Project: Spice Simulation of Power-Amps with real loads

[AnalogSteph]

What does CCS topology or VAS topology mean? I never heard of those acronyms.

You have gotten this far without ever reading Douglas Self? Heresy.
Distortion In Power Amplifiers
The guy has literally written one of the two or three standard books on power amplifier design, next to Bob Cordell.

 

[tvrgeek]

@tvrgeek:

Yes, I am aware, that a loudspeaker does not behave as it ideally should do due to non-linearities of quite a few sources...
Actually I have a list of >350 papers (not all are available to me at the moment).
In fact I would be highly interested in the papers, that you mentioned; be it as a "author/title" input to my list or even as full paper pdf.
Give me your input. Thx in advance.


I will add your potential candidates to the list of TBDs. Thanks.
What does CCS topology or VAS topology mean? I never heard of those acronyms.

Constant Current Source and Voltage Amplification Stage ( stage two in a standard amp) I did a huge analysis of CCS topology in Spice. Bottom line, cascode FET is the best of the best limited to two devices, but the big step is a single BJT and green LED which is what I used as "good enough, go fix the next problem" Matching the transistors in the current mirror was fruitful. Again, you can play with the models in Spice to see what small miss-matching will do to you. Hint, why modern OP-Amps have such good performance. Hint, look at ALL the curves for the different LEDs. You will see why green.

I came across details on the non-linearities of the motor inductance when researching horns and transmission line loading. I think it was Martin King or others. I don't care for horns for home use and consider transmission lines obsolete, so I didn't bother to mark them.

Read Self, Cordell, and Leach for basic understanding on amplifiers. Jan Didden gets really technical in his publication. I have not had the time, but for a small fee, I can get access to the NC State library where they have IEEE paper access. Don't know about AES. I bet if you lived in Berkley, they have everything as many of the pioneers in solid state attended. I could be a hermit in their engineering library if given a chance.

I'll give you the same hints I was given: Look very carefully at the method of compensation in the VAS. Look how it behaves across frequency including above the audible range. Look very carefully at the curves of FETs and BJTs. Look at the quiescent current in all the stages. Only hints as I am not qualified to give a lesson on it.

Here is a little page with some good details. :
http://aetechron.com/Controlled-current_vs_controlled-voltage_operation.shtml

I am not sure I know of any current production transconductance amps. Maybe that is a hint. Maybe the no-load risk is too much for a commercial product.

 

[sepp2gl]

You have gotten this far without ever reading Douglas Self? Heresy.
Distortion In Power Amplifiers
The guy has literally written one of the two or three standard books on power amplifier design, next to Bob Cordell.

Yes...shame on me.

 

[sepp2gl]

Constant Current Source and Voltage Amplification Stage (stage two in a standard amp)

Thx 4 explaining CCS and VAS. I just wasn't aware about the acronyms.

Hint, why modern OP-Amps have such good performance.

Would you advise to use integrated Op-Amps or CFAs in the input stage rather than discrete components?

I came across details on the non-linearities of the motor inductance when researching horns and transmission line loading. I think it was Martin King or others. I don't care for horns for home use and consider transmission lines obsolete, so I didn't bother to mark them.

I could not agree more. I used to build TLs with KEF-Drivers for quite some time, but today sealed enclosures can benefit more from progress in amplifier technology.

Read Self, Cordell, and Leach for basic understanding on amplifiers. Jan Didden gets really technical in his publication.

To Do's: Need to buy Cordell. Need to read Self (~700 pages). Look for Jan Didden.
Already done: Read Leach (quite some years ago)

I'll give you the same hints I was given: Look very carefully at the method of compensation in the VAS. Look how it behaves across frequency including above the audible range. Look very carefully at the curves of FETs and BJTs. Look at the quiescent current in all the stages.

Thx. I'll buy that.

I am not sure I know of any current production transconductance amps. Maybe that is a hint. Maybe the no-load risk is too much for a commercial product.

Transconductance amps need to have either a feedforward compensation of the load impedance to work in a reasonable way. As this is highly dependent on the loudspeaker, it is not the piece of cake of an amp-manufacturer. Another solution is a voltage feedback as Abacus Rieder Ampino provides. We might doubt, if current-drive is still effective then. Another trap for current-drive might be a passive cross-over in the loudspeaker: if current has to be maintained i.e. for the bass-driver the resulting fast voltage increase might end up in the mid-range speaker.
Not sure about it. So I expect to use current-drive in active speakers exclusivly.
This would also fix the open-output issue with current-sources as there will be no disconnect. Otherwise open-output could be detected and handled by electronics as well as short-circuit for voultage-drive amps.

 

[sepp2gl]

@tvrgeek:
P.S.: Jan Didden is an Elektor-Guy.
Elektor is a DIY-Elektronics Magazine in Europe since the 1960's.
As I am a frequent reader, I am well aware about their publications.
Some made it into my list of candidates.

 

[sepp2gl]

You have gotten this far without ever reading Douglas Self? Heresy.

Yes...I agree: how could I grew up without it...
I am just reading Self's 6th edition and I am really impressed by the systematic way, that he approaches all aspects of amp-design.
A very useful almanach.

 

[sepp2gl]

Hi All!
Quite some weeks have passed by. Without any update to you - but not without any progress.
This is what happened till now (besides pandemic restrictions):
1. I established a complete open-loop test for power-amps as automated ngspice-script.
Open-loop in this respect means without NFB, but can include local feedback.
This test especially looks at built-in linearity, slew-rate, bandwidth and THD,
that comes with the amp-concept without any correction by NFB.

2. Based on the open-loop script I am now working on closed-loop test-script.
The open-loop tests were performed with 4 and 8 Ohm resistive loads only.
With closed-loop a simulated loudspeaker is added.
The resistive load characteristics are used to compare with available measured data.

With only a real bass-driver load the simulation is limited to <<10kHz. Above the simulation breaks due to convergence issues caused by instability.
So I need to extend the single bass model into either a 2/3-way crossover loudspeaker or a single wide-band loudspeaker model.
Both solutions make perfect sense to me.
As the proposal from this board have proven to be very helpful to extend my portfolio of power-amps,
I would like to ask the members of the board to give me some proposals for both multi-way and wide-band loudspeakers.
Otherwise I will pick from my own database.

If there is any additional closed-loop characteristic, that should be added to the list, please feel free to propose.

Would you be willing to continue supporting this project?
Thanks in advance
sepp2gl