Another Scope Question

[surfsup]

Long post. Liverpool-ish build. 2xEL84 output. 100mV input signal. Gain/T/B/MV all max (no mid pot, just a 10k). 335 V plates, 329 V screens.

PI has 1k2 bias R and 47k to ground. I put all the scope shots into one JPG for ease to upload so here are the descriptions of the top 6 scope shots:

LEFT = Output at speaker vs one side of PI
RIGHT = Both sides of PI

Top Row: Gain=5, PI is clean and symmetrical, output is clipping. You can see the Vpp on each trace is exactly 24.08V...

2nd Row: Gain=8, PI is well into distortion and signal is stronger on one side (40Vpp vs 35Vpp), about 15% increase due to clipping. PI still resembles a normal wave on the negative blue side

3rd Row: Gain=10, PI on negative blue trace between 8 and 10 has this shelf that develops because the latter half of the peak gets squashed but not the forward half. Also on the output there seems to be some oscillation due to this on the yellow positive trace

4th Row: Ignore this shot, I just realized I might not have had MV at 10 for this one...crap.

Anyway, if I strum the low-E/A/D strings lightly, I get a flubbiness. If I play a power chord normally, it sounds great, smooth, very nice. So its like if I don't input enough signal, its flubby, but great if I jam on the strings...???

If I play a high string, I get a little harshness (at least what I perceive to be a slight harshness) at the trail end of a single string note. If I play two strings at once, the harshness is more prevalent and up front at the picking of the strings. The decay of the note is definitely buzzy.

For filtering, scoping the B+ rail AC:
Plates: 13.0 Vpp
Screens: ~80 mVpp
PI: ~50 mVpp

I have tried different CCs on different stages, different grid load Rs, I have put in snubbers across Ra's, removed them, put em back, changed the caps in the tone stack multiple times, etc. So I've tried lots of stuff and it is about as good as it is going to get at this point without outside help from someone.

-Output is not symmetrical, to cure this should I bias the tubes a bit more centered on the load line?
-What could cause the negative wave-form to do that and could this be the source of my problem?
-Is that an oscillation and what techniques besides a zobel can cure that?

[img:960:936]http://chicagocadcam.com/ChrisHahn/scope/ScopeShots.jpg[/img]

[Joe L]

Your waveforms look a lot better controlled than what I am getting out of a Marshall 18 watt build. I was having problems with a huge 10v bias shift and crossover distortion before I added the "Paul Ruby Zener Mod" which is detailed in the file section of 18watt.com. But I'm not seeing any crossover distortion on your waveforms. I wonder why not?

Anyway, from the research I have been doing, I am suspecting the output transformer I used. Which are you using?

..Joe L

[surfsup]

OT is CXPP25-MS-8K edcor, rated 25watts so this 2xel84 should be no problem for it, regarding actual output.

I do have another OT but geez it is going to be tough to swap it. I'll try. Any other ideas? Maybe i'll try that rubymod. I'll post some shots of the output in more detail with different frequency inputs later this weekend (i hope)

[surfsup]

Actually I should be able to swap the OT fairly quickly. I don't even need to really unmount the one I have now. Just unsolder the ct, two plate supplies, and the 8R tap. If the other's leads are long enough.

Here's some more shots, to my untrained eye, the other frequencies I used as an input signal seem to be fine but the 1kHz has that hump in the negative trace.

[surfsup]

A better shot of the 1kHz...

[Joe L]

A better shot of the 1kHz...

From what I can see, the time interval is about 100us (10KHz) on that spike which is the tenth harmonic of the 1K signal. That can very well be what you are hearing and I don't think you are going to get it out of there if you are not seeing it on the input to the grids of the power tubes.

I am using a different brand of transformer than you. However, as a way to get back in to bench work, a tech friend and I built matching Deluxe Reverbs. Everything was identical except I used Heyboer iron and he didn't. His amp sounded exactly like mine on the same speakers except when we would drive it really hard, his would buzz on the trailing notes and mine would not. After I built this 18 watter, using the same brand he did, I heard the same type of buzz.

From the tests I have done (speakers, Paul Ruby Mod, etc.), and also the opinion of Paul Ruby, is that some transformers pass the nasty harmonics and some don't. I have a GDS OT coming in a day or two and his are made by Heyboer and reportedly don't buzz.

I'll let you know how it turns out. In the meantime, I don't see anything bad broke with your waveforms. Maybe just too much bandwidth on the top end.

ps. Just looked up your tranny. Looks very nice... if you are building a hifi. - "Frequency Response 20~20K Hz., <1dBu"

..Joe L

[Structo]

Hi Joe,

Can you tell me what brand OT that was that buzzes?
I seem to have some of that with an amp I built and I want to rule out the OT.

You can PM me if you don't feel comfortable posting the name on the open forum.

Thanks,
Tom

[surfsup]

Didnt swap the ot yet. But, i was testing stuff and noticed after a CC between the PI and the dual gang MV, i got upwards of 4vdc on my dmm. On the other side, i was getting 300mv tops. The side that was getting the several volts dc was the same side the screen was being fed from the previous stage. So thinking the cap was bad, i replaced it. Afterwards, i dmm'd it again and same result. Anywhere from 3-4 vdc after the .022uF 630v coupling cap...is this normal to have several volts dc on this one side of the PI after the cc?

[Joe L]

Anywhere from 3-4 vdc after the .022uF 630v coupling cap...is this normal to have several volts dc on this one side of the PI after the cc?

It should read 0 with no signal present. If you disconnect the capacitor from the pot and check with your voltmeter to ground you should read 0 volts. If so, you should still read 0 volts to ground across the pot and at the control grid of the power tube. If you have voltage on the tube, it is bad. Swap the power tubes and see if the problem follows the tube.

And I did get my new transformer in from GDS and the high frequency trash that was present with the old transformer now sounds much more musical. I am going to try it out tomorrow and see how much difference the Paul Ruby mod makes with it but I have a feeling I will put it on a switch labeled "Smooth" and be done with the amp.

The next transformer I am going to try is from Trinity Amps but that might not be until the next 18 watt build.

..Joe L

[surfsup]

There's <50mV on my DMM (not quite zero) at idle. The 3-4 volts is when I am playing. At the power tube there is no DC, from memory, but now I'm going to re-check that this weekend. The laws of physics seem to be different on my bench...

[butwhatif]

At a quick glance it looks like a CC plus bias on preamp tubes--you will have to experiment with that stuff. See what it does w/ a sig directly into the PI and take it from there. if u r lookin for bal, then start there.

[Joe L]

There's <50mV on my DMM (not quite zero) at idle. The 3-4 volts is when I am playing. At the power tube there is no DC, from memory, but now I'm going to re-check that this weekend. The laws of physics seem to be different on my bench...

Nah.... Don't worry about the DC you are reading with a signal. Just so you have ~0 volts at idle. The signal screws up any DC reading you take and all the distortion screws up any AC measurements.

In order of effectiveness and cost, I would try:

Increase the size of the power tube grid leak resistors - they can be up to 300K with EL84s using cathode bias. This reduces the current available to charge the CC when drawing grid current.

Try the "Paul Ruby" mod to prevent bias shift/blocking distortion.

Get a "better" output transformer.

..Joe L

[surfsup]

I should clarify, the OT is not the problem (well I'm 90% sure at this point but haven't built enough amps to be 100% yet). I discovered a problem with a cap I had installed on the first tube for the VVR scaling. The cap was going bad, eventually (thankfully) it completely shorted and I was getting no sound from the amp, so when I scoped it, I saw nothing was coming out of V1. So I knew to look there. I rewired the whole socket for the input tube, and now the amp is much more stable. Unfortunately, until that happened, the scope wasn't showing me the problem which I thought the scope would!

So I still have a little "looseness" on the lows and a little fizzy on the highs. Which could still be OT related but nothing like it was doing before, and its better than some of my other amps - but I will continue my search for perfection to my ears and learn along the way.

I will try the PaulRuby mod. I gotta see if I have the right components "in stock" or if I gotta pay shipping once again...

[Joe L]

That is good that you found that cap. That sure would have caused the distortion at lower levels.

As far as the Paul Ruby mod, I found that 16v zeners were optimum. To find that value, I injected a 200Hz signal while monitoring the grid of the power tube. Without clipping the preamp/PI, I could see the grid current clipping the upper portion of the signal. I used the zener value that would clip the lower peak of the signal just after the upper portion starts clipping.

Just as Paul Ruby says, clipping the signal at that point doesn't affect the output because the power tubes have passed cutoff. What it does do is keep the CC from charging up from the grid current. That stops bias shift and blocking distortion.

My bias shift under hard distortion before - -10V. After - -.3v. And no more crossover distortion.

And creds are due to Paul. His discovery is worth adding to the textbooks.

..Joe L

[surfsup]

JoeL, thanks so much for this quick explanation! I will check that PRMod out this week (almost missed this reply of yours) and scope my amp to see what's going on. I don't see crossover distortion on the waveforms of my output but what the heck...