Class D Modulator

[KD1SH]

A couple of days ago I took a break from other projects and revisited this thread:
http://amfone.net/Amforum/index.php?topic=47350.0
Using a Hammond 1628SEA transformer, and a MPAB1X100-TPA3116 Sure 100 Watt Class D amplifier module, I built an external modulator that I can experiment with.
This morning, using my Gonset G-76 as a lab-rat, I modulated the B+ from the Gonset's power supply and looked at the results on my scope. As you can see, it looked very nice at both 20KC and down in the basement at 100Hz, although you can see just a bit of distortion on the negative peaks at 100Hz.
Odd, but as you can see by the cursor data on the upper right corner of the scope shots, I wasn't quite able to double the B+, but with voice rather than sine it might be quite different.
I also looked at the actual modulated RF output from the Gonset, and it was lousy, but then I remembered (duh) that my test configuration applied modulated voltage to the plate only, not the screen, so that's that.
Now I just need to try it on a properly configured transmitter.

[KD1SH]

Okay, with the Gonset properly set up to apply my externally modulated B+ to both the plate and screen, here's a scope shot of the RF output under modulation. The modulating frequency is a 12KC sine wave, and I increased the amplitude until the negative peaks looked about 100% - difficult to be precise with a digital scope's "jaggies". Maybe I'll dig out the old analog scope for some better looking pictures. The yellow trace is the modulated B+. The actual amplitudes aren't really equal; I adjusted the scope to overlay the two traces. Note the slight phase shift between the modulated RF and the modulated B+. Down at 50Hz there's no shift. Also note that there's a bit of asymmetry in the modulated B+. The negative peaks are bit stronger than the positive. I'm thinking that this asymmetry is what prevents me from getting full 100% positive peaks on the RF envelope. It's supposed to be a 100 Watt module, and the transmitter only puts out 60 Watts of RF, so I would have expected plenty of excess headroom. But, so far I've only connected the module to the 4 ohm winding on the 1628SEA. Maybe it would like 8 ohms better?

[KD1SH]

I experimented a bit more today, connecting the amplifier module this time to the 1628sea's 8 ohm and then 16 ohm windings instead of the 4 ohm. I didn't notice any significant difference in the actual modulating amplitude, but I did notice a bit more distortion of the sine waves when using the 8 ohm winding at 12khz, although it wasn't present when using the 16 ohm winding, oddly. The 4 ohm winding appears to provide the cleanest sine wave.
I am coming to the conclusion, though, that since this amplifier module was designed for 100 watts output into a 2 ohm impedance, it's probably giving me around half that with the lowest impedance winding, 4 ohms, which probably accounts for its ability to just barely but not quite reach full 100% positive modulation with my 100 watt DC input Gonset G-76. Other than that, it's a nice modulator for lower powered rigs; perfectly clean sine waves from 50hz to a blazing 20Khz. Even a triangle wave didn't look too ugly.

[KD1SH]

More experimenting: the Sure amplifier board has two features, mute and standby, which you would ground to activate, but I'd misunderstood the instructions (which, in my defense, were pretty meager) and tied the two terminals together, and this seems to have an effect on the total output. Now, using those terminals correctly, I can achieve full 100% positive modulation of my 100 watt input Gonset G-76. Still, I find this 100 watt amplifier just a bit weak for driving a 100 watt input transmitter. I need to push the amplifier board pretty hard to get 100% positive peaks; almost to the point where some notable distortion kicks in.
Now, on to the next phase: I also have the Hammond 1628sea's big brother, the 1642se, and another Sure class D amplifier board, this one rated for 400 watts audio output into 2 ohms. The 1642se isn't rated by Hammond for that much power - Hammond rates it for 75 watts - but I suspect it's over-designed, based on its formidable weight, and my 400 watt audio board is rated by Sure for 162 watts into a 4 ohm load or 81 watts into an 8 ohm load. Besides, since my plans don't include modulating anything greater than a 100 watt input transmitter, I won't by applying full output to the transformer.
Fun stuff to play with.

[KC2ZFA]

nice work ! do you employ and R and/or C between the amplifier output and the xfmr ? some threads talk about some kind of filtering, other threads do not.

Peter

[KD1SH]

No, I haven't tried any output wave shaping yet, but I might experiment with it. I do notice a very slight tilt to the sine waves - as observed on the modulated B+ - at higher frequencies and when nearing maximum output. It doesn't seem to be notable on the actual modulated RF waveform, though.

nice work ! do you employ and R and/or C between the amplifier output and the xfmr ? some threads talk about some kind of filtering, other threads do not.

Peter

[KD1SH]

My inspiration for this project was these two threads:
http://amfone.net/Amforum/index.php?topic=45973.0
http://amfone.net/Amforum/index.php?topic=47350.0
Both of these threads were started by Tom, K1JJ, and he's got lots of very good information there. I'm still going back to those threads and comparing his results to mine. We both used the same Hammond transformers, but different audio amp modules, and unlike Tom's concept of building the modulator into a transmitter, I figured I'd keep the modulators external, so I can use them with different transmitters.
The cool thing about this concept - and I'd never heard about it before Tom's thread - is that you can not only get fantastic audio quality with standard plate modulation, but do it without having to scrounge for "real" modulation iron, which is getting harder and harder to find.