Once again, this is the amp that did not oscillate.
Went back to the two 10K feedback resistors but left out the cap to ground.
It oscillates just turning it on.
I had removed the 220pF cap at the input to ground for square wave test and decided to connect it
again, what do you know it stopped the constant oscillation.
I remember Keantoken, talking about the input cap having an impact on stability, is anyone able to
explain the theory behind this? I assume that the feedback is coupled through any of the device
capacitances back to the input?
I improved the grounding - no significant change. Put the 10uF cap back to ground on the feedback path.
I lifted the input ground from the chassis and connected it through a 10R resistor to the output ground
on the circuit board - no difference. Perhaps I should have run a dedicated wire to the star ground.
I have a hunch that it is actually oscillating at the 1 - 5 MHz range when kicked into oscillation but I'm not
seeing it on the scope, rather I see it in the 15 to 150 KHz range.
In the process of running the sim, I tried many common methods for stabilizing the amp, looking
at the loop gain.
Miller cap didn't help much.
I won't go into all the details but a zobel (330pF + 100R) from the VAS collector to ground and a 330pF
across the VAS emitter resistor - SEEMS to make it stable. I got hints to try these from the Tigersaurus,
JE990, and the Bryston.
I tacked these into the amp and now it is reasonably stable, square wave response of the loop gain
has maybe 5% overshoot. When I touch the hot connection of the input it bursts into 30Vpp
oscillation at 15 KHz - strange could this be pickup of some local radiation? I set the scope very
sensitive and touch the probe, nothing like it. It just hit me that the UT seems to be like a pos feedback
regen receiver.
Something weird, the input jack is connected by twisted pair, if I squeeze the insulated twisted pair
the overshoot on the loop gain path with a square wave increases to about 30%. The input is strangely
sensitive.
I tried adding a 2.2K base stopper to the input diff transistor - no change.
At this point I wonder if there's local instability in the output stage, I should simulate the local feedback
path from output to driver. I could also try base stoppers on the output transistors and/or drivers.
Went back to the two 10K feedback resistors but left out the cap to ground.
It oscillates just turning it on.
I had removed the 220pF cap at the input to ground for square wave test and decided to connect it
again, what do you know it stopped the constant oscillation.
I remember Keantoken, talking about the input cap having an impact on stability, is anyone able to
explain the theory behind this? I assume that the feedback is coupled through any of the device
capacitances back to the input?
I improved the grounding - no significant change. Put the 10uF cap back to ground on the feedback path.
I lifted the input ground from the chassis and connected it through a 10R resistor to the output ground
on the circuit board - no difference. Perhaps I should have run a dedicated wire to the star ground.
I have a hunch that it is actually oscillating at the 1 - 5 MHz range when kicked into oscillation but I'm not
seeing it on the scope, rather I see it in the 15 to 150 KHz range.
In the process of running the sim, I tried many common methods for stabilizing the amp, looking
at the loop gain.
Miller cap didn't help much.
I won't go into all the details but a zobel (330pF + 100R) from the VAS collector to ground and a 330pF
across the VAS emitter resistor - SEEMS to make it stable. I got hints to try these from the Tigersaurus,
JE990, and the Bryston.
I tacked these into the amp and now it is reasonably stable, square wave response of the loop gain
has maybe 5% overshoot. When I touch the hot connection of the input it bursts into 30Vpp
oscillation at 15 KHz - strange could this be pickup of some local radiation? I set the scope very
sensitive and touch the probe, nothing like it. It just hit me that the UT seems to be like a pos feedback
regen receiver.
Something weird, the input jack is connected by twisted pair, if I squeeze the insulated twisted pair
the overshoot on the loop gain path with a square wave increases to about 30%. The input is strangely
sensitive.
I tried adding a 2.2K base stopper to the input diff transistor - no change.
At this point I wonder if there's local instability in the output stage, I should simulate the local feedback
path from output to driver. I could also try base stoppers on the output transistors and/or drivers.
Last edited:
The input RC filter attenuates incoming high frequency noise, so a marginally unstable amp like this one
is less likely to oscillate if there is not enough HF energy to trigger it.
is less likely to oscillate if there is not enough HF energy to trigger it.
Well, yes of course it is an input filter, but Keen talked about it causing a pole or altering one
from a loop gain perspective. I suppose I should ask him. @keantoken
Also, to add to my previous post, since the main negative feedback path is shunted from an AC
perspective it should have no way to oscillate and perhaps there's a "sneak" path through the
rails or ground. Or have I built an oscillator with the 10uF on the 10K resistors, and the shunt
on the negative input? This didn't show up in the simulation.
from a loop gain perspective. I suppose I should ask him. @keantoken
Also, to add to my previous post, since the main negative feedback path is shunted from an AC
perspective it should have no way to oscillate and perhaps there's a "sneak" path through the
rails or ground. Or have I built an oscillator with the 10uF on the 10K resistors, and the shunt
on the negative input? This didn't show up in the simulation.
This guy also finds KHz oscillation when he touches the input of a Tigersaurus, hmmm interesting:
And I thought that this one was stable:
He also thinks that it is in the output stage.
He also has one amp that works fine.
And I thought that this one was stable:
He also thinks that it is in the output stage.
He also has one amp that works fine.
I believe that the input filter more or less affects the loop gain phase response. My previous post.
https://www.diyaudio.com/community/...se-margin-of-an-amplifier.137436/post-7164930
Thanks! That's exactly what I was wondering about, the thing is that this amp has 220pF to ground
at the input. I will read that more carefully. I've seen amps (OP amp based) where something like
47pF was included across the differential bases, does this help in anyway? I'd think that it would
only make it worse. I'm going to read that entire thread.
Also, just last night I tried grounding the input and it calmed the amp right down, maybe I need to
increase that 220pF or choose very low capacitance diff pair transistors.
at the input. I will read that more carefully. I've seen amps (OP amp based) where something like
47pF was included across the differential bases, does this help in anyway? I'd think that it would
only make it worse. I'm going to read that entire thread.
Also, just last night I tried grounding the input and it calmed the amp right down, maybe I need to
increase that 220pF or choose very low capacitance diff pair transistors.
Bob Cordell wrote this in that thread post #6, for designs without a lead cap:
Not applicable to this amp since it has a lead cap.
"Let's say that your exiating gain crossover frequency is fc. Let's also assume that your feedback network has a flat frequency response (e.g., no lead capacitor, etc.) Change the feedback network to increase closed loop gain by a factor of ten. Measure the input-output phase of the amplifier at fc. This measurement must be performed absent any input low-pass filtering on the amplifier. This I/O phase measurement at fc will get you fairly cloase to the amount of open-loop phase lag at fc. If you measure 90 degree lag at fc, your phase margin is about 90 degrees. If you measure 135 degree lag at fc, your phase margin is about 45 degrees.
Another test is as follows: Add a pole in the feedback loop at twice fc and decrease closed loop gain by 1 dB. If the amplifier does not oscillate, you have at least 22 degrees of phase margin. This of course is not enough. Next, put a pole in the feedback loop at fc and decrease closed loop gain by 3 dB from its design value. If the amplifier does not oscillate, you have at least 45 degrees of phase margin. This whole approach is based on how much lagging phase shift an additional pole adds to the feedback loop and by how much that pole decreases the loop gain. By reducing the closed loop gain setting, we are compensating for the decrease in loop gain at fc caused by the introduction of the pole, leaving only its lagging phase effect."
Not applicable to this amp since it has a lead cap.
"Let's say that your exiating gain crossover frequency is fc. Let's also assume that your feedback network has a flat frequency response (e.g., no lead capacitor, etc.) Change the feedback network to increase closed loop gain by a factor of ten. Measure the input-output phase of the amplifier at fc. This measurement must be performed absent any input low-pass filtering on the amplifier. This I/O phase measurement at fc will get you fairly cloase to the amount of open-loop phase lag at fc. If you measure 90 degree lag at fc, your phase margin is about 90 degrees. If you measure 135 degree lag at fc, your phase margin is about 45 degrees.
Another test is as follows: Add a pole in the feedback loop at twice fc and decrease closed loop gain by 1 dB. If the amplifier does not oscillate, you have at least 22 degrees of phase margin. This of course is not enough. Next, put a pole in the feedback loop at fc and decrease closed loop gain by 3 dB from its design value. If the amplifier does not oscillate, you have at least 45 degrees of phase margin. This whole approach is based on how much lagging phase shift an additional pole adds to the feedback loop and by how much that pole decreases the loop gain. By reducing the closed loop gain setting, we are compensating for the decrease in loop gain at fc caused by the introduction of the pole, leaving only its lagging phase effect."
You may want to look at the last two pages of this document.
This was the method I used.
I measured the honey badger with it. Check out my build album for more details.
HB Build Album
I used a python to drive my function generation and scope to get the data for the plot.
Over all it worked well. However I don't think the phase margin was completely accurate possibly do to the bandwidth of the op-amp.
This was the method I used.
I measured the honey badger with it. Check out my build album for more details.
HB Build Album
I used a python to drive my function generation and scope to get the data for the plot.
Over all it worked well. However I don't think the phase margin was completely accurate possibly do to the bandwidth of the op-amp.
Thanks for that stuartmp I'll have to dig into it more.
jcx wrote this, and I thought that an amp would be close to minimum phase at least below 10MHz:
Again from this thread: https://www.diyaudio.com/community/threads/how-to-measure-phase-margin-of-an-amplifier.137436/page-4
"The amount of negative feedback you can apply is limited by delay, which is the non-minimum phase component of the system response, and by the margin that has to be allowed for the variability of both the non-minimum and the minimum phase components with load and operating point variation
Minimum phase response of the usually limiting output stage, for example single pole RC roll-offs, can in principle be equalized out (or compensated by tailoring the feedback transfer function) to the extent they are stable, repeatable
Distributed R-C lossey transmission lines and LC ladder networks formed from parasitic packaging and device R,L,C or geometric layout limitations can add limiting non-minimum phase (~= delay) even when fundamental device physics would allow more speed – power MOSFET polySi gate resistance may be an example
From my readings I still believe that high voltage audio power output BJT have minority carrier base transit delay that limits their phase response – although the fastest types may be approaching packaging and internal distributed RC limits as well
as for Class AB stages I believe "shoot though" increase in output stage current at high frequencies is well documented in inadequate designs"
jcx wrote this, and I thought that an amp would be close to minimum phase at least below 10MHz:
Again from this thread: https://www.diyaudio.com/community/threads/how-to-measure-phase-margin-of-an-amplifier.137436/page-4
"The amount of negative feedback you can apply is limited by delay, which is the non-minimum phase component of the system response, and by the margin that has to be allowed for the variability of both the non-minimum and the minimum phase components with load and operating point variation
Minimum phase response of the usually limiting output stage, for example single pole RC roll-offs, can in principle be equalized out (or compensated by tailoring the feedback transfer function) to the extent they are stable, repeatable
Distributed R-C lossey transmission lines and LC ladder networks formed from parasitic packaging and device R,L,C or geometric layout limitations can add limiting non-minimum phase (~= delay) even when fundamental device physics would allow more speed – power MOSFET polySi gate resistance may be an example
From my readings I still believe that high voltage audio power output BJT have minority carrier base transit delay that limits their phase response – although the fastest types may be approaching packaging and internal distributed RC limits as well
as for Class AB stages I believe "shoot though" increase in output stage current at high frequencies is well documented in inadequate designs"
Yes. I have also seen capacitors inserted between the inverting and non-inverting inputs in many particularly commercial amplifiers.
They are inserted as an EMI countermeasure. I have also experienced 60Hz demodulation buzzing under a strong electric field near a TV broadcast transmitting antenna (NTSC analogue broadcasting at that time), which was greatly reduced by inserting those capacitors. However, the phase characteristics almost always go in the undesirable direction. (However, in the case of R+C, it is often used as a compensation method to stabilise the unity loop gain frequency down without reducing the amount of feedback in the audio band.)
Equipment manufacturers have to comply with a number of regulations, including EMI, when developing commercial products. They must therefore market their products in compliance with them, even at the expense of general audio characteristics.
Last edited:
I have a theory that I've been meaning to mention. It's possible that touching the hot side of the
input jack provides a large 60Hz hum signal that drives the outputs into saturation, slowing them
down and kicking off the instability. Just a theory. Going to try to run sims of the output stage
stability and try a few things on the bench.
It is also interesting that hanging a foot long test lead off of the input hot sends it into constant
oscillation. Shorting the input seems to calm it down completely.
input jack provides a large 60Hz hum signal that drives the outputs into saturation, slowing them
down and kicking off the instability. Just a theory. Going to try to run sims of the output stage
stability and try a few things on the bench.
It is also interesting that hanging a foot long test lead off of the input hot sends it into constant
oscillation. Shorting the input seems to calm it down completely.
Well, the problem is that we have to make the design stable even in overload so
the first thing is to solve the stability problem. If it had hum, then yes a shield
would help and also perhaps the capacitance of the shielded cable might help.
This has me thinking more, I mentioned in a past post that just squeezing the
twisted pair wires for the input caused the square wave overshoot on the loop
gain to go from 5 to 25% so obviously that change is making a big difference
in the loop gain. I also need to sort this out. I'm sure that there are multiple
problems with this design. What is very interesting is that the input sensitivity
is seen in this amp, the larger and different Tigersaurus, and the Tiger .01 with
a triple output stage. They all do it. I've seen these amps with a note on top,
DO NOT PLUG IN INPUT WHILE ON, lol.
the first thing is to solve the stability problem. If it had hum, then yes a shield
would help and also perhaps the capacitance of the shielded cable might help.
This has me thinking more, I mentioned in a past post that just squeezing the
twisted pair wires for the input caused the square wave overshoot on the loop
gain to go from 5 to 25% so obviously that change is making a big difference
in the loop gain. I also need to sort this out. I'm sure that there are multiple
problems with this design. What is very interesting is that the input sensitivity
is seen in this amp, the larger and different Tigersaurus, and the Tiger .01 with
a triple output stage. They all do it. I've seen these amps with a note on top,
DO NOT PLUG IN INPUT WHILE ON, lol.
The BE capacitance of the input transistor forms a filter with the source impedance at it's base. So the input voltage drops by some factor of the differential voltage. The effect is similar to that of degeneration. If say the source capacitance at the base were equal to the capacitance between differential inputs, then the input voltage would drop by an amount equal to the differential voltage since both capacitances are conducting the same current. This assumes the source capacitance dominates at the test frequency of course.
Thanks Kean, mason_f8 also explained it here: https://www.diyaudio.com/community/...in-of-an-amplifier.137436/page-7#post-7164930
I wonder if the diff pair transistor(s) gets damaged with age, this amp has new old stock 1970s MPS6566 there perhaps I
should try some better, modern production parts.
Also, I realize that this diff pair should have degeneration for the obvious reasons, but I wonder if that resistance
helps to decouple the BE capacitance between the pair. My only reason for not putting them in was to keep the mods
to a minimum.
I wonder if the diff pair transistor(s) gets damaged with age, this amp has new old stock 1970s MPS6566 there perhaps I
should try some better, modern production parts.
Also, I realize that this diff pair should have degeneration for the obvious reasons, but I wonder if that resistance
helps to decouple the BE capacitance between the pair. My only reason for not putting them in was to keep the mods
to a minimum.
I’m not trying to advocate adding resistors, but only suggest a possible technique: sever board trace and span the gap with a surface-mount part.
Best.
Best.
I actually have an old MOT data book but it doesn't have Ft data. It does have capacitances Cob is 3.5pF
and Cib of 4.3 pF.
and Cib of 4.3 pF.