Understanding scope traces

[defaced]

I've got some scope traces that I'm trying to really understand what they mean with regards to the circuit. From what the few traces I've seen online, it looks like everything is ok until the signal gets to the PI. From there I'm not sure what it's supposed to look like, but my hunch is that my output scope trace should not look the way it does.

So my question is, does these scope traces look they way they should, and if not, is there something from a design standpoint that needs to be fixed?

Input signal 1 kHz, 1 v/div:



Input signal after 68k grid stopper, 1 v/div:



Signal after V1, master volume at 0, 20 v/div:



Signal after V1, master volume cranked, 20 v/div:



Phase inverter, input side, master volume cranked, 20 v/div:



Phase inverter, ground side, master volume cranked, 20 v/div:



Signal into 8 ohm resistive load, master volume cranked, 20 v/div:


Schematic:
PI is actually a 12AU7
http://img398.imageshack.us/img398/6634/entireampschematicwithodw5.gif

[BackwardsBoB]

I'm surprised I didn't hear you run the test...

What is this thing with no tone controls? If it's a guitar amp, you need to scale back your input to around .1V initially. You certainly don't want 80V P-P going into the PI unless you're studying how it flips out.

Are your "signal after V1" traces before or after the master volume? If they are before, the signal shouldn't change that much, and if they are after, the signal should flat-line when you ground the wiper of the master volume.

Where are you grounding your scope probe? Is your circuit ground tied to the chassis somewhere, and is the chassis tied to earth ground somewhere?

Are your "phase inverter" traces at the grids or anodes?

[defaced]

This is a power amp. In a former life it was a Peavey Deuce, but the only thing remaining from that is the iron and the power tube board.

What kind of signal voltage should I be sending to the PI for max clean output to the grids of the power tubes? This is the first amp I've built or scoped so I don't have a good feed for what's reasonable yet.

The "signal after V1" is before the master volume. One thing omitted on the schem is that there's a 10nf cap across the master volume. When that's removed, the signal amplitude doesn't really change, it just flat tops when it distorts.

If I scope after the master volume, the signal is flat when the wiper is grounded and just as pictured when the wiper is maxed. At intermediate points, the positive half of the signal is clipped. This goes away when the volume is maxed. Is there a way to get rid of this clipping?

The preamp and phase inverter use a bus ground, which is taken back to the ground side of C18 in the power supply. That point is where I have my star ground and it is joined to chassis ground. I do not know if the wall outlet is wired properly as this is an apartment. The probe is grounded to the preamp bus ground.

Those phase inverter traces are at the anodes.

Here's what the PI looks like with the input signal scaled back to 0.1 v p-p as suggested, the master volume is wide open:

Grid, input side, 5 v/div:


Grid, ground side, 5 v/div:


Anode, input side, 10 v/div:


Anode, ground side, 10v/div:

[loudthud]

I think you need a little more B+ at the PI. The 22K R38 should be something like 4.7K. The thing you have to watch out for is the voltage on C15 which may go over it's voltage rating before the tubes warm up when the amp is turned on (no standby switch is shown). Most caps can tolerate a brief overvoltage surge but they live longer if you can avoid it. The second group of waveforms look OK but I don't think you have enough available to drive the outputs to full power. You need to be able to get about 2x the bias voltage peak to peak on each side of the PI.

[BackwardsBoB]

OK. I get it now.

For a power amp, you want to be able to handle a line input, which will be somewhere around 1V P-P. Some headroom for stronger and weaker inputs is always good.

If you turn up the voltage on the PI by dropping R38, the preamp voltage will go up too, and I bet it's a bit higher than intended already with the missing preamp section. You can increase R39 to drop it back down. R39 has less current through it, so you need to increase R39 by more than you decrease R38.

Since you have a scope, you can have fun varying R8 in the PI. This will change the operating point and gain in the PI.

The preamp looked good at 2V input. You need to be able to get around 80V P-P into the power tube grids to get full output. With a 1V input, this might happen with the volume control less than all the way up.

How does it sound?

[Rob Mercure]

Hey Defaced,

What are you using to feed the signal into the amp? If it's a commercial signal generator you should have a square wave output available. If you were to send the same information while using a square wave input it would be easier to analyze the amp's frequency response and to determine if there is any parasitic oscillation "ringing" in any of the stages.

Some of the other foks may work well with sine waves but there's scads of information out there - google "square wave analysis" - on breaking down what's going on using a square waves. I admit that I'm so "spoiled" that trying to analyse the trace's you supplied makes me want to reach over and switch the signal generator to square wave <grin>.

Just my 2 cents but I'd like to see what your amp does with a flat top.

Rob

[loudthud]

One thing that you should do if you are using a 10x scope probe is adjust the compensation with a square wave, usually 1KHz. Somewhere on the probe body or the connector you should find a little screw adjust variable cap. It compensates for the input capacitance of your scope. Misadjustment can lead to measurement errors of 2x or more on sine waves that look perfectly normal except the amplitude is wrong.

Running square waves through an amp takes a little getting used to. You want to use small amplitudes like 1/2 to 1/10th full power and look for ringing on the leading edge. But make sure your generator and scope are working correctly first.

[Mark Black]

I'm a fan of the square wave too as well as sine wave input. Squares can tell you a lot about the frequency response of the amp and really let you see EQ controls working, and then of course you want to use sine for onset of clipping and power measurement.

Square waves also seem to be good at poking an amp into misbehaving if it's on the edge of an oscillation problem or something.

[Rob Mercure]

Mark,

I think we pecked our way out of the same egg on this one <grin>. Generally there's nothing that a sine wave reveals that I can't hear as I increase voltage from a signal generator or monitor voltage with a DMM (or a VTVM years ago). You can generally hear the onset of clipping when it hits 10% or so and you can usually see it on a 'scope around 3-5% - while that seems like a lot of difference for audio trouble shooting it's "ball park" for me. But with a square wave you can pick up parasitics and power supply funk and udder such like an impedance mismatch between stages that creates a treble cut when the volume control is turned down - and that just the "few instances."

I guess if one has an audio sweep generator sweeping a sine wave would be great but I've never had one so I'm "square wave dependend" (is there a 12 Step Group for this? <grin>)

Rob

[defaced]

Hmm, I think I figured it out (thanks!!!) - finially, after months of messing with it. The problem I was having is that the amp sounds ok, it just had very low output. Using everyone's suggestion and shooting for 80v P-P on the PI plates, I ended up taking up the PI B+ to the 440v range, dropping the plate resistors to the 12K range, and getting approximately 140 plate volts on one side of the PI and 110 on the other side. All of these voltages are approximates because that amp isn't one the bench per se right now. Being that none of these values are close to what I see in schematics (this may be due to my using an AU7 at the moment), I probably really screwed up the operating point of the PI. During these alterations, I did not adjust the 470 ohm (R8) resistor, so I'm going to go back and see what playing with that does - before it didn't help matters.

Now on the topic of square waves, no, I can't produce a "proper" one. My signal generator is the audio software Sound Forge. The software has a signal generator in it which can do sine, absolute sine, square, triangle, and different kinds of noise (pink, white, and brown I think). In the software, the square waves look square, but when you scope them you can see humps on the corners of the wave from the harmonics being added together (http://upload.wikimedia.org/wikipedi...sis_square.gif).

So in the spirit of square waves, how does this little circuit look? http://www.talkingelectronics.com/ht...reWaveOsc.html