I use the 317-337 regulators to convert the 50-60Hz to DC and set an initial DC Voltage. Then, I put the shunt regulator after a Caddock 10 ohm power resistor that is in series with the first regulator, and this resistor keeps the first regulator seeing a constant resistance and nothing complex, and something that the active shunt regulator can work against. Both regulators are in the power supply box. Then, in the main box, each channel has its own built in cap multipliers ending with a .1uF Rel RT polystyrene.
Last edited:
The case is that currents of output stage halves sum to almost constant. Upper and lower device currents do vary (that was your point probably), but sum is constant. And currents of both supply rails are constant, each.
Yes I see. I didn't really look at the circuit since they are often without values and such (that's why I asked about the load R), but only to the curve and misread the scale. Ahh well, another learning experience for free
jan didden
See John that's what I don't understand. After that shunt you have an almost ideal supply with close to zero ripple, an output Z that's close to zero, and then you add this cap multiplier which undoes the low output Z again, and doesn't do anything for the already non-existing ripple etc. What's the point of the cap multiplier?
It would be much better to have the shunt close to the load, in the amp box, to keep any varying (admittedly low) supply return current local, to avoid any ground contamination in the umbilliocal between supply box and amp box.
jan didden
The cap multiplier is open loop local feedback, where as shunt still operates on loop-NFB.
Guess which one is better for coping with HF ?
Patrick
The shunt. The NFB always reduces the output Z to below what any cap multiplier could do, way out to above audio frequencies. The cap multiplier has always the intrinsic emitter resistance from the pass device, which in itself rises with freq meaning it appears inductive.
If a shunt has higher Zout than a cap multiplier, it's badly designed to begin with.
Do this thought experiment: take a cap multiplier with a certain Zout vs freq. Then wrap NFB around it, and you'll see the Zout drop. Dramatically.
jan didden
That's not generally the case. Why would NFB loops be slower? It may result from an often repeated misconception (not by you) that feedback comes 'after the fact' and therefore always comes too late.
This has been shown to not be the case over and over again but especially people with no engineering background appear not to be able to grasp the concept.
Fact is that if you use too much feedback you may make your circuit overshoot and ring. This results from the fact that when frequency rises, the delay through the amp (NOT through the fb network because in general that's just two resistors) increases. That means that the phase shift between the input signal and the feedback signal increases. If you go so high in freq that the phaseshift approaches 180 degrees, your nfb turns into pfb. If that happens, AND you still have gain in your circuit, you have an oscillator.
The remaining phaseshift before 180 degrees, at the point where your gain drops down to one (so it can't oscillate), is called the phase margin. Now, even with some phase margin (less than 180 degrees phase shift and still gain > 1) you start to see ringing on fast signals.
But, it is very straight forward engineering to design your circuit to any stability point you want. Sometimes engineers accept some ringing at say supersonic frequencies because a) there will never be any signal in that region except on the test bench and b) it allows just a bit more fb in the lower freqs to make the circuit a bit more linear or transparent or lower Zout, whatever the circuit is.
jan didden
An open loop shunt is equivalent to a load resistor (to set DC shunt current) plus a resistor (the shunt dynamic Z) in series with a capacitor (to isolate the DC). It's not a regulator at all. 'Regulation' implies feedback to regulate an output parameter.
jan didden
I know John, nobody doubts that, least not me.
I can agree to that difference, I have no strong opinion whether local feedback is essentially the same as global fb or not.
I assume that with 'local feedback' in case of your shunt, you mean some sort of (emitter) follower configuration with the (emitter) follower load providing the local feedback or, as it used to be called (avoiding the confusion), degeneration?
Still I have difficulty to accept such a circuit being called a regulator. Remember how the regulator term came into use? Steam engines were initially 'open loop' circuits with the output (rpm, torque) set by how much coal you shoveled in the firehole.
Then someone invented the 'governor': a mechanical type of regulator that measured the output (RPM) and if it rose, throttled down the supply of steam, and vice versa. A true regulator, 'regulating' a circuit parameter to keep the output at a setpoint.
This mechanism is absent in an open loop (emitter) follower type circuit.
jan didden
Last edited:
Sorry for the delay, been away from the thread for a bit, but Kovar is a much better match to most glasses than invar. Invar will lead to cracking because it expands LESS than the glass, most other metals expand more than the glass, plus it wets to glass quite well.
Wrinkle
It appears to me that even 1cm of wire, from the AC inlet to the power transformer, if not properly isolated and/or filtered, may well radiate RF that resides on the mains supply to the entire circuit. Or am I wrong?
A new audio measurement?
Not to knock this study in emi noise reduction off track (I have opinions but this may be more interesting) here is a really novel amp measurement from the upcoming AES:
P7-5 New Techniques for Evaluating Audio Amplifiers via Measuring for Induced Wow and Flutter and Differential Phase Distortions—Ron Quan, Ron Quan Designs - Cupertino, CA, USA
In the past, mechanical systems were measured for Wow and Flutter or frequency modulation but not amplifiers. Instead, amplifiers are typically measured for intermodulation and harmonic distortion. A new method for audio amplifier/device performance measures frequency modulation effects and differential phase distortion. Frequency and phase detectors are used to evaluate induced frequency and phase modulation from an amplifier under two conditions. The first condition has a low frequency signal inducing the modulation on a high frequency signal. The second condition has a high frequency AM signal inducing the modulation on a lower frequency signal. Practical design topologies for the new test methods are shown and the results of the new testing methods are tabulated.
Convention Paper 8194
Not to knock this study in emi noise reduction off track (I have opinions but this may be more interesting) here is a really novel amp measurement from the upcoming AES:
P7-5 New Techniques for Evaluating Audio Amplifiers via Measuring for Induced Wow and Flutter and Differential Phase Distortions—Ron Quan, Ron Quan Designs - Cupertino, CA, USA
In the past, mechanical systems were measured for Wow and Flutter or frequency modulation but not amplifiers. Instead, amplifiers are typically measured for intermodulation and harmonic distortion. A new method for audio amplifier/device performance measures frequency modulation effects and differential phase distortion. Frequency and phase detectors are used to evaluate induced frequency and phase modulation from an amplifier under two conditions. The first condition has a low frequency signal inducing the modulation on a high frequency signal. The second condition has a high frequency AM signal inducing the modulation on a lower frequency signal. Practical design topologies for the new test methods are shown and the results of the new testing methods are tabulated.
Convention Paper 8194
I will make sure I attend that presentation, it sounds too intriguing. I just hope it is more than AM/PM in a new coat
.jan didden
EMI RFI isolation
I see a lot of poor unserstanding of EMI management in the audio field and a fair amount of wishful thinking about reducing its impact on audio. The first common mistake is thinking that the third wire on a power connection will do anything constructive for removing noise. its more likely to introduce more noise than remove any, but it serves a very important safety function. Second, thinking that a classic EMI filter on the power inlet will help. Mostly they couple the ground leakage currents into the audio grounding making it all more difficult.
For anything above 1 MHz I don't think active components will help. Effective LC networks are really the only choice. You must take into account the parasitics of your components, inductors that have effective bypass caps and caps with effective series inductors. Active components will only amplify these effects if they can amplify at those frequencies. John's trick of the transformer with isolated windings helps a lot. I did that on a preamp with a transformer with a huge core. the coils were 2" apart on separate legs. It made a large difference that a shield can't.
I am also a fan of cascaded regulators and shunt regulators simply because they force you to figure out the loop currents. There is another paper being given at AES on essentially this issue:
P13-1 Neutral-Point Oscillation Control Based on a New Audio Space Vector Modulation (A-SVM) for DCI-NPC Power Amplifiers—Vicent Sala, Luis Romeral, G. Ruiz, UPC-Universitat Politecnica de Catalunya - Terrassa, Spain
In this paper the oscillation or flotation in the DC-BUS neutral point in the DCI-NPC (Diode Clamped Inverter – Neutral Point Clamped) amplifiers is presented as one of the most important distortion sources. This perturbation is characterized and studied, as well as its causes and distorting effects. It also presents two techniques of vector modulation for audio. The intelligent use of these techniques in the process of vector modulation allows the redistribution of the charge of the two capacitors in the DC-BUS, allowing the control of the voltage in the neutral point of the DC-BUS, and therefore, the cancellation of the flotation and its distorting effects. Experimental and simulation results that verify these strategies are presented.
Convention Paper 8227
I see a lot of poor unserstanding of EMI management in the audio field and a fair amount of wishful thinking about reducing its impact on audio. The first common mistake is thinking that the third wire on a power connection will do anything constructive for removing noise. its more likely to introduce more noise than remove any, but it serves a very important safety function. Second, thinking that a classic EMI filter on the power inlet will help. Mostly they couple the ground leakage currents into the audio grounding making it all more difficult.
For anything above 1 MHz I don't think active components will help. Effective LC networks are really the only choice. You must take into account the parasitics of your components, inductors that have effective bypass caps and caps with effective series inductors. Active components will only amplify these effects if they can amplify at those frequencies. John's trick of the transformer with isolated windings helps a lot. I did that on a preamp with a transformer with a huge core. the coils were 2" apart on separate legs. It made a large difference that a shield can't.
I am also a fan of cascaded regulators and shunt regulators simply because they force you to figure out the loop currents. There is another paper being given at AES on essentially this issue:
P13-1 Neutral-Point Oscillation Control Based on a New Audio Space Vector Modulation (A-SVM) for DCI-NPC Power Amplifiers—Vicent Sala, Luis Romeral, G. Ruiz, UPC-Universitat Politecnica de Catalunya - Terrassa, Spain
In this paper the oscillation or flotation in the DC-BUS neutral point in the DCI-NPC (Diode Clamped Inverter – Neutral Point Clamped) amplifiers is presented as one of the most important distortion sources. This perturbation is characterized and studied, as well as its causes and distorting effects. It also presents two techniques of vector modulation for audio. The intelligent use of these techniques in the process of vector modulation allows the redistribution of the charge of the two capacitors in the DC-BUS, allowing the control of the voltage in the neutral point of the DC-BUS, and therefore, the cancellation of the flotation and its distorting effects. Experimental and simulation results that verify these strategies are presented.
Convention Paper 8227
Ron Quan is a friend of mine, and once worked for me, as he finished his degree in Engineering at UCB. I spoke to him at length, about this paper, last night. He is very aware of Cordell's work. This is ALSO the person I mentioned on this website about one year ago, and then I could not get any more, because people were demanding more info from me about his pending patent and his method. Finally, we will see what he has found.
Thank you John
Slightly OT:
http://www.diyaudio.com/forums/club...festival-audio-happening-207.html#post2350136
jan didden
Slightly OT:
http://www.diyaudio.com/forums/club...festival-audio-happening-207.html#post2350136
jan didden
Hi John,
I'll look forward to hearing about this work as well. Can you tell us what two frequencies he is using? It does sound a lot like the PIM measurement that Matti concocted and for which I developed the coherent IM analyzer that uses synchronous detection techniques.
The three things I would first look to hear are:
1. Is the measurement technique itself basically the same; i.e., something like Otala's where 60 and 7000 are applied in some ratio and phase modulation on the HF signal is measured?
2. Is the instrumentation largely the same; i.e., use of synchronous phase detection?
3. Are the results similar for reasonable amplifier variations being tested? I.e., is PIM small when AIM is small? Is PIM ever found to be large when AIM is very small (this would go against previous findings).
From your discussions with Ron Quan, do you know the answers to #1 and #2?
BTW, will you be at AES? If so, I'll look forward to seeing you.
Cheers,
Bob
- Status
- Not open for further replies.