Need Modification Guidance

[mitchwolos]

I'm using a Quantasylum QA401 interface. The DSP is not in the signal chain of this measurement. This is directly out of the QA401 into the amp and back into the interface with a 4 ohm 300w resistor.

I haven't modified any of the stages. I just changed the BJTs because they were out of spec. The original Sankens are no longer available. I've increased the amount of feedback and eliminated the preamp board. If I

I replaced literally every part in this amp with top-quality components. To be clear. I don't think anything is necessarily wrong with the amplifier. I'm just confused as to why it performs so well without any bias and I'd like to see if I can fix the high-frequency distortion.

 

[NareshBrd]

It has peaks, so try checking the supply, and put a small pF in parallel with the main supply.
And those peaks have increasing distortion along with the peaks as the frequency goes higher.
Try and use a battery as power source, clean DC, see what happens.
SMPS can cause weird issues.
And if possible, vary the input to the main amp, this can be caused by the wrong signal level as well.

 

[mitchwolos]

It has peaks, so try checking the supply, and put a small pF in parallel with the main supply.
And those peaks have increasing distortion along with the peaks as the frequency goes higher.
Try and use a battery as power source, clean DC, see what happens.
SMPS can cause weird issues.
And if possible, vary the input to the main amp, this can be caused by the wrong signal level as well.

The peaks are multiband stimulus. The rail caps have bypass caps.

I can try an opamp on the input stage. But, It already has excessive gain. I did try a battery for power source.

I'm thinking it has to do with the phase of the feedback because the feedback goes from the outputs to the first stage and it appears to have 4 or 5 stages. So, Feedback is too slow .

 

[JMFahey]

Hi guys. I'm trying to modify an amplifier. I'm definitely a beginner. So, I started by increasing the bias to 10mv across the emitter resistor.

WHY????

But, It doesn't make any noticeable difference in the FFT measurement.

Of course.
You applied a non solution to a non problem, what did you expect?

I bypassed the entire preamp board.
I've tried lowering the value of R145 as well as R121.

WHY?

It does seem to have excessive gain.

Why do you think that?

It's a car audio amplifier and the DSP has up to 4v of output. So, I'd like to make use of it and have a low noise floor.

Are you qualified to custom program that DSP?
If not, your question makes no sense, nobody will program it for you.
Not for free at least.

I've also replaced the sanken 2sa1303/2sc3284 with MJL0302/MJL0281 because they were out of spec.

Ever heard of manufacturing tolerance?
I bet they were well within factory spec or they would never have passed Factory QC.

I matched the Beta of the devices within 1 or 2.

Not bad but irrelevant/undetectable.
Your own measurements say so.

I also replaced the rail caps with MKP Nichicon and all small value caps with WIMA and all through hole resistors with 1% or less low PPM value. All electrolytic caps with polymer. Emitters with Ohmite etc.

NO design equations include BRAND as a parameter, guess why.

I'm trying to reduce the high-frequency distortion. Attached a multiband measurement. It measures nearly exactly the same with 0mv across the emitter and 10mv and 15mv.

Proof your Mods have no useful effect, more specifically bias current, so why insist on that particular point?

Should I experiment with local feedback? I've been told that it probably has too much gain in the front stages.

Ask the Genius who told you so to help you.
Or he just talks but never grabs a soldering iron?

Dude. I already upgraded every part inside this amplifier with high-quality components.

"High quality" components do NOT improve designs because you are leaving DESIGN untouched, why would it change?
Again:
NO design equations include BRAND as a parameter, guess why.

It does seem to have excessive gain.

Why do you think that?
Did that Genius measure it?

It's a car audio amplifier and the DSP has up to 4v of output. So, I'd like to make use of it and have a low noise floor.

Are you qualified to custom program that DSP?
If not, your question makes no sense.
Who do you expect to program it for you?

I've also replaced the sanken 2sa1303/2sc3284 with MJL0302/MJL0281 because they were out of spec.

Ever heard of manufacturing tolerance?
I bet they were well within spec or they would never have passed Factory QC.

I matched the Beta of the devices within 1 or 2.

Not bad but irrelevant/undetectable.
Your own measurements say so.

I also replaced the rail caps with MKP Nichicon and all small value caps with WIMA and all through hole resistors with 1% or less low PPM value. All electrolytic caps with polymer. Emitters with Ohmite etc.

NO design equations include BRAND as a parameter, guess why.

Instead of trying to lecture me about wasting my time. Why don’t you help me with what I ask.

because you are following a dead end path and following it makes US waste time.

I'm asking for help. With THIS amplifier. It has to be THIS amplifier.

What part of not practical or possible you don´t understand?
Besides redesigning and fully rebuilding from ground up, but I doubt anybody will waste hundreds (or even dozens) hours doing your homework for you.

If you look at the provided multiband measurement. Distortion is near -110db in the midrange. 1k fundamental third harmonic is 0.009%. This measures really well compared to most other car amplifiers.

THEN LEAVE AS IS.

You can´t redesign it.
Some here are definitely qualified, ask quotations for Consultant/Design work.
Hope you have deep pockets.