scope waveforms?

[teemuk]

I think i know...rounded is clean, sharp edges show it's starting to clip and at full clipping it becomes more or less square. I've observed this so far.

That's a good basis. Yes, that is pretty much what happens in clipping where higher amplitudes are limited.

However, when you EQ that kind of signal with a non-linear frequency response you will notice that the effect of the EQ will alter the waveshape quite drastically. Usually anything resembling "squares" will dissappear because the "flat", DC-like, signal components are attenuated and rising edges are either enhanced or slew limited. This usually does quite a bit of mangling and the waveshape usually ends up looking very different from the usual idea of how a clipped signal looks like. e.g. try googling how a clipped waveform looks like after the tone controls and how it responds to changes of tone controls.

Another issue is feedback. In an amp where you have stages enclosed into a global feedback loop the stages within often try to error correct and they do this by distorting their output signals, basically to counteract the distortion in the overall circuit. So, when you run PA to clipping you may examine clipped waveforms in PA output and waveforms in PI output that look like being distorted. They are not always. The PI's just tries to correct the distortion and increases gain to compensate the clipped portions. This may again look like distortion (because people automatically tend to assume the PI should output just a higher magnitude version of what goes in it) but in essence the PI's actually operating normally and doing it's thing with negative feedback.

Scope just allows you to see the waveform. Equally important point is understanding what you see.

Oscillation? That usually looks like very high frequency signal added up to the original signal. It can be continuous or spurious, where you have just little, short spurs of oscillation striking at signal points where amp's operation is unstable (e.g. clipped or crossover regions, edges of the waveform, etc.)

[daz]

I can see i have a lot to learn. But i likely won't anytime soon since the amps sounds as good as it is now. I will however play with it and post some waveforms to ask what i'm seeing and get thoughts on where to look based on what i'm posting. I know if i'm told there are issues i will no doubt want to address them. I just hope this doesn't cause me to redo the entire amp. I'm a big believer in if it ain't broke don't fix it. But i'm helpless when it comes to screwing with this thing when i have an idea or am given one.

That's a good basis. Yes, that is pretty much what happens in clipping where higher amplitudes are limited.

However, when you EQ that kind of signal with a non-linear frequency response you will notice that the effect of the EQ will alter the waveshape quite drastically. Usually anything resembling "squares" will dissappear because the "flat", DC-like, signal components are attenuated and rising edges are either enhanced or slew limited. This usually does quite a bit of mangling and the waveshape usually ends up looking very different from the usual idea of how a clipped signal looks like. e.g. try googling how a clipped waveform looks like after the tone controls and how it responds to changes of tone controls.

Another issue is feedback. In an amp where you have stages enclosed into a global feedback loop the stages within often try to error correct and they do this by distorting their output signals, basically to counteract the distortion in the overall circuit. So, when you run PA to clipping you may examine clipped waveforms in PA output and waveforms in PI output that look like being distorted. They are not always. The PI's just tries to correct the distortion and increases gain to compensate the clipped portions. This may again look like distortion (because people automatically tend to assume the PI should output just a higher magnitude version of what goes in it) but in essence the PI's actually operating normally and doing it's thing with negative feedback.

Scope just allows you to see the waveform. Equally important point is understanding what you see.

Oscillation? That usually looks like very high frequency signal added up to the original signal. It can be continuous or spurious, where you have just little, short spurs of oscillation striking at signal points where amp's operation is unstable (e.g. clipped or crossover regions, edges of the waveform, etc.)

[daz]

[daz]

ok, heres one waveform. I've looked at lots of things but i only shot this one because i'm unsure of whats pertinent. But at the pre PI master the form looks like this. i have much more to say but i have to make this short because i seem to only have access to the site for a few minutes and i fear i will try and post this and no longer have access. So real quick, this is the wavform at the master but it seems like the gain stages give ma a round wav that gradually gets pointy which i believe is right. At the CF is where things seem to get squared off even with the gain pot after the 1st stage way down. It's round till about 8:30 then quickly goes to a form of square or pointy after the CF. More to say if and when the site is working normally again. I soldered a 1uf cap to the CF input to look at the signal there and it's isn't near as bad there, so it's happening somewhere after that.

[Chuck H]

That looks pretty good actually. The interesting thing is that there is an obvious dip and recovery. I usually see this on power tube circuits. It's probably an affect of the cathode followers current (even though it's not driving a load?). There's also a nice amount of asymmetry in the duty cycle that usually translates to something good tone wise. That wave form would explain why you're getting the tones you want at lower volumes. For a preamp distortion wave form that actually looks really good daz

[daz]

Hmmm....i thought that was bad and that it should look more rounded. It is a killin sounding amp tho. When at it's best, which it is at the moment, it's better feeling and sounding than any marshall i've had, and i've had most up to the late 90's. But can you tell me what to look at in order to understand why the PI caused the amp to sound better and have less OD and more dynamics with a 2k cathode in place of the 470R ?

[daz]

Ok, maybe this is a issue, you tell me. If i connect the scope to the speaker out, that wave stays like that only till 9:30-10:00 on the master after which that dip you mention becomes a straight line that stays like that all the way to max volume. Does this maybe mean i need to reduce the signal to the PA? It IS a very loud amp. If so, where would be the best place/way to lower it after the master w/o altering the tone, or at least minimally ?

[Chuck H]

But can you tell me what to look at in order to understand why the PI caused the amp to sound better and have less OD and more dynamics with a 2k cathode in place of the 470R ?

Nope. But YOU can wire the amp both ways, take scope notes at different stages and see for yourself what might have changed now

[Jazz P Bass]

That is a very nice looking 'distorted' signal.

[loudthud]

One thing frequently missed by the novice scope user is Probe Compensation (sometimes called probe calibration). An X10 or X100 probe will have an adjustment somewhere to adjust for flat frequency response which will change based on the input capacitance of the scope. A probe must be connected directly to the scope's input jack, you cannot use any kind of extension cable. The usual procedure is to connect the tip of the probe to a source of 1KHz square waves (many times on the front of the scope) and adjust a trimmer capacitor on the probe for flat tops. A mis-adjusted probe can make a perfect square wave look almost like the waveform daz posted in #19 above. It will also distort the amplitude of sine waves.

The scope came with an X1-X10 probe. Verify that you get a good looking square wave in both positions of the switch. Turn up the Volts/division switch until you get a nice big waveform on the screen of 4 to 8 divisions with tops that are dead flat.

[Chuck H]

The scope came with an X1-X10 probe. Verify that you get a good looking square wave in both positions of the switch. Turn up the Volts/division switch until you get a nice big waveform on the screen of 4 to 8 divisions with tops that are dead flat.

With the 1kHz square wave. Not your amp daz. Then you can be sure what the scope is telling you is accurate. Your scope seems to have this built in. It's the little metal tab on the front marked "cal". Just touch your probe tip to it.

[Enzo]

And there is a tiny hole on the side of your probe with a screw slot for trimming. Turn that back and forth to get the tops of the square wave flat, not curling up or down on the ends.

[daz]

not sure what you mean chuck. I touch the tip to that calibrate thing and whats supposed to happen? Enzo....tried that but not sure what you mean ether. Theres always some degree of curvature ate the corners. Are you just saying the top should be level? It will go MORE level, but not 100% and the corners never get totally square unless i find a place in the amp where the wave is more square.

Also, the tone controls effect the shape hugely. I just have them where i usually play it. Where should i set the amp to check things anyways?

[nashvillebill]

Touch your probe to the Calibrate post on your scope (hook it on there if you can). Now set your scope with the v/div and time dials until you see the square wave. Now with a trim pot screwdriver adjust the screw on the scope probe until the signal on the scope actually looks like a square wave. Your amp is not involved in this procedure.

[Chuck H]

Do not use the amp for the probe adjustment. When you touch the probe to that calibration point there should be a 1kHz square wave present. It may be such a pure square wave that it looks more like two dashed lines with little or no rise and descent lines visible.