The Holy Grail Follower Output Stage

Jump to Latest
#21
The ZD25-Mini (aka. the Holy Grail Follower).
The ZD25 is an amplifier I built that is similar to the Pass Labs XA25, that was discussed in the "F4 Beast Builders" thread: https://www.diyaudio.com/community/threads/f4-beast-builders.300233/post-4907053.

The ZD25-Mini is a simplified, lower power implementation of the ZD25 output stage.

Implementation:
The amplifier combines three experiments (for me):
  • The optocoupler bias circuit described in post #1.
  • Using a opa551 high voltage opamp for the front end, thanks to EUVLs “The M2 Output Stage in Class A/B, and maybe a Power WHAMMY?” thread
  • Use of inexpensive SMPS for the power supplies. A regular linear supply would be fine, but cost more and be heavier. More about this later.
After many simulations to verify the circuit topology, I designed a PCB containing the output stage, the optocoupler bias circuit, and the OPA551 opamp based front end. After receiving PCBs and parts, I assembled, bench tested, adjusted a few part values for the amplifier. Recently, I finally began listening tests and I am very pleased with the result.

This is a FirstWatt scale amplifier with a PCB that conforms to the diyAudio UMS and can operate with either linear or switching power supplies with +/-23V to +/-27V rails. The output stage FETs are not degenerated which has the benefits described in the Nelson Pass article "Art of the Square Law": https://firstwatt.com/pdf/art_the_square_law.pdf.

Specifications: (as configured Jan 16, 2024)
  • Voltage Gain: 22dB
  • Global Feedback: 12dB
  • Distortion: THD <0.01% at 1W, THD < .05% at 25W
  • Input Impedance: 46K
  • Damping Factor: 120
  • Frequency Response: 1Hz to -3dB @270 kHz.
  • Power Dissipation: 60W-70W per channel
  • Noise: below 100uV
Here is the circuit schematic as of Jan 16, 2024:

1705433285800.png

1705433307386.png


1705433320883.png
 
  • Like
Reactions: decramer, dantwomey, chermann and 9 others
#30
I tried IRFP140/9140 and IRFP240/9240. Both exhibited the IRF P-channel problems that Nelson has described in the past.

I happened to have the Fairchild FQA28N15 and FQA36P15 FETs in stock so I tried them and they look very good for this output stage.

I have on order some IXYS IXTQ36N30P and IXTQ36P15P FETs that I will also try.

I really want to use FETs that are in current production.

Do you have other suggestions?
 
  • Like
Reactions: DIYHarry
#33
The opa445 would be good if I needed rail voltages over +/-30V. Otherwise it has a lower gain bandwidth product.
I preferred the IXTQ36N30P partially based on its lower Crss. I don't know if the higher transconductance of the IXTQ75N10P would be better or not.
 
#34
What I have posted so far is work in progress. There are many component options parameter values that can be tuned.

What I found amazing is how the amplifier sounds and measures. I keep rechecking my instrumentation because the measurements seem unbelievable.

In future posts, there are three major subtopics to discuss:
  • Performance of the magic-modified optocoupler bias circuit. Hint: It works as predicted in simulations.
  • Performance of the opamp front-end.
  • Performance of the SMPS supplies. Hint: No filters (except to RC filter at opamp) and I cannot detect SMPS noise.
 
  • Like
Reactions: Murdoc
#39
Yes, I the SMPS noise is very low without serious filtering.
  • Output stage has good PSRR
  • RC filters at the opamp
  • Careful star grounding
  • opamp common-mode rejection by RCA shield ground-loop breaker resistor.
Note that the spectra shown in post #22 show a spike at around 42kHz. It goes away when if I turn off the overhead fluorescent lights. I do not know how that couples into my measurement system.
 
Last edited:
  • Like
Reactions: DIYHarry, dkfan9 and codyt
You must log in or register to reply here.
Top Bottom