Kofi Annan in: "Oscillatin' in the Bevois Valley'"

[SY]

What's the square wave frequency in the photo? (Or, better yet, time per division)

The asymmetry between the ringing on top and bottom suggests that the transformer isn't perfectly balanced side-to-side in terms of bandwidth.

Transformer imbalance is very common in less-than-top-end iron. And sometimes there, too. Generally, it's fixed by bringing things to the lowest common denominator, that is, crippling the better side so that it matches the poorer side. If you measure the DCR of the transformer primary from center tap to each end, one will have a higher resistance. That's the side which is likely to be poorer. Across the other "better" half, tack a small high voltage cap, say 680pF/1000V. See if that makes the top and bottom of the square wave more symmetrical.

If it does, you'll have a couple options to get that bit nailed down, one of which is the old Dynaco trick of running a small feedback cap from the UL tap of the "better" side to the feedback summing junction.

 

[EC8010]

Also, I noticed that my plate voltage for the first stage is a little high-- about 310VDC. This could also be causing a problem, I'd guess, so maybe I need to drop a few more volts.

I hope you don't really mean the voltage on the anode of the first stage? I hope you really mean the HT for the first stage that should be 285V? It's a little high, but probably not worth getting too worried about.

Your square wave isn't amazing, but it's probably as good as you're going to get.

Edit: I was going to add "without fiddling with the output transformer" but I see SY has beaten me to it.

 

[Kofi Annan]

What's the square wave frequency in the photo? (Or, better yet, time per division)

The asymmetry between the ringing on top and bottom suggests that the transformer isn't perfectly balanced side-to-side in terms of bandwidth.

Transformer imbalance is very common in less-than-top-end iron. And sometimes there, too. Generally, it's fixed by bringing things to the lowest common denominator, that is, crippling the better side so that it matches the poorer side. If you measure the DCR of the transformer primary from center tap to each end, one will have a higher resistance. That's the side which is likely to be poorer. Across the other "better" half, tack a small high voltage cap, say 680pF/1000V. See if that makes the top and bottom of the square wave more symmetrical.

If it does, you'll have a couple options to get that bit nailed down, one of which is the old Dynaco trick of running a small feedback cap from the UL tap of the "better" side to the feedback summing junction.

Well, let's make sure we're talking about the same thing here. I set the function generator to a 10K square wave with the cool mad-scientist dial thingee to "1.0", which I'm hoping is a multiplier for the frequency which, in turn, would mean this is a bona-fide 10KHz square wave.

So, messing with the transformer may yield some positive results, eh?

Hmmm....

I hope you don't really mean the voltage on the anode of the first stage? I hope you really mean the HT for the first stage that should be 285V? It's a little high, but probably not worth getting too worried about.

Your square wave isn't amazing, but it's probably as good as you're going to get.

Edit: I was going to add "without fiddling with the output transformer" but I see SY has beaten me to it.

No, of course I don't mean the voltage on the anode. You know when you read my posts you have to translate from Kofispeak to actual reality, right?

The voltage on the anode of the first and second plates are 105VDC and 217VDC respectively. Yes, it's a little high and I could certainly drop some voltage by adding some resistance between the "regulator" (I opted for the easy-to-implement choke and cap combo) and the E88CC. If you still believe that this is not high enough to cause concern, then I'll leave it alone.

Kofi

 

[EC8010]

I'd leave the DCs alone. If you're feeling courageous, have a go at the output transformers.

 

[Kofi Annan]

OK-- so I finally received my new capacitors and soldered them in yesterday. Everything is working and the oscillation is gone, but it took a little fiddling even with the measurements.

I had initially used alligator clip jumpers to test the 68pF mica capacitor paralleling the feedback resistor and it worked fine. Once I soldered it in, though, some oscillation appeared. I added a 38pF in series and the oscillation went away.

I thought that this was odd, but I guess maybe I'm picking up some crud near the socket that I didn't pick up when I had the cap alligator clipped into place. Anyway, the oscillation appears to be gone and I think we're in good shape.

I tested the square wave from the right channel for the first time (I had assumed that this would be similar) and it was quite dissimilar to the left channel. One thing I noticed is that the top of the square wave seemed to roll off (it angled downward a bit instead of being horizontal).

I noticed that when I put the amp back in service, the bass was more pronounced such that I had to rearrange the speakers in my room to avoid excessive boominess (this is likely due to the Buschhorns' rear firing bass port in a small room) and that the highs seemed a bit rolled off.

I think I'll reduce the value of the grid stoppers (currently at 1.5K) and see if I get some of the highs back.

Any thoughts?

Kofi

 

[SY]

The grid stoppers will have minimal effect on the audible treble range. Perhaps you've gotten too accustomed to the high frequency hash that some other amps overlay on the music? In any case, live with it for a few weeks before deciding to change things further.

 

[Kofi Annan]

live with it for a few weeks before deciding to change things further.

Good advice. It now sits at a friend's house and it probably ain't coming back for a while. Gorgeous sound.

The help I get in this forum is invaluable. I know I can be a real pill when I don't get it and you're forced into a puppet show, so thanks for your patience.

Kofi

 

[colinB]

Kofi,

I've recently completed the build of Morgan Jones Bevois Valley EL84 PP using pretty much the same Hammond output trafo's as you. As you would expect it is oscillating badly!

As I don't have an oscilloscope or signal generator (yet) would you be willing to post your 'best' set-up for R & C in the feedback loop and input stage anode resistor & Cap (at the E88CC) so that at least I can get somewhere close to stability before I start playing when I get the necessary kit. (I guess my hammond 1608's won't be exactly the same as yours but they should a closer match than to MJ's Leak trafo's so your R& C values will be helpful.)

Thanks for your help
Cheers
Colin.

 

[Kofi Annan]

Well, unfortunately the amp is till at the friend's house, but my booze-addled memory and order history from Mouser may help...

I ordered a pair of 68pF and a pair of 39pF mica capacitors from Mouser, so one of those is in the resistor / capacitor that's in parallel with the 22K load resistor and the other is the feedback capacitor. I also reduced the feedback, which means there's a third variable, which sucks...

OK-- so I'd say that I used a 1.5K resistor / 68pF cap in parallel with the 22K load and the 39pF job parallels the feedback resistor. If that doesn't work, swap the capacitors and try again. This may not work for you, however, as I reduced the feedback, which changed the value of my feedback resistor.

If I get the amp back this weekend, I'll take a look under the hood and let you know exactly what I wound up with.

Sorry I can't be of more help right now, but stay tuned.

Kofi