I sent Dave a note and asked him if he would show how a NPC/RC mod should look on a scope. I know that a scope cannot detect the asymmetry of the form; just what the scope reading should look like in general. As well as discuss other pros and cons of using a scope. We will see what happens.
Maybe unit248 can do it instead?
Sadly, I don't have a sacrificial radio on which I could perform an NPC mod for demo purposes. The 3 HF transmitters I have are my Kenwood rigs, which I don't want to chop up in the first place, and which are all low-level modulated. The TS-430 and TS-850 both do AM using an AN612 balanced modulator chip, and the TS-950SDX does AM modulation inside the DSP unit. I'm not sure that the typical RF-stage NPC/RC mod would be feasible with this kind of design.
However, a while ago I did capture a few images from the TS-850 (with the DSP-100 unit turned off) showing what low-level over-modulation looks like. (Be warned: these images are kind of big.)
This first photo is an unmodulated dead carrier:
http://people.freebsd.org/~wpaul/ts-850/modulation/balanced/PICT0179.JPG
The second photo shows the same carrier at 100% modulation:
http://people.freebsd.org/~wpaul/ts-850/modulation/balanced/PICT0180.JPG
The third shows over-modulation:
http://people.freebsd.org/~wpaul/ts-850/modulation/balanced/PICT0181.JPG
These were captured using my Kenwood SM-230 station monitor. It's designed to operate as an IF band scope (with either the TS-850 or TS-950 rigs), TX signal monitor and a simple single-channel scope (10Mhz). It also has a built-in two tone audio signal generator for testing, which is what I used to generate the 1000Hz tone to modulate the carrier in the above pictures. The TX signal monitor is basically the same as scope mode, but the signal is sampled via the PL-259 input/putput connectors on the back (you put it in-line with your coax). In regular scope mode, you can just use a standard probe with the BNC input in the front.
Unlike high-level over-modulation, low-level over-modulation produces "bumps" at the negative peak points where a high-level over-modulated signal would just flatten out. The distortion you get is different, but still undesirable. This is an artifact of the balanced modulator chip behavior. Note that the 100% modulation image actually shows very slight over-modulation: if you look very closely you can see the beginnings of some small bumps. I didn't quite get the mic gain control set right.
Although the TS-950SDX uses the DSP for modulation, it produces similar results when over-modulating, so I always use the SM-230 in TX monitor mode when setting the mic gain control to make sure there's no distortion. (This is a good example of where "playing it by ear" just doesn't cut it: it's very hard to tell just by listening if the audio is really clean. There's no missing it with the scope though.)
It think that with this type of transmitter, the only way to get negative peak compression might be to do it at the audio stage, and then use RF amplification to make up the difference. I don't have the right kind of audio processing gear to do that though. Maybe some day.
As far as being able to see asymmetric modulation on a scope, yeah, you should be able to do it, but I think you sort of need to sit and carefully count the divisions in the reticle in order to gauge it. With a symmetrical sine wave, you know the crossover point is right between the peaks, but once you start compressing the negative peaks, it can be a little hard to just eyeball it. I think the best way to do it is to start with a dead/unmodulated carrier and set the transmitter power and/or scope signal attenuator controls so that it fits exactly between to major divisions on the screen. Those division lines are the crossover points for the audio waveform that you'll be superimposing onto the carrier. Once you've made a note of those reference points, you can start cranking up the mic gain until you reach 100% modulation (the positive peaks should now be two major divisions wide instead of one). Then as you keep turning the gain up, the positive peaks should keep growing and ideally the negative peaks should remain at their current positions and retain their curved shape, but start to look more "smushed" compared to the positive peaks.
There's one thing I've never been to clear on about this though: some NPC mods prevent the negative peaks from reach 0 output entirely (i.e. from letting the waveform on the scope turn into a flat line), but still allow the negative peaks to attain a flattened shape. But supposedly it's that flat shape that causes most of the unwanted harmonics to be generated (the closer you get to a square wave, the more harmonics you get), so I would think it would be better to compress the negative peaks while preserving the sinusoidal shape too. But people don't seem to worry about maintaining the curved shape though, and I'm not sure why.
My guess is that it's probably preferable to use the latter approach, but it's technically easier to accomplish the latter, and while allowing the negative peaks to flatten out is not as good as keeping them curved, it's considered a reasonable trade-off.
-Bill
