than ever. Voltage is all the way up to 9-VDC, modulation is firm and loud. The bias type put on the base of Q2 is called Collector Feedback Bias, but I do not know what CLASS of amplification that is. Take a look.
http://kdxradio.com/am_files/pixie2_1.0.2.JPG
All parts are the same as my V 1.0 with two exceptions
R3 is now back to the original 33k and C10 is a newly introduced part based on previous discussion.
I still have the GE-20 transistor, so this needs next to be tested with 2n2222A, which will probably be later in the day.
POWER INCREASE
Because this transmitter is so stable, I increased power from 9-V to 12-VDC. Of course increasing the carrier power caused the modulation level to drop, however this is easily compensated for by boosting the modulation drive coming from the audio amp.
POWER DECREASE
Running at 12-VDC for close to an hour made Q2 too hot to touch, so I’ve come back down to 9-VDC.
Ongoing Tests V1.0.2
Removed GE-20 as Q2, replaced with 2n2222A.
Relationship between audio modulation and RF carrier is different, running at 9-VDC the carrier seems stronger than the modulation is able to match, so I rolled voltage down to 4.5-VDC and am able to achieve full audio modulation.
The only change available for this version of the circuit is to change value of R3 from 33k to a different value. Before trying that I will wait awhile to see if anyone has a chance to comment on this version
http://kdxradio.com/am_files/pixie2_1.0.2.JPG
IMPORTANT CHANGES
Now the circuit can be driven with the intended 9-VDC and achieve full modulation….
Change R3 from 33k to 100k
This will be Version 1.0.3
During Saturday (today) I’ll update every link to point to this version, since it is the solution of the moment.
Modify
Carl,
Waiting for the ECG123AP, do I need to modify the circuit as you did ? 100 k is definitely the value of R3 ?
Do I test Pixie V 1.0 with that new transistor before ? Or can I use 100 k remove the other resistor and keep 2N2222A ?
Thanks
Bertrand
Hello France
RTBE I’m glad you’re following this. I have tried two more changes with R3, since this is a work in progress and may change for awhile….
With R3 as 330k the modulation wasn’t able to be loud enough.
With R3 as 150k it may be slightly better than 100k, it modulates very well across a VDC range of 4.5 to 9-VDC.
Maybe more voltage could be applied if a cooling fan was added, but I don’t know if that would add to the distance or not.
Yes…. if you change to this new circuit, RTBE, you should also get modulation as I have, and I am using 2N2222A, but the ECG 123A would (I think) also work. …. yes …remove R5.
Please let me know what you get.
Not Final
The changes made so far are a definite improvement, because now we can achieve amplitude modulation with more than just an obsolete transistor for Q2, but there may be more improvements to be made.
Input from Radio8Z and other more qualified engineers are looked forward to, with their comments on the design we’re dealing with at this stage, using “Collector Feedback Bias.”
I have just gone outdoors and walked the same 120-feet as I did last night, only this morning the signal is not so strong along that path as it was last night, with R3 = 33k. On the other hand, the difference may be the time of day and not circuit changes.
Testing Continues
Talking Pixie2 V 1.0.5
R3 = 47k
I am still trying different versions of R3 because I suspect that there is an inverse relation between signal strength and modulation strength.
It seems, unless I’m only imaging it, that the lower the value of R3 the RF signal becomes stronger, the higher the value of R3 (up to a certain limit) the higher the audio strength, but with less RF power.
As a compromise 47k seems to balance the two.
Comments welcome.
DONE
Ladies and Gentlemen,
The objective has been met for now. We have constructed a 2-transistor version of Talking Pixie2 Shortwave Transmitter for low power 13.560mHz AM broadcast, using the original 2N2222A as Q2
Check it out
http://www.kdxradio.com/pixie2.html
The next project in this series will follow a suggestion by Bruce MICRO1700 to produce a 3-transistor version featuring a buffer amplifier between the oscillator and final.
See you then.
Fantastic!
Carl, this is great! I still am
not feeling well, but I will have
comments soon!
Congratulations!
Bruce, MICRO1690/1700
Buffer Stage – Do you need one?
Congratulations again Carl!
Maybe you don’t need a buffer
stage. It would depend on how
well the transmitter is performing,
I guess.
I think you should get that thing
into a dummy load so you can check
stability, modulation, and that sort
of thing.
Just an idea. More comments will be
coming.
Best Wishes,
Bruce, MICRO1690/1700
Under Discussion
Hello MICRO1690/1700 and welcome to Sunday.
We can discuss the idea of the buffer stage. To me, I like the concept, even if this version proves to be very stable.
I am definitely going to build a #2 because I have a second crystal, and it seems like the chance to try something different.
Another version worth designing is one which modulates without a transformer. There are at least two well proven ways of doing that.
Just in using the transmitter it “seems” stable. The audio fidelity is very clear and “present.” The signal around the yard is constant and covers better than the AM transmitter.
I would like to have a simple tuning capacitor to “tune” the antenna resonance, so let me know if that is as easy as just adding one. What value-range?
Once this laziness passes I’ll do the scientific testing under dummy load.
ALSO I would love to install a “power” control like some AM transmitters have, to adjust from very weak up to maximum output.
OUTPUT METERING
The new Version 1.0.5 Talking Pixie2 SW Transmitter has been running now for 2-days at 15.530mHz, and I just stuck an AC meter on the RF output for the first time.
The meter swings high and low just like a VU meter, following the modulation. The low side, silence, is about 1/10th of a Volt, the high swing goes up to about 1.75-Volts. The average modulation is about 1-Volt.
I didn’t know it would follow the modulation that way. Previous measurements were always taken with no modulation.
Circuit Tweaks
Carl, nice work and thanks for the blow by blow as this project unfolds. I looked at your circuit and wish to point out that the DC bias depends on R3, the supply voltage, and the hFE of the transistor so others who replicate this will probably have to adjust R3 for their particular transistor due to the variation from one device to the next. This is not a problem as long as others are aware of the possible need for this adjustment.
Here are some calculations which may help:
The DC base current is the supply voltage (VCC) minus .7V divided by R3. IB = (VCC-.7)/R3 = (9V-.7V)/47k = 0.177 mA.
The collector current is hFE x IB = 150 X 0.177 mA = 26.5 mA assuming a typical hFE of 150.
The power dissipated by the transistor is IC x VCC = 26.5 mA x 9V = 238 mW.
This may help you follow the what happens as you change VCC and R3. For example, lower values for R3 increase the power dissipation and lower values for VCC decrease the power dissipation. Without looking I believe the power rating for a 2N2222 plastic is 600 mW and you could calculate back to get a combination of VCC and R3 which would be less than this. The problem is that the hFE can vary quite a bit and unless you measure it for a particular transistor the calculations may be inaccurate.
Interesting meter readings for the output. Be aware that most meters will not respond accurately to RF signals but it could be useful for comparison purposes.
If you want adjustable power a simple way to do this is to use a LM317 IC with a pot and a fixed R. It could be installed between the power feed to the oscillator and the final so the voltage to the oscillator remains constant. The LM317 datasheet shows how to connect this and how to calculate the resistors. I did this in my prototype transmitter so I could keep it at 100 mW input without disrupting the oscillator supply voltage.
Keep us appraised of your progress.
Neil
Woe
Well, Neil, with the grace of a diplomat and humanitarian you have allowed me to realize that my quest of finding a universal circuit good for all situations has not been reached. And to make things more grave, I could not have realized the situation without your input, which is therefore priceless.
Still, you offer some clues as to repairing the situation, or, on the other hand, we could look for the “Class C” output you earlier suggested.
I cannot find a single example of a class c circuit. In fact I don’t know what “class” the circuit we now have amounts to.
For now I’ll go quiet while I contemplate the options.
Many thanks.
Class C
Carl,
I am not trying to pull you off track since your circuit is working but if you want to try to operate class C it wouldn’t be too hard to modify what you have.
My guess is that your circuit is operating Class A. Class C would be with no bias to the base. Remove R3 and place a 100 ohm or so resistor from base to ground. Add a diode (1N914 or 1N4148) in parallel with this resistor with the banded end toward the base. The unknown is if the signal voltage from the oscillator is high enough to drive the final. The other problem could be that a class C has a low input impedance which might overload the oscillator but so does the circuit as you have it now so that is a good sign.
I don’t have a way to post a picture or a drawing but will try to do so tomorrow.
Neil
Power Limiting ?
Hello Radio 8Z,
Sorry for approximative english and very limited electronics knowledge.
I’m in France, and there’s no part 15 law in this country, so I’m not allowed to broadcast.
I’m very fond of broadcasting, I have my own Radio Station 24 Hrs a day here in Paris, with a SSTRAN and AM88.
This morning, I put a 2N2219 as Q2 on the schematic. It seems more powerful, and I increased the modulation.
It sounds loud and clear, with a HiFi amplifier and a telephone audio transformer.
I’m very interested in improving the circuit. But I’m not limited by HF power. If I put a diode down there, would the power decrease or be limited ?
As you can understand, the more power I can get, the more happy I am.
Thanks for your answer.
PS : I’d like to thank here very much Mr Carl Blare for his kind help since I started the project.
1.0.7
Hello all,
Tested the 1N4148 diode and R5 100 ohms.
It works, but bad modulation.
I kept the new diagram, but put back R3 between base of Q2 and L2
The modulation seems to be VERY MUCH louder and clearer than before.
More linear I think?
Thanks Carl and Radio 8Z
Diagram soon on Carl’s KDX site.
Let’s Do It
Fabulous, Neil…
Class C is back in our sights.
What you have said is clear, but the drawing will make it perfect.
I will make sure to have the parts and try this ASAP.
We are very indebted to you for this help.
Two Countries Working Together
Thank you France for testing the Class C version of the Talking Pixie transmitter.
Today I drove way on the other side of town to find a diode so that I can verify the results.
Here is a question….RTBS…. when you replaced R3 did you then remove the diode and R5?
Kept the diode
Hi Carl !
No I kept the diode and R5, just put back your R3.
I sent to you the diagram.
It’s 1.0.7
Bertrand
Great Job!
Great job you guys. I’m going to ‘
start building a whole new 13.560 MHz
circuit soon.
It will be on a big board so parts can
be easily changed.
Thanks again.
Bruce, MICRO1690/1700
Strange Waters
The international team has some unexpected results.
As RTBE has noted, his laboratory version of Talking Pixie2 is operating in combined Class A and Class C mode.
Stumped again.
The Present Version
Here is the diagram uploaded by RTBE showing a hybrid of the past two versions.
Is this a legitimate circuit?
http://www.kdxradio.com/am_files/Pixie2_1.0.7.JPG
Legitimate Circuit?
If it works then the answer is yes.
The addition of R3 along with the diode and R5 could (depending on the R values) bring the base bias up to about cutoff for the transistor. This would result in class C operation if the drive signal from the oscillator is strong enough. Is the transistor running cooler than before?
Neil
To Carl or Bertrand or Neil…
I just want to be sure.
What are the final values for R3 and
R5?
Thank you and Best Wishes,
Bruce, MICRO1690/1700
Final Values
The values for R3 and R5…
R3 = 47k
R5 = 100-ohms
I think France is using the same.
Thank you Carl
I appreciate that.
I’ll be getting the pieces
together.
It will take me longer, but
this is on the front burner
now.
Thanks again!
Bruce, MICRO1690/1700
Transistor Temperature
Radio 8Z
Here, I use 2N2219 as Q2. AT 18 volts, temperature is cool.
RTBE
Trouble Report
Radio Bon Espirit was the first to try the changes including diode and R5 (100-ohms), and he found that keeping R3 (47k) gave the best modulation experienced in all tests, calling this Version 1.0.7
But, here in America, I just caught up by adding the diode and R5 (100-ohms), and my experience is different. The carrier is on the air, but the modulation is crushed to a barely audible garble.
Measuring AC at the antenna I see VU-like swings up above 2-Volts in tempo with the audio, but this is not audible on the radio.
In trying to find why mine is so different from the one in France, let me ask you, RTBE, to double-check your R5. 100-ohms should be color coded brown-black-brown.
Covering Every Base
To be sure to try every variation I backed up to Version 1.0.6, which includes three changes:
1.) Removed R3 (47k)
2.) Added D1 (1N914)
3.) Added R5 (100-ohm)
The audio modulation improved only slightly, but was subdued and slightly crackly.
Until we have a new focus, I’m returning to Version 1.0.5 which has only R3 as the bias for Q2
Carl…
In any of the versions where your
audio is bad – was the radio really
close to the transmitter? Was it
being overloaded? I have put radios
next to my 1690/1700 AM transmitting antenna,
and they go silent because they are
overloaded.
Or is the transmitter FMing? In that
case if you tuned to the side of the
center frequency – you might hear the
audio with slope detection.
Best Wishes,
Bruce, MICRO1690/1700
The Physical Setup
Hey Bruce:
The Sangean ATS 505 radio is about 25-feet from the transmitter, and about 10-feet from the end of the antenna, which is strung horizontally up at the 8-foot height for the length of the house.
I do have a portable Grundig FR-200, so maybe I’ll test again using it.
Neil will probably have an idea of what we can try next.
The differences between my transmitter and Bon Espirit’s are as follows:
My Q2 is 2N2222A while Bertrand’s is 2N2219A
My T1 is exactly the one shown in the original Talking Pixie2 diagram, whereas Bertrand’s is one he selected by trial and error, with unknown specs.
We each have different audio amps.
Re-Tests
While the transmitter was operating with R3 in place I carefully pushed R5 and D1 until they were back in the circuit, and the audio dropped to a very clipped low level.
Tried 150-ohms for R5 and it was still not allowing the audio to modulate.
During these tests I moved the Grundig to different distances, but the audio was suppressed continually.
The truth about 1.0.7
Hello All
I confirm I use a 2N2219A as Q2, 100 ohms at R5, 47 K at R3, and the diode is 1N4148, I think the same as 1N914.
About the audio section I have an old french public telephone 600 ohms audio transformer I modulate very strongly from an HiFi audio amplifier. I can ear the modulation IN THE TRANSFORMER ITSELF !..
However, the modulation is not on the same power suply than the transmitter, for the moment.
I know, it’s not very “orthodox” but it works fine. It works really better than the audio kit Velleman i’ve tested before
.The modulation is absolutely pure, loud and clear, as strong as any commercial short waves radio station.
I’m planing to use the transmitter this summer in the fields at the countryside, so I bought a 12 volts motorcycle audio amplifier, and will use the system on a small 12 volts battery with an inverted V 1/4 wave antenna.
When I tried 1.0.6 version, by removing R3 47 K and puting R5 and 1N4148, the audio was bad and I decided to put back (?) the 47 K resistor and keep the diode and 100 ohms, and it worked much better at once. It’s just a chance. May be with more linearity in the modulation, and no more spurious signals from 12 mHz to 15 mHz I had before.
May be the difference comes from my alien type of modulation, or from the specifications of the 2N2219A. I can not say why, but it does work. I’m absolutely sure of this. The only slight problem I have is a light hum coming from the transmitter power supply. It will disappear with the battery.
I precise that my antenna is a piece of wire and I put a 47 ohms resistor in the filter to avoid problems or may be burning the 2N2219. At this time, the circuit works with 18 volts, and Q2 is cold. Total consumption is 100 mA. I have no idea if there is 100 mW HF or not, and here in France, it’s not a problem because there is no part 15 law, and personal broadcasting is prohibited.
But as we say in french : if you have no right, take the left.
Bertrand
A Few Comments
Hi Carl! I don’t know what to say. I
guess we’ll have to think about it for
a while. There is one thing – I also
have an FR-200. I find mine to be
great on the AM band, but not very sensitive
on SW and FM. Although for short range
testing it would be very convenient because
of the small size.
Hello Bertrand! If you have no right,
take the left. I like that very very much!
By the way, although I have never heard
France here in Connecticut on the U.S.
east coast on the AM BCB – I have heard the BBC on 693,
Spain on 684, Germany on 1593, and the
U.K. again on 1215 kHz. I have also heard
a few longwave 153 – 285 (?) kHz stations
from Europe, here. I remember now – I have
heard France, I think, on 162 kHz. Does that
sound correct?
Best Wishes,
Bruce, MICRO1690/1700
162 khz
hey Bruce !
Yes we have France Inter, our “NPR” on 162 khz omnidirectional 2000 kW.
When you live very near Allouis Town, where the transmitter is located, you can really ear the modulation in the razor when you shave yourself !
Tweaking the Tweaks
Radio8Z put a lesson in how to tinker with the mathematics of the output transistor, under his title CIRCUIT TWEAKS.
I’ve been studying and practicing and put forward this assumption based on what is said.
“The power rating for 2N2222 is 600mW,” and the power dissipated by Q2 in my Version 1.0.5 (R3=47k biasing the base) works out to 238.mW.
Is that to say that if I “tweaked” using the numbers, and got total dissipation to, say, 500mW, that Q2 would then be at the top of its game? Does that mean it would put out the most gain available to it?
Tweaked
Using the math recently learned I ventured ahead and did what I just talked about.
By making R3=22k we are now getting an estimated 508.95mW dissipation out of Q2, which is purported to have maximum dissipation rating of 600mW.
The audio is crystal clear.
This type of biasing is called “Collector Feedback Bias.”
This change makes Version 1.0.8
By the way, 9-VDC is part of the set of properties that make this circuit work, therefore a different voltage would change the balance and require another math calculation.
Readings
Back with Version 1.0.5, where R3= 47k and Q2 power dissipation was .177mA the AC meter showed RF output average of about 1-V with negative swings to 1/2-V and high swings to 1.75-V;
Now Version 1.0.8 with R3= 22k and Q2 power dissipation of 508.95mW the RF is averaging 1.5-V with pos/neg swings in a much tighter pattern close to the average of 1.5-V.
Here are some additional notes to several members:
— MRAM asked for an ON AIR schedule so he can check 13.560mHz. The transmitter will run during this whole schedule as seen here…
http://kdxradio.com/schedule.html
— PhilB reminded us to put a proper filter on the output, and I inform that SCWIS supplied a correct filter, which is part of the diagram
http://kdxradio.com/pixie2.html
The present diagram needs to be updated.
Reply to Carl re Tweaks
Edited to Add: You posted twice while I was composing this so what I say here may not be helpful in light of your new information.
Carl,
The equations I posted were for the DC biasing of the transistor and the AC performance requires a different analysis. The 600 mW (actually 625 from the data sheet) is the power dissipation rating of the transistor and does not indicate an optimum operating condition, just the maximum.
Here’s an example: The Class E design I reported on before draws a DC average collector current of 42 mA at 2.36 volts collector to emitter and outputs 86 mW to the antenna so the transistor is dissipating no more than 13 mW and probably not this much due to filter loss. How is this possible since 42 mA X 2.36 V = 99 mW? The answer is that the voltage and current are not present at the same time in the final transistor and these are average volts and milliamps rather than instantenous. The point is that AC analysis is different from DC analysis and the equations I posted don’t predict AC operation.
I don’t have much to advise regarding the modulation problem you have. An oscilloscope would be a great aid in diagnosis but lacking this other troubleshooting methods may help. Can you monitor the audio into the transformer with a headphone? Is it distorted? Perhaps the DC current through the transformer secondary is saturating the core? Maybe the transformer step up ratio is too high?
Measure the collector to ground voltage using the DC scale on the multimeter. It ideally should be about half the supply voltage. It could be that the transformer is not able to swing this voltage up and down without hitting a limit (zero volts) causing the distortion.
Neil
Discussion So Far
Have been reviewing this thread and several branch threads referred to in some places, and I learn that the Pixie2 shortwave transmitter is only one thing that needs refining.
In my opening post I referred to the Pixie2 diagram on my website at that time, and much of the ensuing conversation talked about the voltage divider at the base of Q2: R3 & R5.
But as time moved forward the diagram has evolved to match each change made along the way.
The point is, if someone started at the beginning of this thread, they would not see the correct opening diagram.
Having more than one problem to solve at the same time can be very distracting and perfectionists become frustrated by their own mistakes.
Can you all donate to a fund to buy me a lovely assistant who would have her own desk, telephone, computer and plenty of free lunch?
Problem – Solution
In the last message I described the problem that someone first starting to follow this thread would no longer be able to see the schematic mentioned at the beginning.
The solution was to put two schematics on my webpage, showing the starting version (1.0) and the most recent version (1.0.8).
http://kdxradio.com/pixie2.html
All in service to the great Part 15 Lifestyle.
Cross Reference
In the continuing effort to keep referenced data accessible throughout this thread, I have installed links on the Pixie2 web page for each of the Versions of the developing schematic. The most recent Version appears right on the webpage.
http://kdxradio.com/pixie2.html
Reflection
Upon reflection it seems possible that, while RTBE was able to operate his Pixie2 in Class C mode but we were unable to make ours work with Class C, the difference in voltage may be the reason.
RTBE was running 18-VDC, while we were locked at 9-VDC.
Consider this statement by Radio8Z: “In a Class C amplifier the oscillator signal needs to “punch” the base of the final pretty hard to drive it into saturation.”
Does this seem like a reasonable theory?
Saturation
Hi Carl! We both sent in comments at about
the same time, so I have one above your last
that you might like to look at.
What you say about saturation and the difference
in Bertrand’s voltage (18 DC) and ours (9 DC)
makes sense to me. Perhaps someone else
will comment. As you know, saturation is the
complete turning on of the transistor, it is like
big kick into full conduction and that is all –
nothing less. If we add a voltage regulation
system for our needs, the transmitter can
probably operate over a range of voltages
above 9 volts, up to a certain point that is
practical.
I have to run, talk to you soon.
Bruce, MICRO1690/1700
Good Idea Carl
Having all of the versions on your
website helps us understand where
things are going and that is good.
My Pixie 2 transmitter kit was ordered
from HSC electronics. (I actually have 2.)
So I can look at the original Pixie schematic
and plug in the changes right on the document.
There are a few things about what I have here
that are different from you and Bertrand.
Aside from the original Pixie kits, I have an audio
transformer that is very close to the recommended
one, and I also have an unknown transformer. Both
of them work in the one transistor version. I think
the unknown transformer is an AC stepdown type
for powering a radio. It looks like 120 VAC in and
9 VAC out. It is amazing to me that it works.
One other situation I have here. I am going to have
to make the output coils myself. The various websites
where you can order parts are very hard for me. For
instance, I do not have the eyesight to navigate
though the Digikey site and order from there. There is
too much fatigue that sets in looking through all of the
charts for the correct component. I know there are lots
of coil calculators on the web. So I may have some
questions about this once I start making coils.
When the circuit is done here, and I can measure the
output power with my low power wattmeter, this is
what I plan to do. I will just build an attenuation
network to bring the power output down to the
proper level. That will be a lot easier for me than
trying to make other circuit changes to bring the
power output down. And the transmitter will see a
more constant output load as well.
I do not have an amp like the Veleman. But I do have
a powered speaker here that is kind of banged up, and
I can use the amp from that. I’m sure that the output
of that unit is at least more than a couple of watts. It
is probably something like ten.
So thank you, and I’ll let you all know more when I have
something to report.
Bruce, MICRO1690/1700
1-Transistor AM
Bruce, here we are working on a 2-transistor version of the Pixie2 SW transmitter, but I think it’s very significant that you have amplitude modulated with only one transistor. Do you have an estimate what your output power is?
Meeting
Meeting.
My office.
O.K., here’s the deal. The work on Talking Pixie2 2-transistor AM SW transmitter is, for my part, done.
We have RTBE’s Class C Version 1.0.7, and we have my Class A with “Collector Feedback Bias,” Version 1.0.8.
Of course, anyone is welcome to go forward with other versions and if they do, we hope to be kept apprised.
However, we, that is me, at this time, do declare a new 3-transistor version is now officially in the planning stage, under a new name.
The Pixie name was good for a small device, meaning “wee little fairy” from mythic folklore. But we’re upsizing just a tad with a new name.
Introducing BIG TALKER.
Features will include…
buffer amp stage at low impedance with extra “punch” to excite the final stage…
Class C Final stage…
LM317 power variable stage with user controllable output power control.
Suggestions are welcome.
BIG TALKER will open for business on a new thread titled “Big Talker.”
Finally, this thread is not closing, as new information or comments may happen at any time.
Lunch?
Afterthought
Radio8Z, my brain has gone over some of your informative notes and I want to re-visit the entry you titled CIRCUIT TWEAKS.
At first I jumped to the notion that you were talking about my “Collector Feedback BIAS” circuit, featuring R3 in Class A mode.
But there is a particular number used in your calculations which resembles the voltage held by the diode in Class C mode, -0.7 volts.
In CIRCUIT TWEAKS you show .7V subtracted from supply voltage (VCC), and I am guessing that the figure “-.7V” wouldn’t apply in a non-class c situation…?
I don’t mean to over-think, but I try to be precise.
Thank you.
-.7v in the Equation
In CIRCUIT TWEAKS you show .7V subtracted from supply voltage (VCC), and I am guessing that the figure “-.7V” wouldn’t apply in a non-class c situation…?
That is correct. The .7V accounts for the base to emitter voltage drop in a class A circuit where the base is biased into the linear operating region. The diode clamping at -.7V is a different term and would apply to a class C amplifier but since there is no bias resistor (R3) in a class C circuit the equations I posted would not apply.
The .7V figure for both the B to E in class A and B to GND in class C comes from the voltage which appears across a silicon junction (B to E in class A and the diode in class C) when it is forward biased.
Neil
33 K ? No! Carl’s 47 K !
About 1.0.7
Tried this morning 33 k instead of 47 K as R3 …
Bad Idea.
Carl was right : 47 K sounds really better.
Now, I put 1.0.7 in a box.
http://www.radiobonesprit.com/Pixie107.JPG
One Transistor Transmitter For 13 MHz
Hi Carl!
Sorry this answer isn’t sooner – life got
in the way here.
My 13 MHz one transistor “transmitter”
is just an RF oscillator with the antenna
attached to the output component.
The transformer is just placed in the
B plus line. Because of simplicity of
the circuit and the fact that the output
impedance is unknown, I don’t think
one can figure the RF output value. Input
is a few milliwatts,
Getting interested again in 13 MHz a
few weeks ago prompted me to put the
little rig back “on the air.” I realize now
that the circuit is FMing even when no
output wire is connected. No wonder it
didn’t sound very good. Because of this,
I really want to eventually complete
one of the 13 MHz designs you and the
guys have come up with recently.
Thank you for coming back and doing more
work on this whole 13 MHz thing.
Best Wishes,
Bruce, MICRO1690/1700
Power Pixie
Hello users !
Tried new version of Pixie : the Power Pixie !
http://www.radiobonesprit.com/Power_Pixie.JPG
Q1 is 2N2219
Q2 is 2N3053
R3 is 47 K
It works. Waiting to install good antenna at the countryside, but … Anyone knows how to build a loading coil for using with a 1, 80 meter telescopic antenna ? It would help me very much for using the transmitter in town with antenna at the window.
Greetings
Bertrand
Pixie Reproducing
Like friendly rabbits, the Pixie projects are reproducing!
Love the Power Pixie!
Very good question about a loading coil for the 13mHz frequencies.
Help National Industry
Dian Xing.
http://www.radiobonesprit.com/DianXin.JPG
A very good amplifier modulator that modulate now Power Pixie.
Stereo.
FM Receiver.
Fine EQ Bass Treble.
Works on 12 – 18 volts
10 euros brand new on Ebay. Perfect for motorcycle, boat, and part 16, the french part 15 that doesn’t exist (remember, no right, take the left).
Still no spurious, better signal with 2N2219 2N3053, and the amplifier and transmitter are linked to the same 12 volts battery.
Bertrand
Power Pixie Super Amp
The better amp looks good….
From a French left-side-of-the-brain Part 16 project.
I listen to 13.870 many times a day, so far no reception.
At least, no antenna
There is no antenna, only 47 ohms resistor, and tests will be in august 24 hours a day on 13870 kHz. Hope it will reach my local receiver … 😉
Hi Carl and Bertrand
Sorry – I have been away for a while.
Bertrand, your transmitter looks great!
I don’t know the answer about your
loading coil, but I’m sure there are
websites out there that can probably
help with that.
Best Wishes, Bruce, MICRO1690/1700
Updates and Stuff
The 2-transistor Talking Pixie2 built here has been running ever since, all day with beautiful programs from KDX.
I don’t know if it’s “FMing,” how can I check that?
The audio quality on the radio sounds too good. That is, maybe I’m exceeding the narrower bandwidth of shortwave, since the high frequencies are zinging wonderfully with no splash. I mean “zinging” in a pleasing way.
I just stuck an AC meter on the antenna output and the reading says 2-Volts, swinging nicely with the audio.
Antenna Cut to Size
Using 1/4 wave lengths calculated by SCWIS and PhilB, 17.6-feet for 13.560mHz, I carefully measured and cut my hap-hazard experimental dipole to the exact length (X2).
Right away the A.C. hum detected by radios near the antenna was many times louder than before at 9-feet away, but that says the strength is increased.
The A.C. hum is only present near the antenna at certain angles and goes away entirely out in the field.
As yet the transmitter is not grounded, but that’s yet another thing to look forward to.
This is an indoor antenna horizontal near the ceiling reaching into three rooms.
Fascinated By Hum
Now with the indoor dipole cut to length, 17.6′, I mentioned that the Talking Pixie 2 (2-transistor transmitter) at 13.560mHz has a loud hum or even a buzz on the carrier when monitored indoors, but which disappears with a portable at some distance outdoors. With the added throw of the signal brought by a tuned antenna, the hum is detectable farther out than before. Naturally I would like to eliminate the hum.
I just read many threads here on the site about hum in AM (medium wave) installations and all the possible causes and cures, and I tried a number of experiments to try and find out exactly what this buzz or hum is.
One new idea occurs to me that I haven’t seen discussed…… this antenna length, 17.6′, is the 1/4-wavelength at 13.560, but it is the 1/2 wavelength or full wavelength of some other frequencies, and if there are strong communications transmitters on those frequencies located nearby, I could be receiving and re-transmitting someone else’s carrier. That happened to me when I was at 960kHz….. my MW antenna received and re-transmitted 850kHz which was 5kW and close by.
I have ruled out A.C. power supply noise. The diodes are bypassed with .1 caps, and when unplugged the 2200uF filter cap takes a few seconds to discharge but the hum does not go away during this DC only phase.
I disconnected audio input, shorted the input to the modulation transformer, and the buzz does not change.
I disconnected one half of the dipole at the antenna output which weakens the RF signal at the radio, but the buzz is still there.
I changed to a different phase of the AC 115v power line… no change.
Grounded to the ground system used by the AM MW with the MW transmitter first on, then off…. no change.
Tried turning a few household items off and on and unplugged… no difference.
This is a very interesting hum-buzz with as yet no obvious source.
Shortwave Studies
Last night before concluding the inquiry into trying to identify the buzz on my dipole antenna transmission, I hooked the antenna to the oscillator, leaving the final stage transistor out of the circuit, and the exact same buzz was audible. This has me wondering if there’s a bad diode in the power supply.
Then, just now, I attached a signal tracing device to the antenna with the transmitter disconnected. This little box is basically an audio amp with a tiny speaker and some clip leads, but it also has an RF listening capability, with a diode in the circuit to detect the presence of RF at a test point.
The antenna has some interesting sounds, listening to its RF component. The warble of compact florescent bulbs is noticeable, and they sound so space-alien-like that I believe Alex Jones when he says they are designed to program our brains. That’s all I know about that.
There was one totally mysterious sound that makes no sense at all and maybe one of you will have a theory… I plainly heard a Beethoven Symphony playing without interruption… and the reason that is unusual is that we have no classical station in this city. No one ever plays Beethoven Symphonys, unless you count my station, KDX, which has the exact same Beethoven Symphony (No. 3) scheduled later today. But if we are hearing that broadcast one hour ahead of schedule, we are hearing the future. Has anyone ever heard the future on a bare antenna wire?
The hum is still not explained, so I’ll stick with that mystery, since the paranormal is a different field not related to Part 15 radio.
More Humming Antenna
The paranormal Beethoven detected on the raw antenna has been explained. It seems a nearby Christian AM station was playing the classical music in some kind of a tie-in with a religious retirement center. Forget about it.
Back to the humming antenna. I tried feeding only half of the dipole with RF output, and it became super-sensitive to hum influences from everything… electric lights turned on and off, refrigerator starting, even my hand passing near the electrical socket strip. By re-attaching the negative side of the dipole the hum steadied to a constant background buzz.
Yes, this is a temporary indoor horizontal antenna hung a few feet below parallel AC lines up in the attic. No, the lead-in to the antenna is not proper RG58, but is simply plain wire (which I am counting as part of the 1/4-wave length). Now I suspect that getting the positive and negative sides of the antenna balanced with respect to one-another the hum may minimize.
PhilB’s best of all solutions (described above) is a vertical antenna 17.6-feet high, necessarily an outdoor project unless you set it up in a basketball court.
SWR I’ll Swear
The un-schooled belief I just stated, that the hum on the Talking Pixie2 dipole antenna is related to an out-of-balance condition, has become a devout belief.
In fact, now I think that all the antenna related hums had by everybody (if not due to ground loop, power supply trouble, audio cable problems and all that sort of thing) are obviously coming from bad S.W.R.
“Standing Wave Ratio” is the difference between the amount of RF power dissipated by an antenna vs. the amount of energy reflected back to the transmitter because of mis-tuning.
Is there a SWR meter capable of Part 15 application?
One more thing I forgot to say earlier…
I removed all antenna from the Pixie2 and listened to the RF output with both the plain audio and the RF detector of the signal tracer, and both ways I could hear the modulated audio with a perfectly clean RF carrier and no hum at all. It’s an antenna thing.
antenna and power supply
Hello Carl
I have the same problem, it possibly comes from the power supply. I have here in town a very bad antenna, because I wasn’t able yet to translate feets and inches into meters and centimeters (but I will) for the loading coil and the Tx hums. But it disapears when I switch to battery.
Just in case, I put VK200 RF chokes in the 12 volts line.
For me, paranormal phenomenon would be an email from somebody who listen to the 13870 kHz next number or from my street … or Kansas City.
Bertrand
Two Normals
Hi Bertrand.
A paranormal experience
is two normal experiences.
The Transmitters Wall
May I introduce to you … The Radio Bon Esprit’s transmitter room !
http://www.radiobonesprit.com/Transmitter_Wall.JPG
At the top, the big Volt Meter : 16 volts is shown
Dian Xing : the Chinese strong modulator (10 $)
The Power Pixie on 13870 kHz in his big East German style grey box
The busy power supply, amp meter swings with modulation peaks
A Fostex monitor from the Final Cut video station
All the music 2To Hard Disk
AMT 3000 in a metal box. Amplifier is off.
Now, you know all about the french “pirate” station. Waiting for a french part 15 law.
😉
Bertrand