Carl,
We need to keep separate the concepts of power and voltage. It is true that with high level modulation that the audio needs to be supplied at 1/2 the power of the unmodulated carrier. But when looking at the modulation voltage, this translates to the audio modulation circuitry being able to swing +/- above and below the carrier peak voltage sufficient to take it down to zero without clipping and upward without distortion or clipping. To a limit, this happens in the AM-25. If the unmodulated power into the final is 100 mW then the audio amplifier needs to supply 50 mW. This is trivial with today's electronics and is certainly not a "disadvantage". If R23 is set to 7.1 V then your observation is correct, with a 12 or 13.8 supply the upward voltage swing will be insufficient. I suggested earlier that a setting between mine and theirs may be optimum. Allowing for the approximate Vcc-3.3V drop I observed this would suggest the best setting at the test point would be 5.25 volts rather than 7.1 volts. I have not verified this so it is, for now, only theory.
I do not intend to bash the Ramsey unit for I think it has great potential. It is almost insensitive to the antenna load, it has very effective 7 pole RF filtering on the output and produces a very good looking sine wave at RF frequencies.
Though it runs well above the limits for final DC power input, perhaps it meets the requirements of 15.209 in terms of field strength with a 3 meter antenna. I don't know, and disappointingly, when I posted this question on their board a year or so back, I did not receive an answer.
I am now thinking that if R33 and R34 are replaced with higher resistances that the unit will meet 15.219 requirements and also provide close to 100% modulation. If this turns out to be true (another experiment for me or you to do) then the Ramsey AM-25 just might get back on the air at my QTH.
Here's one thing I found strange from the manual, page 20:
11. Adjusted as indicated, your AM-25 will produce 100 mw of RF power which conforms with the FCC rule of 100 mW. If you live outside of the USA.....etc.
The RF power produced depends upon the load which is not specified nor known. Seems pretty ambiguous to me. If they were referring to the final amplifier DC input power as is used in the FCC rules then they really missed the mark according to my tests yet the use of the word "produce" implies output power.
I really like the SSTRAN but have found that its performance is highly dependent on the proper adjustment and tuning of the antenna system. The Ramsey unit doesn't seem to care about the antenna load but it lacks the SSTRAN's control of the audio levels. The best of both worlds for me would be to bring the advantages of both units together into one. I have already experimented with tapping the SSTRAN processed audio signal out and feeding it into the AM-25. This worked well but because of the modulation limits I have seen in the AM-25 and the higher than allowed power I took it off the air a few years ago in favor of the SSTRAN which always measures compliant in terms of 15.219 according to my tests. But, I am equipped to adjust and operate a legal and clean system. Many are not and it would be good to have a "turn key" box for these folks.
Your thoughts or those of others on this?
Neil
As these notes on AM25 accumulate there will be a Manual Addendum fit for publication.
Yes a "turn key" product made easy for the non-technician is a worthwhile objective. The AM25, for me, has been the perfect unit for tinkering which is exactly why I have it. I do respect the design of the unit and I think it was modeled after a higher power design and is therefore a challenge to operate at Part 15 limits. But the responsibility is with the end user, and that's us.
The difference between "voltage" and "power" is important to realize, but what did the manual author mean? Did he/she mean voltage when saying power?
To build on everything said so far, I paraphrase something in the SSTran AMT3000 Manual, that its output is intentionally designed for the high impedance of a 3-meter wire (unless modified for the optional low impedance grounded system). By comparison the AM25 employs the standard 50 ohm output found in professional transmitters, which expect to see at least a 1/4 wavelength antenna plus ground-system. I believe that's exactly why it doesn't compare with the SSTran in simple installations.
Every issue we've discussed is part of the overall theme of this thread, "Modulation," because everything interacts within the transmitter system. The idea of combining the two transmitter is very interesting.
One strength shared by both transmitters is frequency stability and good harmonic rejection.
According to the Ramsey Manual, R33 & 34, two 10-ohm resistors, are in the circuit to prevent exceeding 100 mw output power at the input to the final. If they are insufficient, would a 50-ohm potentiometer in their place provide a way of fine-adjustment of output power?
Otherwise, maybe, output power could be fine-tuned by adjusting overall voltage to the transmitter.
Tonight I removed the Nortronics Bias Oscillator Transformer from AM25 (previously described), which provided such crisp clear modulation, and checked the other thing that was asked: how is the signal strength affected? Now I have a better answer than before.
Restoring the original L4 to the circuit the noise quieting on radios 75' distant is solid, whereas with the Nortronics a good deal of background noise crept through. Conclusion, the Nortronics transformer was "choking" at 1640 KHz and reducing RF power to the antenna.
To be continued.
Carl,
This project is on my bench now waiting for me to get inspired to do some more.
The power in question is the DC power delivered to the final amplifier, Q3 with no modulation. I measured this by measuring the DC voltage from the drain to the source (ground) using a digital voltmeter on the DC setting. I measured the drain current by measuring the voltage across R34 and dividing by 10. The power is Vds x Id.
The current through Q3 is set primarily by the conditions on the gate and the effective way to modify the power is to change the DC voltage from drain to source. This can be done in two ways. Increase the resistance of R33 and R34 and/or lower the DC output of Q9 by means of R23. So your pot idea would seem to be OK providing the pot can handle the current but it will still require quite a high output voltage capability from the audio system which could clip the audio peaks.
I think a better approach is to change R21 so the voltage set by R23 will be lower than the present minimum (about 4.5 volts) at the R34 test point. This will give the audio amplifier system more headroom and should prevent distortion on the modulation peaks. This was evident when I lowered the test point voltage from 7.1 to 4.7 V.
Unfortunately, the power input and the audio performance are not independent and adjustment back and forth will probably be needed to find the best modulation performance along with the unmodulated input power being 100 mW.
The RF output power measurement was done using a digital oscilloscope which measures RMS voltage. The power out is Pout = Vload^2/Rload. One can also do this with the peak to peak voltage reading from a scope by converting the peak to peak to RMS. Don't trust a digital meter for this since most of them do not respond to frequencies much above 10 or 20 kHz. One can build a simple diode rectifier capacitor filtered probe and use a DVM on DC to get relative voltage measurements however.
I will try to do more on this later. If I can get my unit to 100 mW with close to 100% undistorted modulation I will seriously consider putting it back on the air. I think the robustness of the output circuit to changing antenna load and the effective filtering of the RF output are real pluses. I have an Alesis Nanocompressor which I use for my FM audio chain and this could serve as well to control the audio into the AM-25 to achieve high average modulation without the danger of overmodulation.
Neil
On Jan. 18 I added a comment to this thread about Radio Frequency Interference being picked up by the AM25's 3-meter antenna and back-modulated into the audio. At that time I suggested that frequencies above the AM band could conceivably be encountered owing to the short antenna, and now I have discovered FM 107.7, 100 kw, 4-miles away, is in fact clearly present on my silent carrier.
The Part 15 rules require accepting interference from licensed stations, so the solution needs to be home-made.
Three Choices:
1.) move my antenna;
2.) build an FM trap;
3.) live with it.
THEN I began receiving a christian message on my carrier at 1640 KHz. After careful scanning of all the bands, I found the source at about 9500 KHz shortwave, which I think is Family Radio. The host was a mush-mouthed older sounding guy who was being very authoritative about explaining what was meant by the Bible.
This event only happened on one radio in particular, so it may not have been the Ramsey AM25 per se, but a quirk of the radio itself.
In earlier stages of this forum on AM25 modulation, Neil pointed out an important semantic point of language when he mentioned that wording in Ramsey's Manual referring to "voltage" and "power" meant two entirely different things. I had been interpreting the words as referring to only voltage, but Neil caused me to take a step back. However, I've found another line in the manual that makes it reasonably obvious that Ramsey did mean "voltage" when they said "power." That line is on page 5, second paragraph, and says: "These transistors are biased to be operating in the linear region with a no-signal output level of exactly one-half the supply voltage as set by pot R23."
Carl,
In the quote you provided they are referring to the voltage when they say "no signal output level exactly one half the supply voltage...". The output level means the voltage output level. It would make no sense to describe a power as being one half the supply voltage would it?
I can see a bit of confusion since "output level" could refer to the RF output power but in the context of the entire paragraph their statement is correct. Perhaps they should have worded it "no signal output voltage from Q9....". The Ramsey board has a topic regarding manual errors and if you want you might post a comment about this there.
I have had to suspend further experiments on the AM 25 for now. Maybe someday in the future life's events will get out of the way and I can get to it again, but please let us know of your results.
Neil
Neil, I'll keep looking forward to your tests when you have time.
I ended up trying again with the Nortronics Bias Transformer in place of L4, only this time I used only half of the windings of the primary and secondary and the result is good. Modulation is equal in level to an SStran AM3000. For A - B comparison the Ramsey is at 1640 KHz and the SStran at 1680 KHz and the sound quality and level are equal.
I have another question regarding R33 & R34, two 10-ohm resistors that can be removed for higher power mode, but there are two questions about those resistors aside from their affect on power.
1.) When they are in the circuit they would tend to slightly attenuate the modulation being fed to L4 where it mixes with the RF;
2.) When they are in the circuit the impedance of the combined RF and modulation (Q3 & Q9) is 20-ohms, whereas the combined impedance without R33 & R34 is 0-ohms.
That difference could change the matching and distortion characteristics, and I'd like an opinion on that.
Hi Carl,
Thanks for the report on your test. It sounds encouraging.
Though I have not done anything with R33 and R34 shorted I'll give an opinion.
1.) When they are in the circuit they would tend to slightly attenuate the modulation being fed to L4 where it mixes with the RF;
This is true but they also attenuate the voltage applied to the final amplifier drain. The lower modulation voltage along with the lower drain voltage (and thus lower RF at the drain) should track. I don't think in practice this tracking will be accurate and some adjustment of the modulation will probably be needed.
Here's how a circuit model would predict what is happening. The FET is a current source and this current is controlled by the gate. This current will be constant regardless of the voltage at the drain (within limits). Shorting the resistors raises the drain voltage but not the drain current.
2.) When they are in the circuit the impedance of the combined RF and modulation (Q3 & Q9) is 20-ohms, whereas the combined impedance without R33 & R34 is 0-ohms.
Not exactly. If one looks toward the load from the drain of Q3 at RF frequencies one does not see the resistors because capacitor C18 shorts the RF to ground leaving only the Z as transformed by L4 from the antenna circuit. The resistors, in or out of the circuit, will not change this impedance at RF. In other words, the Z at pin 1 of L4 is already zero due to C18.
At audio, the impedance seen at terminal 1 of L4 would be V/I at this point (to ground) since the audio is decoupled by C1 from the antenna. Changing R33 and R34 would change this impedance since this changes the voltage. Interestingly, theory predicts that shorting the resistors will raise the voltage and also the Z for audio. The output Z at Q9 emitter is very low and the change caused by shorting the resistors should not affect the voltage here unless the current capability of Q9 is exceeded.
These comments assume perfect performance from Q3 according to the model but this seldom happens due to higher order effects in semiconductors. They only guide us toward practice.
I wish we could sit down with some coffee and a pen and a napkin to pour over this. It would be much easier than via the forum, but we can still kick this around here.
Neil
Thank you Niel. Coffee, pens and napkins are the highest style of civilized society and I am there. In fact this gives me an idea, since the best ideas begin as jots on a napkin. How about a pack of Note Napkins from the office supply store? Let's share the patent.
I've gotten more than money's worth from the Ramsey, since I first saw it as an experimental toy, which it will remain for years to come.