If you aim for low distortion, you have to include opt in the NFB loop. You do need to have a pair of good transformers to start with.
Also an amp should be stable even without speakers connected. If not, it is just a bad design.
The optimal proportion could be dynamic. If you exams dynaco schematics, primary is compensated with a cap to the cathode of the input tube, which dominates high frequency. Feedback resistor from secondary takes care of the low frequency stuffs. The transition point is about 100KHz. This approach can be applied to other amps. It is relatively easy to follow.
Last edited:
I guess the aim is always for low distortion, but on the IM (Intermodulation) distortion front, I'm not convinced that the global NFB from the secondary does the job.
Of course I could be wrong (I often am 🙂 ), perhaps my view (Thorsten's idea) is best described as a philosophy that flows, rather than a THD measurement. The philosophy being that I aim to drive the primary as best I can, and let the rest do what they want, no feedback bouncing around - just a flow of information at the primary, that via the secondary+voice coil finds it's way into the room, warts and all, unmolested, without fiddling with 'corrections' that may or may not be valid.
Sort of a 'libertarian delegation' to the secondary and loudspeaker, of 'Do your best with that!', rather than the micro-control we have with a secondary feedback loop, that in reality must always be a little late, and must play with voltages coming back from the cone flapping about at the other end 😀
My main surprise when I tried this (on a GU50 single ended amp) was how good it sounded. Effortless, open, powerful. Powerful as if a 100W SS amp was sitting there. I don't think I'm qualified to sell the idea here, but I can say I liked the result, and would suggest people try it! 🙂 .
So all I can suggest is that I found it very much worth trying, and of course it makes the phase lag of transformer - and hence that tweaking of 'how much stability / ringing' all go away nicely.
I put my ear closer to the output valve area and turned on my amp and I heard a little noise, like an oscillation, it doesn't come out of my speakers.
Is my amplifier oscillating, or are they the I E of the OPT?
Is my amplifier oscillating, or are they the I E of the OPT?
Some years had this happen and discovered it was the warming up tube heaters buzzing with AC heater current inside the cold output tubes before electrons are on the move. Two solutions; change tubes/types or use DC to stop the problem.
The output tranny laminations won´t buzz from amp cold start; more like mains tranny E&I laminations creating chassis noise. Solution, check clamping, dunk transformer in varnish or paint varnish on outside and hope.
BB
Apologies, I did not see this comment before.
I agree with you, 'fast' and 'slow' etc to describe a sound is nonsense 🙂
In this case however I was really talking of rise time: with my local feedback, around two tubes, the step response is very fast as the tubes can happily go to many MHz. The OPT secondary on the other hand, will be much slower to react due to the magnetism etc.
My view is the 'tidy' (with good phase margin perhaps?) GNFB around tubes plus OPT is less harmful that that of much solid state, because I consider tubes to be more linear in the first place (less harmonic multiplication occurs), but the OPT vs Stability balance will always result in a compromise....
...the question perhaps, is whether that compromise is worse that letting the OPT do it's own thing, I tend to favour letting the OPT do it's own thing now having tried it, and heard the result. It's also IMO a great way to avoid oscillation, at least at audio frequencies, haha! But as always there are many variables 😀
Once one audiophile-DIYer invited me to listen to his new creation, a preamp with zero feedback and transformer coupling. I brought mine to compare, with parallel feedback by voltage. His preamp produced audible intermodulation distortions, frequency response was audibly limited. He could not understand how I hear intermodulation. When we connected mine, everything came in place. He said that mine is too fast, but I could not understand what that means. He tried to explain to me that it plays too fast to follow the music, but I still could not understand that.
Synesthesia is very individual thing, it causes many incompatible languages, so language of physics and math is still the best to share experiences.
Synesthesia is very individual thing, it causes many incompatible languages, so language of physics and math is still the best to share experiences.
Just play a pure 1khz sine wave. You will hear distortion easily if there is any harmonic distortion. Bad amp sounds fuzzy under pure sine wave. Once you hear it, you will never unhear it. The only option is to get rid of the amp.
globulator,
Output transformer RF frequency output is complex. As frequencies increase, primary to secondry leakage inductance gets in the way. Then, depending on the winding layers, the primary to secondary capacitance couples all the way to UHF.
All of that can affect the performance of global negative feedback from the secondary windings.
For thinks like Schade negative feedback, RF effects are complex. The leakage inductance of the primary to secondary, and the capacitance of the primary to the secondary, and from the primary to the laminations affects the primary plate tap impedance versus frequency, all the way to UHF.
Again, that can affect the performance of Schade negative feedback.
Given the above factors, and the rest of the amplifier topology . . . If an Oscillation is possible, it will happen.
Output transformer RF frequency output is complex. As frequencies increase, primary to secondry leakage inductance gets in the way. Then, depending on the winding layers, the primary to secondary capacitance couples all the way to UHF.
All of that can affect the performance of global negative feedback from the secondary windings.
For thinks like Schade negative feedback, RF effects are complex. The leakage inductance of the primary to secondary, and the capacitance of the primary to the secondary, and from the primary to the laminations affects the primary plate tap impedance versus frequency, all the way to UHF.
Again, that can affect the performance of Schade negative feedback.
Given the above factors, and the rest of the amplifier topology . . . If an Oscillation is possible, it will happen.
I have added compensation to Schade feedback circuits to fix oscillation issues.
There are 2 kinds of stability issues. One is the high frequency stability. Actually most amps are quite stable at high frequency, as this is usually heavily tested and validated during the design stage.
The other one is the low frequency stability issue. It exhibits as motor-boating at low frequency. People talk about motorboating but only few people really know what is going on. The best resource that I can find is the wiki page.
In short, both capacitor coupling and transformer coupling would cause phase shift at low frequency. They can’t pass DC signal, so they roll off at low frequency with 90 degrees phase shift.
The total phase shift could reach 180 degrees easily. You only need one output transformer and one cap inside the global negative feedback to get 2x 90 degrees phase shift. When the phase shift reaches 180 degrees, if the total gain is still greater than one, you’ve got a low frequency phase-shift oscillator.
The other one is the low frequency stability issue. It exhibits as motor-boating at low frequency. People talk about motorboating but only few people really know what is going on. The best resource that I can find is the wiki page.
In short, both capacitor coupling and transformer coupling would cause phase shift at low frequency. They can’t pass DC signal, so they roll off at low frequency with 90 degrees phase shift.
The total phase shift could reach 180 degrees easily. You only need one output transformer and one cap inside the global negative feedback to get 2x 90 degrees phase shift. When the phase shift reaches 180 degrees, if the total gain is still greater than one, you’ve got a low frequency phase-shift oscillator.
Last edited:
That’s not two “kinds” of stability. They are both the same kind (Nyquist) and have the same fix (reduction in phase shift while loop gain >1). Local instabilities are different, because it does not depend on the global NFB. The fix for those must be local to the loop that is causing them. IV kink, requiring a screen stopper, for instance.
Even a power supply related motorboat is the same as the first “kind”. The feedback path is just not obvious, but it’s there, and it has 180 degrees of phase shift and loop gain >1.
Even a power supply related motorboat is the same as the first “kind”. The feedback path is just not obvious, but it’s there, and it has 180 degrees of phase shift and loop gain >1.
So true. There is however value in identifying the various loops involved as different types as they point to how an issue may be analyzed and resolved .
One of the main causes of motorboating in an amp is insufficient bypassing/filtering of the supply rail, signal gets propagated back to the input via the +V supply rail, gets amplified, etc, etc, etc, essentially, the un-bypassed +V supply rail turns the amp into an oscillator by acting as part of a positive feedback loop, remember, if I remember my electronics theory from my days at uni, regarding oscillators, there are at least two conditions that need to be met in order for an amp circuit to self-oscillate, the overall gain needs to be greater than 1, and, the total phase-shift needs to equal 360 degrees so that the output ends up in-phase with the input, here's something I remember reading somewhere "Amplifiers only oscillate, and oscillators only amplify".
Absolutely correct, with the exception of your last statement that should be detailled: The phase thing depends on how the amp works, i. e. 360° with a non inverting, 180° with an inverting amp.
Best regards!
Best regards!
It is part of the original definition of motorboating, but it is not the main cause based on my experience. When you trouble shoot any faulty amps, the power supply and the filter caps are the first thing to look anyway.
Bad power supply can cause all sorts of symptoms. Motorboating is just one of them. We cannot just say all sorts of issues are mainly caused by bad filter caps.
I want to say, after rule out the power supply issue, the next thing to investigate is the loop stability at low frequency.