DIY Audio Projects Forum

Emitter and collector resistorless output stage, you mean like Q5 and Q6 in this:


This is not a good topology, I never seen people using this since the early 70s.

This and common emitter output stage is not good. Output impedance is too high. You can only lower the output impedance by GNFB, that is a really bad way to do that. You need as low output impedance without feedback, never depend on closed loop feedback to do the job. Like any closed loop feed back, loop gain is low at 20KHz, no loop gain at frequency of oscillation. Amp like this cannot drive capacitive load as output impedance and load capacitance form another pole.Output impedance of my design is 0.03ohm without feedback,

Its deffintly not the output stage you have shown in picture. That has emitter resistors.

I mentioned before where the cross over like distorion is happening at high frequencies you cannot hear it, also speakers don,'t output any of those kininks due to limited frequency response.

The distortion your seeing in 24khz,124 khz sine waves are due to nature of the resistorless oputput stage.
The last picture is 124khz square wave and that is not cross over in the part you circled. That is due one gain stage running without any compensation and purely reling on pole canceling and have small residuals.
This indicates high gain and is a good sign. Without frequency reduction the square wave has no overshoot(90degree phase margin) but still has that small kink during rise of square.

Listening exprience has proved that my 4 stage op amp. Not only sounding better than previous 3 stage op amp.
My latest valve only op amp with low openloop distortion and high gain 2300+ has dominating distoriton. (using 6 vaccum tubes and requires compensation)
It has zero visible distortion on scope, however the distorition is much higher and dominates the distortion when used with 4 stage head amp.
The distortion was much lower when I used a hybrid valve pre amp and higher openloop gain impression of pure zero distoriton music.

As I saw in your scheamtic your using mosfets as relay.
Did it offer you auditible sound quality improvements or reduced distortion on testing. What brang you there.
IMO mechnical relays are good enough long enough life span and using higher amperage ones will reduce the contact resistance so its totally neglitible

I never try using mechanical relay on the pcb. I can tell you I have huge relays for a amplifier mux so I can have two amplifiers switching one by one to drive the same pair of speakers. The idea is to compare two amplifiers with identical setup using the same source and speakers. It did NOT pan out, the sound obviously degraded. Those are big 40A relays already.

I can tell you the THD did not get worst before and after the MOSFET relays. Those are very expensive transistors, like $5 each. Very low Rdson. I use 3 pairs in parallel. I design for the best.

You really need to get a spectrum analyzer, I don't think you can claim anything looking at those waveforms. They don't look good to me. I'll take a set of picture when I have the new boards up.

No circuit, not much to say.

You need to do large signal, you are missing the point here doing 2V signal.

I don't like idea of having music pass through mosfet DS before going into speakers, Perosnally would prefer mechnical relays.

It's a lot lower resistance than contact resistance of relays. The resistance is 2mohm or less. I have been using this from day one, all the measurements are done through the SS relay. I got 0.002%THD large signal. You can't do better than this.

Some good op amp designs allow you to have zero power on off pops. My head amp falls into this category.
This means you don't need a relay but if any fault and dc appears on the oputput then the speakers are smoked.

the pop is the last thing I worry. SS amp are DC coupled output, anything goes wrong you pump large current into the speaker. Burning the speaker is the least of my concern, it can catch on fire and burn the place down. One time I accidentally shorted the output, it burned my screw driver. It's like arc welder. I have rail fuse on the pcb, It didn't have time to burn the fuse before it burned the screw driver with a big spark.

On the side note, even with this protection, I still pull AC coupling caps to block the DC to the woofers. Like 20 of the 3300uF caps each side. Too much money burning the speakers.

I would have dumped in a mechinical relay for my head amps however I don't have enough space thats why I didn't