Its better but it not good.

With 2n2 it has improved, going to 4n4 is fractional better than 2n2.

Similarly a small series resistor from the output to the opamp decoupling cap (c3) would likely help too. Maybe 21-47R

Edited by user Sunday, March 23, 2014 10:47:00 PM(UTC)  | Reason: Not specified

Test will be carried out tomorrow (GMT+2, CET).

Avr, where are you ?

Ok , nice,,,and the test you promised us ?

Today the Beta kits arrived (finally). A job at the customs office over here must be extremely relaxing: if we don't process it this week, there's always next week or the week after that.

I will be putting together some mod kits to adapt the initial beta kits for the mods that have been worked out (details of which are to follow).

Sorry Leon... More waiting, unless you have a few extra parts handy...

Ok Guys, prepare for a long post.

AVR300 and I have been working hard testing and modding the beta kits to get the ultra low noise we were expecting.

What we found initially is that the error amp needed better isolation from the output to prevent some unwanted feed-forward, then we noticed that the error amp did not particularly like being directly connected to the VREF, and that the VREF itself needed some love.

This involved a few days of back and forth and measuring. :)

In the end we found that supply was very stable, but needed tweaking to really get the low noise we want.

So here it is - it requires some PCB surgery, and it is definitely easier to do these prior to stuffing the boards. You can do it after, but it will require some desoldering and careful work.

I will post pictures to aid in modding (forgive the state of the boards, they have been through a lot).

Here are the details:

film caps used are 220nf

Rs at vref are 470R (221R to 680R should be ok)

vref filter caps are 47uf(at least 16V) I used 25V.

Rs at Vout are 47R (22.1R to 75R should be ok)

Steps, cut traces from outputs to C3 close to the output.

Install 47R from C3 to output.

Remove VREF2 - snip pin 1;

Cut trace from VREF2 to VREF_TP

Install VREF2 as per the picture.

Install 47uf cap in place of VREF2 observing correct polarity as per picture.

Install 470R resistor as per picture.

Install 220nf across VREFs as per the picture.

Repeat for other half of the placid in case of placid BP.

The production boards will incorporate these changes.

Thanks so much to AVR300 who worked some late nights with me on this. You rock man!

I have also attached the last scope shot of the output - perfectly clean. :)

Edited by user Friday, March 28, 2014 10:42:47 PM(UTC)  | Reason: Not specified

Perhaps I should mention that it actually do works.

Man - it works.

In another threat I have been writhing that the Placid 1.0 (Q) modified was way better than the 1.0 original. It was / is.

But - this one is better than the 1.0 Q


Russ description is straight forward - as long as you already have done it ;-)

Feel free to claim even more pictures describing how to implement "beta-to-production".

It wasn't oscillation - it was noise. To me unacceptable noise.

Noise that was coming from the voltage reference and the error amplifier - or the combination of the two.

Both cured by the modification.


If your Placid isn't suffering from noise - then the modification perhaps isn't necessary for you to make. It is completely up to you.

You could try to scope the output to see for yourself.

Remember it's a beta run. Changes are expected.