diyAudio (
-   Group Buys (
-   -   The SLB (Smooth Like Butter) Active Rect/CRC/Cap Mx Class A Power Supply GB (

xrk971 5th May 2019 01:28 AM

I am using 32v transformer for the above test and DC was not too high. Under 2.15A load, the voltage above the cap Mx was about 41v.

xrk971 5th May 2019 03:19 AM

It's Like Buttah!!!
2 Attachment(s)
After some discussions we decided to remove the 10R resistor at R17 and R18 to buy back some gain that should squash the ripple even more. R17 and R18 are there to prevent oscillations should they occur, but do so at the detriment of ultimate ripple eating performance. So I removed the 10ohm resistors and replaced with a jumper. You could keep a 2R2 or even 1R5 there if you are chicken. But if you fry a $2 Toshiba 2SA1943N or 2SC5300N, no big deal. I monitored the output voltage with an o-scope during the power up and saw a small 50mV MHz range oscillation that was in a burst mode, but as soon as the voltage stabilized (5-10 seconds) the oscillation went away. What we are left with is pure magic. Here is 35.4v output into an 8ohm load for 4.43A current with a 3.1v drop across the CFP slave 2SA1943N transistor, and even at high current, the ripple is about 1mV. I measured 1.2mVrms on the postive rail and 0.9mVrms on the negative rail for the same condition. The negative rail had no startup oscillation.

This O-scope shot is on the 5mV per division vertical scale, and you can see the output. It's a tall order to ask a PSU with only 30,000uF bilk capacitance to give 1mVrms ripple under a 4.4A load. This means that one of these PSU's will have plenty of ripple eating ability for powering a stereo amp. The only thing one needs to be careful with are ground loops when sharing a PSU between two channels.

Here is the schematic of the production board (v2, May 3) with extra sets of output spade terminals for use as a stereo PSU:

In the above schematic, replace R17/18 with either jumper if you are brave or a 2R2 if not. Although a !R would probably squelch the startup oscillation nicely and give about same performance.

AKSA 5th May 2019 05:03 AM

Brilliant, X. It seems that with ripple at 1mVrms even a low PSRR Class A would sound utterly quiet. It occurs to me that using a single darlington like a BD679/680 in place of the BD139/140 we would be able to get the ripple down another ten times for a stellar result. But we chose BD139/140 for good dissipation, low cost and slow speed as we knew the CFP configuration is famous for instability.
Do you have any emitter resistors, ballast, on the slave transistor? We only use one device here (Pd is around 13.5W, only one device needed) and if it is stable, arguably ballast resistors need not be either reduced or even used.

Thank for letting us know what and exactly and how you gone about it. This is a primer for anyone studying linear power supply design.


xrk971 5th May 2019 05:11 AM

Hi Hugh,
There is no ballast resistor on the 2SA1943N slave BJT. It is exactly as it appears in this schematics (except jumpers provided at R17/18):

And even at 7mVrms, a low PSRR SE Class A amp does sound so quiet I cannot tell is is even on by listening to speakers when source is not playing music. 1mV must sound even more quiet.

AKSA 5th May 2019 06:07 AM

Very impressive, X!


nautibuoy 5th May 2019 06:37 AM

Hi X. Reference your post over on the MoFo thread;

will this group buy cover provsion of the single rail version?

If so I'll take a couple (ad be happy to combine shipping with the dual rail boards I've already ordered even if it means waiting).


xrk971 5th May 2019 11:24 AM

Hi NB,
Sure, we can start thinking about the single rail version of the SLB now that we know the layout is good. It’s basically the top half of the SLB board. We will have to wait for Jhofland to make the layout.

Let’s call it the SLB-SR, and simply specify SLB or SLB-SR in the GB list. Pricing will be $17ea for the SLB-SR. Note that cost of boards is mostly based on the single maximum dimension, and not the surface area.

Blackmani 5th May 2019 11:52 AM


Whenever I read about capacitor multiplier it sounds like a simple but an effective power supply solution for ripple rejection, mainly for class A amplifiers. But is there any disadvantage in using it instead of normal crc or clc power supply?

xrk971 5th May 2019 12:09 PM

If you are using a Class AB amp that may require a current draw beyond the output impedance of the cap multiplier (and beyond 5A continuous draw) that may not be desirable. The output impedance of the CFP slave transistor - in this case, is quite low and equal to the output stage of most amps using a similar BJT. For most Class AB amps in the 50w to 100w range, I don’t believe that the SLB is going to be a limiting bottle neck for the music program. Hugh may have more comments regarding use of SLB for other amps. Most well designed Class AB amps will have higher PSRR so may not benefit as much from having an ultra quiet PSU like a SE Class A amp would.

Blackmani 5th May 2019 12:21 PM

Great.. thanks for the quick explanation. Thanks for all the hard work you are doing for us all..

All times are GMT. The time now is 06:55 AM.

Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2020 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 17.65%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2020 DragonByte Technologies Ltd.
Copyright ©1999-2020 diyAudio