Page 19 of 21

#271 Re: Dieselpunk Preamp of Doom and the Beast with Two Backs

Posted: Sat Nov 09, 2019 5:09 pm
by izzy wizzy
Might be a silly question but have you swapped valves from the good one to bad one yet? Or tried another valve in that position that oscillates?

#272 Re: Dieselpunk Preamp of Doom and the Beast with Two Backs

Posted: Sat Nov 09, 2019 5:52 pm
by Thermionic Idler
Not yet, but I think I have identified the culprit. The question is - what is the solution?

Shorting pins 2 and 3 of the XLR input, i.e. the primary side of the line input transformer, stopped the oscillation dead in its tracks. This was consistent and repeatable. As I said previously, there is a length of twisted pair interconnect between the XLR and the transformer on this amp, but not on the other one.

My suspicion is that the oscillation problem will go away if these are now connected to the preamp. However, clearly we have some interaction between the feedback loop going into some sort of resonance with the transformer secondary, when any length of paired wires are connected to the primary, but left open circuit at the other end. I suspect on the other amplifier, the wires being much shorter wasn't enough to 'tip' it into oscillation.

The question is now how to resolve it, as this could potentially damage tweeters I need to ensure it's safe agaiinst having the input cables pulled out. Something as simple as a resistor or capacitor across the primaries?

#273 Re: Dieselpunk Preamp of Doom and the Beast with Two Backs

Posted: Sat Nov 09, 2019 6:05 pm
by Ray P
Why not just use a cheap sacrificial speaker?

#274 Re: Dieselpunk Preamp of Doom and the Beast with Two Backs

Posted: Sat Nov 09, 2019 6:11 pm
by Cressy Snr
Thermionic Idler wrote: Sat Nov 09, 2019 5:52 pm Not yet, but I think I have identified the culprit. The question is - what is the solution?

Shorting pins 2 and 3 of the XLR input, i.e. the primary side of the line input transformer, stopped the oscillation dead in its tracks. This was consistent and repeatable. As I said previously, there is a length of twisted pair interconnect between the XLR and the transformer on this amp, but not on the other one.
Input/ interstage transformers and feedback are not the best of bedfellows.

#275 Re: Dieselpunk Preamp of Doom and the Beast with Two Backs

Posted: Sat Nov 09, 2019 6:11 pm
by izzy wizzy
Thermionic Idler wrote: Sat Nov 09, 2019 5:52 pm Not yet, but I think I have identified the culprit. The question is - what is the solution?

Shorting pins 2 and 3 of the XLR input, i.e. the primary side of the line input transformer, stopped the oscillation dead in its tracks. This was consistent and repeatable. As I said previously, there is a length of twisted pair interconnect between the XLR and the transformer on this amp, but not on the other one.

My suspicion is that the oscillation problem will go away if these are now connected to the preamp. However, clearly we have some interaction between the feedback loop going into some sort of resonance with the transformer secondary, when any length of paired wires are connected to the primary, but left open circuit at the other end. I suspect on the other amplifier, the wires being much shorter wasn't enough to 'tip' it into oscillation.

The question is now how to resolve it, as this could potentially damage tweeters I need to ensure it's safe agaiinst having the input cables pulled out. Something as simple as a resistor or capacitor across the primaries?
When I first started in this hobby, my first build was an ARC SP3 copy. I built amp after amp after amp trying to get rid of an oscillation that caused the power amp to burn up, literally, blue smoke came out of it when I hooked up the SP3. I rebuilt this thing countless times using different parts until one day there was only one common thing left; a twisted pair from the output to a pair of phono sockets. Changing that stopped the oscillation. Not saying that's it for you but I'd swap valves around first and then maybe swap out the wire for something else. There's no need to use anything fancy or shielded but use something different. And not a criticism of your soldering, but it could be a dry joint.

Cheers,
Stephen

#276 Re: Dieselpunk Preamp of Doom and the Beast with Two Backs

Posted: Sat Nov 09, 2019 8:26 pm
by Wolfgang
My suspicion is that the oscillation problem will go away if these are now connected to the preamp. However, clearly we have some interaction between the feedback loop going into some sort of resonance with the transformer secondary, when any length of paired wires are connected to the primary, but left open circuit at the other end. I suspect on the other amplifier, the wires being much shorter wasn't enough to 'tip' it into oscillation.
In the case of the TS 300B OTL PP this oscillation happens only with one monoblock (both built exactly the same). It starts as soon as it's connected to the preamp but only with Nordost type interconnects. With Acoustic Zen interconnects for example no problem.

#277 Re: Dieselpunk Preamp of Doom and the Beast with Two Backs

Posted: Sun Nov 10, 2019 6:23 am
by Thermionic Idler
Happy days!

So later on last night, I connected up both amps to the preamp, but connected the load resistors instead of the speakers. The bias was set, and stayed absolutely rock steady. I then put the oscilloscope on both channels to monitor them for over half an hour to ensure nothing untoward happened. There was no hint of any oscillation or other nasties, both load resistors stayed cool as a cucumber. So this is definitely an issue that only occurs when nothing is connected to the line inputs.

We then took a deep breath, hooked up the speakers and played some music, still with the scope on the outputs for the rest of the evening to monitor things. Oh yes - these are REALLY good - we got more of the same in stereo that was apparent when we tried the first one on its own. I'll do a more detailed write up later when they've shaken down a bit.

So as long as they are not powered up without being connected to the preamp then everything is fine - nevertheless I think it would be wise to see if there is a way that can be mitigated, without degrading the sound. At least XLR connectors 'lock' into place so there's less likelihood of them being accidentally yanked out. EDIT - I suppose one simple workaround would be to make a pair of shorting XLR plugs to put in place when the amps are travelling to somewhere like Owston.

Image

#278 Re: Dieselpunk Preamp of Doom and the Beast with Two Backs

Posted: Sun Nov 10, 2019 8:19 am
by Ray P
Result! :D

and in time for winter too...

#279 Re: Dieselpunk Preamp of Doom and the Beast with Two Backs

Posted: Sun Nov 10, 2019 8:45 am
by izzy wizzy
Fantastic. Remarkable project and attention to detail. You must be very pleased. They look amazing. Can we see some close ups?

#280 Re: Dieselpunk Preamp of Doom and the Beast with Two Backs

Posted: Sun Nov 10, 2019 9:59 am
by Cressy Snr
Now those are ultra-shoddy! :D

#281 Re: Dieselpunk Preamp of Doom and the Beast with Two Backs

Posted: Sun Nov 10, 2019 1:23 pm
by jack
Stunning! Well done!

#282 Re: Dieselpunk Preamp of Doom and the Beast with Two Backs

Posted: Sun Nov 10, 2019 4:20 pm
by IslandPink
Top notch !

#283 Re: Dieselpunk Preamp of Doom and the Beast with Two Backs

Posted: Sun Nov 10, 2019 6:43 pm
by Thermionic Idler
Thanks everyone. We’re breaking them in with the Lord Of The Rings trilogy :D

#284 Re: Dieselpunk Preamp of Doom and the Beast with Two Backs

Posted: Sun Nov 10, 2019 11:18 pm
by Mike H
Now, when I put the voltmeter probe on V1A pin 6 (the anode of the first stage) the oscillation stopped dead, and I got a 105V measurement against a spec of 100V. Could that have any significance?
Which I think added a bit of capacitance to the anode, and loaded it in the HF / ultrasonic region, this reduced the gain in that region and that's what stopped the oscillation. Because self oscillation is a combination of positive feedback and a gain > 1, both of which are easily done in the HF range by phase delay through the amplifier and in the NFB line.

Now I'd be thinking about maybe adding a small value capacitor between V1A anode and the plus side of C10. Say 47pF or 100pF. This is not unusual and I've seen old designs with exactly this.

#285 Re: Dieselpunk Preamp of Doom and the Beast with Two Backs

Posted: Mon Nov 11, 2019 12:39 pm
by Thermionic Idler
Mike H wrote: Sun Nov 10, 2019 11:18 pm
Now, when I put the voltmeter probe on V1A pin 6 (the anode of the first stage) the oscillation stopped dead, and I got a 105V measurement against a spec of 100V. Could that have any significance?
Which I think added a bit of capacitance to the anode, and loaded it in the HF / ultrasonic region, this reduced the gain in that region and that's what stopped the oscillation. Because self oscillation is a combination of positive feedback and a gain > 1, both of which are easily done in the HF range by phase delay through the amplifier and in the NFB line.

Now I'd be thinking about maybe adding a small value capacitor between V1A anode and the plus side of C10. Say 47pF or 100pF. This is not unusual and I've seen old designs with exactly this.
Thanks Mike - might be worth a try. This bit is on the PCB so adding it could be a touch fiddly. I also wondered about adding a grid stopper to the first stage, although the Jensen application diagram doesn't show one.