Hi,
I've built one channel of this amp PP El84
It sounds very fine, but I've a strange problem:
I noticed by observing the woofer that the cone "breathes" (this term is not very technical but it gives a good idea), that is, it oscillates slowly with a constant excursion of 2-3 mm with a very very low frequency I would say, by eye, about 1 cycle per second (1 Hz) but no noise comes out of the speaker. I measured the output with the multimeter and there is a voltage that fluctuates, between 0 and 200 mV.
Some ideas on possible causes?
Thanks in advance to those who want to help me.
I've built one channel of this amp PP El84
It sounds very fine, but I've a strange problem:
I noticed by observing the woofer that the cone "breathes" (this term is not very technical but it gives a good idea), that is, it oscillates slowly with a constant excursion of 2-3 mm with a very very low frequency I would say, by eye, about 1 cycle per second (1 Hz) but no noise comes out of the speaker. I measured the output with the multimeter and there is a voltage that fluctuates, between 0 and 200 mV.
Some ideas on possible causes?
Thanks in advance to those who want to help me.
It's called motorboating. Try adding another capacitor in parallel to the 2nd capacitor in your power supply. And/or decrease the first cap in after the rectifier.
OK so with you have a few coupling caps and the transformer then is possible at LF 2-3Hz that the negative feedback becomes positive. Its a common problem. Mona suggestion of making the cathode caps bigger may well help. The other thing to try is to increase the coupling caps the grids of the EL84 say .47uF. The 100uF at the input will also have its part to play. Suggest you try changing each in turn and record what makes it worse or better. I would not have thought changing the power supply caps would have much affect as they or not in the feedback path and the PP will reject the supply. However it does not take much to get the unit just into stability or instability - you are just on the edge. The problem tends to occur when the LF break frequencies of the coupling and transformer are similar. Make one pole dominant is the trick.
Last edited:
The first question is about the differences between the schematic and what you are building. Are you using a Dynaco transformer? Any other departures from the PS specifics?
Motor-boating is often caused be ground-loops.
Are the cathodes of the triodes connected as shown? E
Are the cathodes of the triodes connected as shown? E
Thanks everyone, I really appreciate it.
The differences between the schematic and my building are:
- the output transformer it's Hammond 1608A with very similar characteristics to Dynaco 410 (same PP 8k primary and same output 16-8-4 ohm)
- The PSU consist in a transformer with one 230 Vdc secondary, full wave bridge with FRED diodes and the rest is the same. B+ it's 300 Vdc under load and the HT are 6.3 Vac.
This configuration it's only a prototype to evaluate the sound (et problems) before final stereo construction like the schematic but with Hammond 1608A output transformer.
I do some tests as you suggested. I try first with another 100u cap in parallel to the other 100 (200uF in total).
The differences between the schematic and my building are:
- the output transformer it's Hammond 1608A with very similar characteristics to Dynaco 410 (same PP 8k primary and same output 16-8-4 ohm)
- The PSU consist in a transformer with one 230 Vdc secondary, full wave bridge with FRED diodes and the rest is the same. B+ it's 300 Vdc under load and the HT are 6.3 Vac.
This configuration it's only a prototype to evaluate the sound (et problems) before final stereo construction like the schematic but with Hammond 1608A output transformer.
I do some tests as you suggested. I try first with another 100u cap in parallel to the other 100 (200uF in total).
Yes the cathodes of the triodes are connected as shown. Pin3 connected to pin8 with 100R resistor
Is your prototype construction on a steel chassis or just a strungout breadboard circuit?
> about 1 cycle per second
Change the 0.1u caps to 0.02u.
That's still full gain at 50Hz but 5X less loop gain at 1Hz, which may be enough to dodge the instability.
Other possibility: "N"FB phase reversed. Swap the connections to the two EL84 grids (easiest place). Does it get more stable and polite or take off screaming?
Change the 0.1u caps to 0.02u.
That's still full gain at 50Hz but 5X less loop gain at 1Hz, which may be enough to dodge the instability.
Other possibility: "N"FB phase reversed. Swap the connections to the two EL84 grids (easiest place). Does it get more stable and polite or take off screaming?
Yep swapping the output transformer will have shifted the LF pole make an amp that was just stable into one which is just unstable. I don't know which way the pole has shifted by try the approach of cathode caps, and the coupling caps. Oddly making the coupling caps much bigger or much smaller can fix it. If you make the coupling caps much smaller as PRR suggests this becomes the dominate pole. If you make them much bigger then the transformer becomes the dominate pole. You want the NFB gain to drop to unity before the second pole kicks in. Without opening the loop and looking at a bode plot its difficult to give you a fix. I had exactly this problem in a valve amp. The solution was to modify the NFB with pole zero cancellation - but this is more complicated.
Last edited:
It's problematic to let the output transformer's primary inductance be the dominant pole (actually a zero, but yada yada) because its value varies drastically with signal level. Lots of good advice already, but I'd just emphasize the need to separate all other poles from the dominant. Cathode bypass caps should be very large by the standards of Golden Age amplifiers - 10 times larger at least. They're no longer an expensive item that needs cost restraints. Live large.
Folks will have different opinions about the location of the dominant pole, and I'll recommend that it be at the clipping point. For a properly designed amp this is at the output valve grids, where the pole is inversely related to clipping hang time.
YOS,
Chris
Folks will have different opinions about the location of the dominant pole, and I'll recommend that it be at the clipping point. For a properly designed amp this is at the output valve grids, where the pole is inversely related to clipping hang time.
YOS,
Chris
Try it without feedback. That'll tell you whether the problem was situated there, and you might just prefer the sound without.
I built something similar and that's where I landed (for now): 6v6 PP UL bias point
Edit: but I guess you'll have to add a resistor from the lower input triode to ground then.
Simon
I built something similar and that's where I landed (for now): 6v6 PP UL bias point
Edit: but I guess you'll have to add a resistor from the lower input triode to ground then.
Simon
Last edited:
Hi Guys, GOOD NEWS.
I've duobled the cathode cap at 200u but without result.
Then I brought the coupling caps from 0.1u at 0.32u (0.1u parallel to 0.22u) and BINGO!
The cone of the woofer it's finally practically motionless and the previus oscillation 0-200mV, now it's 0-60 mV. Thank you all.
The variation of this cap's does it affect the sound?
I've duobled the cathode cap at 200u but without result.
Then I brought the coupling caps from 0.1u at 0.32u (0.1u parallel to 0.22u) and BINGO!
The cone of the woofer it's finally practically motionless and the previus oscillation 0-200mV, now it's 0-60 mV. Thank you all.
The variation of this cap's does it affect the sound?
The original coupling caps value should not be a factor. Many of these amps have been built with no trouble. You have an instability problem that needs correcting through the general construction and feedback tuning for your specific output transformer. You still have the instability, it's just attenuated.
Last edited:
Super glad its sorted. You have moved it from just unstable to stable. You may still find there is a peak in the LF. Many valve amps suffer from this problem and go unnoticed. The change in transformer will have pushed it over.
I would go to .47u. Won't affect the sound very much I don't believe. The old coupling cap pole was 5.9Hz you have shifted it down to something like 2Hz. The Hammond transformers have a pole at around 5Hz hence the problem. I think the original transformer must have have a higher pole say 15Hz.
Last edited:
I believe the rolloff point for the .1uf cap is closer to 20hz and the .32uf takes it to 6hz. But that is going in the wrong direction if you want to filter out the LF. You don't want to let the LF through so you need to roll it off at 20hz., as originally designed.
Last edited: