- Posts
I went through all the many Ramsey AM25 threads but couldn’t find one dealing exclusively with modulation level, although modulation is occasionally mentioned in passing along with other aspects like bias and RF output.
After careful observation over more than a year I have come now to believe that the low modulation level of the AM25C, usually described as 80%, is because of the feeble modulation transformer. My guess is that it over-saturates above the 80% point and that a heavier-duty transformer would open the door on better “plate” modulation.
I’d love to hear other thoughts on this idea.
Total posts : 45366Carl,
Help me here. I have an AM25 which is many years old and it does not have a modulation transformer. You mentioned model AM25C. Is this a different design from the original AM25?
Here’s a thought on “plate modulation transformers” in general. They usually have to provide the plate current which is DC and which will bias the core toward saturation. If saturation happens with audio then the peaks of the modulated envelope will be flattened. Is this what you are seeing? This happens on my transformerless AM25 so it must be due to some other effect which I have never taken the time to pursue since my AM25 is retired.
Neil
Total posts : 45366I meant to specify that L4, hand wound “coil” on a toroid form is an 8/2 transformer which does what you said by providing a dc path and it also is the point where the modulation is injected upon the RF carrier.
Total posts : 45366Carl,
Ahhh, you are right, L4 is a transformer. It threw me since it is not a modulation transformer, rather it is a RF stepup transformer but it does have DC and audio going through part of the winding (primary) and thus your comment about saturation may well be valid.
I haven’t done too much testing with the modulation but I have noticed that if the voltage to the final is reduced by setting R23 lower than instructed then a higher undistorted modulation percentage can be had. I used to run mine with the voltage reduced just to get better audio. This supports your theory though I am not sure there is not some other cause for this.
On the other hand, if you short R33 to run in high power mode the unit still works with the higher current which results.
I notice that the power supply/modulation circuit uses an op amp and a Darlington pair which would give a maximum voltage capability at Q9 emitter which would be a few volts below the supply voltage due to drop from the op amp output circuitry and the pair. Since the DC bias is set slightly above 1/2 the supply voltage the swing will be limited in the upward direction and will not be able to provide the voltage needed to produce high modulation peaks. This would also explain my observation above regarding reducing the bias to get better audio Maybe the problem is with this or with its current driving capability. It might be instructive to observe the signal at Q9 emitter on a scope to see if it is distorted. I have not done this.
Neil
Total posts : 45366Neil, this thread is interesting and I want to take it one more step.
L4 is more than the simple coil suggested by the way it’s drawn in the schematic.
Physically there are two windings, one wrapped around the other, with a common point where Q3 both injects the RF Output and receives DC, but the modulated signal on one winding passes by inductance to the other winding for the two reasons, that the windings are wrapped over each other, and through the saturating of the signal through the toroid form.
In other words it could be drawn as a standard transformer with the primary tied to the secondary at one common point.
It is, in the full sense, a modulation transformer.
Tell me if you don’t agree (or if you do)!
Total posts : 45366Carl,
At your invitation I will respond that I do not see L4 functioning as a modulation transformer. In a traditional class C amplifier the modulation transformer is in series with the collector and modulates the voltage at the collector with an audio signal. In this transmitter the audio is added to the DC “B+” through U1 and sent via Q9 to the drain of Q3 through the transformer winding. The transformer functions to step up the RF voltage from the drain of Q3 at radio frequencies, or another way to look at this, it matches the impedance of Q3 drain to the load. The transformer L4 is not transforming the audio frequencies, only the radio frequencies. At audio frequencies the winding between the R33/C18 junction and the drain of Q3 looks like a short.
L4 could be removed from the circuit and replace with a wire and the audio modulation would still be applied to the drain of Q3.
Yes, this is an interesting thread which you have started and I don’t think semantic debates are too helpful. If we can both understand how L4 functions and what possible weakness this brings to the circuit then I don’t really care what we call it.
If I get time, I will try the experiment that I suggested where the waveform at Q9 emitter is observed. I’ll let you and our other readers know what I find. The problem with modulation over 80 percent in this transmitter has been mentioned frequently on the Ramsey board and there has been no good answer on this subject. You and I, perhaps, with some data and putting our heads together can provide one.
Neil
Total posts : 45366Neil, I see your point of view, but still have a different way of seeing it. Earlier I used the term “plate modulation,” which harkens back to electron tubes when giant transformers injected audio onto the RF at the transmitter final stage. In a transistor version the “plate” would be the emitter, and that’s what we have in the Ramsey AM25. At the same time, your reference to collector modulation is valid also, because modulation can be done from that stage, which in the case of a tube transmitter would be “cathode modulation.”
Total posts : 45366Carl,
Again, I don’t mean to be argumentative but it would be good if you and I both used the same conceptions of terms so we can communicate.
The collector of a bipolar transistor is analogous to the plate of a tube, the base to the grid, and the emitter to the cathode. The reason for this is in a tube the plate current is controlled by the voltage between the grid and cathode and in a transistor the collector current is controlled by the current flowing from the base to the emitter. The common reference for the control signal is the emitter and the cathode, thus the analogy.
The final amplifier Q3 in the AM25 is a field effect transistor where the drain corresponds to the plate, the gate to the grid, and the source to the cathode. Here the analogy is clear because the FET is modeled as a voltage controlled current source just as is a tube. The control voltage is applied between the gate (grid) and the source (cathode) and the output current flows through the drain (plate).
So “plate modulation” in a transistor circuit would be applied to the collector rather than the emitter and can be done by placing the secondary of an audio transformer in series with the collector or drain just as is done in the plate circuit of a tube.
I am aware of cathode (emitter or source) modulation schemes but in the AM25 the final amplifier Q3 source is at ground potential and the modulation is applied by changing the voltage at the drain (plate modulation). Q9 is wired as an emitter follower where the audio and DC output is from the emitter similar to a cathode follower used in tube circuits.
I may have some time on Sunday to try the experiment and, if so, will report here.
Neil
Total posts : 45366Neil’s explanation caused me to rearrange my thoughts. To save face I will fall back on having made a “Spoonerism:” whenever I said emitter I meant to say collector.
But I still stand by my transformer idea that a better transformer might improve modulation level.I will enjoy, Neil, your experimental results, and I will also be tampering with the thing and will make reports too.
Somebody said these words once, and they never go out of style: “experimental broadcasting for a better tomorrow!”
Total posts : 45366Laboratory Observation
During modulation experiments with Ramsey AM25C at 1640 KHz using loading coil on an oatmeal box and a 3-meter antenna I also happened to be running SStran AMT3000 at 1680 KHz, with a 3-meter antenna 8-feet apart.
The AM25 antenna received the AMT3000 signal and back-hauled it into the output circuitry where it seriously garbled the audio being generated by AM25. When I ran the AM25 with a silent carrier the audio from the AMT3000 continued being heard on the Ramsey at 1640.
Lesson learned – Strong RF signals intercepted by the Ramsey antenna will play havoc on whatever audio the AM25 is trying to modulate. Because the antenna is so short that could easily include much higher frequencies above the AM band if they’re sufficiently strong.
Total posts : 45366At the beginning of this thread about modulation level and Ramsey AM25C I posed a question: would a heftier modulation transformer improve peak modulation for this microtransmitter?
The answer is now YES! I replaced L4, the small hand-wound transformer on a toroid form, with a Nortronics T70-T2 bias circuit transformer scavenged from an old magnetic tape recorder, and tried two different wiring patterns, based on the available windings. Both arrangements brought modulation up above the nearby AM stations on the dial with no distortion. The 1st wiring trial had a roll-off above 4500KHz whereas the 2nd trial opens up the high end so that it’s the receiver that rolls off above about 8KHz.
Case closed.
Total posts : 45366Carl,
Nice work and thanks for the report. Life has gotten in the way of my experiment but I still plan to proceed.
Any idea if the RF output is essentially the same?
BTW, I am sure you meant 4500 Hz rather than 4500 kHz.
Neil
Total posts : 45366Yes, Neil, your question is exactly the next concern in the chain of improvements on the Ramsey AM25.
I could answer it this way: the output is essentially the same. But I think there’s more to it.
The output of the UNMODIFIED AM25 with a 3-meter antenna has less range than an AMT3000 SStran with the same antenna, therefore it’s less than desirable.
With my transformer modification I “feel” the output is slightly less than it was. The transformer may be reacting differently than the toroid form and absorbing some RF. Maybe it’s acting as a choke.
Therefore I’m going to try winding another 2/8 step-up transformer on a larger toroid and see where that takes us.
I did try one thing you said you’ll also try, namely “listen” to the straight modulation output and it sounded clean and strong. I haven’t run any of this through my scope, so if you check that I’ll be paying attention.
Total posts : 45366Carl and other interested observers,
I got around to doing my experiment and here’s what I found. I had previously adjusted R23 to give a good undistorted waveform so I started with this. The test frequency is 1670 kHz, power supply voltage is 13.77 V, audio test frequency 800 Hz sine wave and the antenna load resistance is 80 ohms dummy load.
The voltage at the R34/C23 test point = 4.77 VDC.
DC input power to Q3 = 201 mW.
RF output power to the load = 32 mW.The audio input voltage was increased until the maximum undistorted modulated RF waveform was observed. Distortion began on the zero level excursion but the maximum excursion appeared undistorted. The maximum modulation percentage undistorted = 86%.
It was noted that the distortion on the waveform at the R34/c23 TP exactly matched the distortion seen on the modulated RF envelop and that both occurred at the same audio level.
I reset the R23 bias to 7.1VDC as instructed in the manual. With this setting I measured:
Pin = 452 mW
Pout = 66 mW
Maximum undistorted modulation = 51%.The distortion under these conditions started with the top of the modulated RF waveform being flattened and with further increases in audio it clipped at zero. It was noted that the waveform at R34/C23 distorted with exactly the same pattern and at the same audio level as did the output. The peak voltage at R34/C23 when this started was 10.5 volts.
Since the peak voltage into the final stage was 10.5 volts with no distortion I conclude that further attempts to increase the voltage peak distorted the waveform because with the 13.77 V power supply the maximum output voltage of the modulator stage was reached.
My data do not indicate any deficiency in L4. According to my test, the limit on % modulation is due to limitations in the audio/bias circuit and not in the RF circuit. There is probably an optimum setting of R23 between the values I used where the distortion appears at both the top and bottom of the modulated waveform at the same time but I have not pursued this (yet).
Since this unit runs well above the part 15.219 limits for DC final input power under my test conditions, my recommendation is to set R23 to achieve the maximum undistorted modulation percentage which will be below the recommended 7.1 V setting at the test point. If one is concerned with the input power (and one should be) then another operating setting needs to be used.
I have left out many details regarding my measurement techniques but I assure all that they were done with accepted engineering practice. If anyone has specific questions about the methods I will respond.
Neil
Total posts : 45366Thank you Neil for your scientific observations. This information could be useful to others because the AM25 transmitter is still available from Ramsey.
I pose to you my belief that the Manual contains an error which can be proven by comparing two different statements:
1.) On page 5 in the Circuit Description I quote one line – “The main disadvantage of high level modulation is that the modulator’s power must be half that of the final transmitter.”
2.) On page 19 under Alignment Procedures “(7) Measure the voltage on resistor R34’s lead closest to C23. Adjust ‘Bias’ pot R23 for 7.1 VDC.”
I’ll let you conclude why both of these statements cannot be true at the same time (except, of course, when the final voltage is 14.2 VDC).
Moving on, I believe the question you raise about the precision of the bias circuit is worth considering. Where you find one error (the manual?) it’s likely there could be others.
- You must be logged in to reply to this topic.
