Building a symmetrical PSU B1 buffer

ra7

Member
Joined 2009
Paid Member
Guys,

The leds are very of good tolerance... I can almost pick any 5 from the lot and they measure about the same.. although, I don't think this is what Salas is refering to. Also, the jfets used in the CCS portion are suspect types. They have GR marked on them, but they measure 8-10ma Idss. I discussed this a few pages back with Salas. Anyway, I took a chance and it is working... although, haven't tested the sound yet. I have a pretty transparent system, so any oddities should show up.

Will post pics of my led mishap soon.
 

ra7

Member
Joined 2009
Paid Member
led orientation pain

Yes, even after reading some things on orienting the leds properly, I managed to orient all of them in the exact opposite direction. Here are some pics.

I fixed it by chopping off the tops and re-soldering new ones on the top side of the pcb, although this was not my initial plan. At first, I tried to desolder a couple of the leads of left in the holes, but the pads themselves started lifting. This can be seen in pic 1. So, I used the top side instead. Pic 2, shows the solution to pads that were lifted..... just solder the successive leds' legs to each other.

I hope this helps some other unfortunate soul who might make such a mistake. More pics when final build is done.

(the power led is yet to be fixed :D)
 

Attachments

  • IMG_5851[1].JPG
    IMG_5851[1].JPG
    793 KB · Views: 540
  • IMG_5853[1].JPG
    IMG_5853[1].JPG
    812.2 KB · Views: 514

ra7

Member
Joined 2009
Paid Member
Salas,

For the input selector, the switching is happening simply by switching to ground? I'm not seeing any other connections to be made. Also read your comments in the GB thread.

So, just connect 1-6 for the inputs on the switch and then connect ground to switch output?
 
One way to get closer than norm, is to have a weaker idss JFET in the plus side. Another way is if the particular Leds are stiff and don't translate easy to significantly more Vf for 5-10mA running current differences. Those Jfets do work both ways but the unorthodox way is normally stronger for self current and worse for regulation. In hot rod CCS currents environment the subjective benefit in THD profile is filled out by the stronger tone and details. In Blue the orthodox will be standard since its reinforced everywhere for hotrod. But unorthodox it helps tone with small currents. Helps with offset too if the audio fets are not that well matched also. Did you measure your offsets BTW?

My gut feeling is that it is likely stiff LED's because the Jets are idss matched between the 2 rails. The jets in the B1 are all at 9ma, and the shunt's are around 7ma. For the time being, I am using 2 wirewound 50R pots to dial the CCS to around 125ma at each rail. The offsets are -1mV and -2mV respectively. No sound yet as I haven't wired the pot and input RCA's. Hopefully, I will be able to hear the nice sound this evening.

Thanks for your support!!
 
Worry is over, it has been singing for 2 hours with my workbench gears. It replaces a passive preamp by Scott Nixon that I bought a few years ago. The first impression was very positive, the added dynamic and the revealing of the detail were quite noticeable.
I will let it run for the rest of the night to break in.

Thanks Papa, Salas, and TeaBag to make this possible!!
 

ra7

Member
Joined 2009
Paid Member
U got me there man! The phono stage was built on a breadboard and underwent several changes (as u and I know) but then I moved and had to dismantle the board. Since then, I have got my hands on the dl103 and am working on the breadboard again. This time starting with your simplistic NJFET design - the 37V version. I'll post there once I finish it. You know I got the LEDs to work in that design on the breadboard but still messed it up in the B1 :eek:

Anyway, back to the B1. Its up and running. I haven't finished all the inputs yet and the output is sitting outside the chassis. I'll post pics as soon as I'm done. I listened for a while this evening and it sounds wonderful. Dark, is the first thing that comes to mind. Everything is dark and silent in the background. Also, the bass is so powerful and has plenty of slam and I was quite surprised by this as I already have plenty of powerful bass. Is the shunt supply playing its part here? And the vocals are simply beautiful. All in all, superb sound. I didn't think introducing another component would have such a positive effect, but it did and I couldn't be happier.
 
...and then I found AndrewT's post #524

select Idss to be similar.
The B1 is designed to run the active device at Idss. That requires the CCS device to pass the same or similar Idss.
If you have a pair of 8.15mA and 8.23mA then that is a selected pair. they are within 1% Idss. You could use <2% tolerance Idss. I suspect you could even use <5% tolerance Idss, but I do not know which should be the higher and which should be lower. There are science reasons for NOT running the active device at well above it's Idss but some builders are reporting that they prefer the sound when the active device is running at >100% of Idss. I will ALWAYS run <=100% Idss. I will NOT try >100% Idss.

But, that selection MUST BE DONE at the SAME temperature. This must be checked when both jFETs are supplied and measured at the same time and they are thermally coupled together.
 
which helps me find an answer to these questions:

Please confirm:

The matched jfets are Q5 & Q6 in the schematics of post 874.

From the builder's guide:

"Select the pair to have similar Idss and the lower run then biases the upper one."

Which one is lower? Q5 or Q6? I can install devices with identical Idss. Is it better for one of them to be lower? ...or the same?

...a little more help and we'll be having a beer.
 
Yes, those 4 jFETs are the amplifiers.
Yes, the middle pair are the CCS and the outer pair are the source followers.
If you have slightly lower Idss and want to use them in the CCS position then the middle pair are where you locate the lower Idss jFETs.
Remember to keep the selected pairs together, either the left pair or the right pair.

If you look at the PCB traces you will see the G & S connected together indicating the CCS duty.
Print off the PCB layout, big enough to fill an A4 sheet. Makes it easy to see where everything goes.
 
Last edited:
hello,

Salas, I have a question for you or anybody who could help me. It is about the delay start circuit of the DCB1. If I put an smaller capacitor (I actually have a 100uF one) can I make the relay to mute the circuit sooner? I mean when you turn it off.

I need the relay kick in it a bit earlier for a Lightspeed attenuator experiment I am doing with the buffer.

Thank you very much for your help, guys. I really appreciate it.

Regards,
Regi