My_Ref Fremen Edition - need help on PCB evaluation

[lce]

Hi,

I saw with great interest the shunt regulator used in the MyRef Fremen .
So I plan to retrofit a MyRefC board , replacing the 1KOhm resistor with a LM317 as current source , like in the Fremen , but keeping the original zener in place.

I read in this interesting article :
http://waltjung.org/PDFs/Sources_101_P2.pdf
that LM337 was really worse than LM317 for rejection in the audio band.
What do you think about using two LM317 : one as a current source in positive rail and one as current sink for negative rail ?

 

[lce]

Hum seems that someone got the same idea.....
I had an old version of Myref Fremen schematic :-(
I just saw the last version and two LM317 are used ....

 

[UnixMan]

Better than any regulator chip, for the CCS (both source and sink) you may use a couple of cascoded depletion-type MOSFETs...

 

[AndrewT]

did Jung say that?
I thought his results put the cascoded 317 nearly on a par with the depletion FETs.
fig13c &13d show and in his conclusion he recommended this topology as best bang for buck.

 

[ClaveFremen]

I saw with great interest the shunt regulator used in the MyRef Fremen .
So I plan to retrofit a MyRefC board , replacing the 1KOhm resistor with a LM317 as current source , like in the Fremen , but keeping the original zener in place.

You will be pleased with results (use OnSemi LM317s).

I suggest you to use KOA SPR Carbon films or Caddock MP915 for current setting resistors.

You can retrofit also the transistor shunt for much better results.

 

[clarkynoel]

Several month ago, I have asked permission with Dario about make new PCB design based on My_Ref Fremen schematic and Dario allow me to do that.

The main issue is about the small SMD components. for some people it's more comfortable to solder non SMD parts. The SMD, I think have several advantages and also disadvantages, but I will not to start debate about SMD and non SMD component. I just want to share the another PCB design using non SMD components

3 weeks ago, I have ordered the 2 boards for my own testing. Yesterday, I have finished to build 2 boards, for left and right channel. The sound is really really good, very natural, although I dont using premium components. I believe this amp can sound more details with premium components.

Many many Thanks to Mauro Penassa, Dario Inserra, Linuxguru, BMCBob, AndrewT, and many people that already contributing in My_Ref design

Anyone interested to evaluate this PCB design?

I will post about the differences with original My_Ref Fremen as soon as possible. Thanks

 

[clarkynoel]

Any comments, any another ideas, any suggestions are welcome for improvements.

The changes I have done :
- diode for limiter signal input to LM3886 (left is the original version, D1 and D2 using BAV99W, and right is D11, D12, D14, D15 using 1N4148.) both of 1N4148 and BAV99 are high-speed switching diodes
zener D3, the original using BZV55 zener 27V, replace with 1N4750 zener 27V

-LM318 the original using SOIC, I replace with PDIP-8 version
-another R and C with SMD, I change to non SMD

I also change value R15 (75K -> 22K), R16(75K-> 68K) and R20(47K->33K) in relay output speaker protection.

The original value needs about +-5.5V DC offset to make the relay switched (the voltage near collector Q1 can switch from 0 to 24V in xx ms), and the new value (already assembled in board, but I have not yet tried, but it works in my PC simulation, I will test it soon) can switch the relay when detecting 1.9V DC offset.

 

[linuxguru]

Bravo! I was planning a PCB respin of the Rev C boards with the regulator circuit from the FE for a while, which is among the things you've incorporated in your new board - good work.

The relay resistor network change is the kind of improvement that nobody has looked at in depth for a while, so that's also a welcome improvement if it works well - probably needs testing for a while to ensure that it doesn't false-trigger on loud low-bass sections.

I haven't looked in detail at the layout yet, but seems to be good work at first glance. Thanks for your contribution!

Edit: One thing that would worry me is the large Power Ground plane, which covers almost the entire solder side and extends below the small-signal ground plane at the input coupling cap and related small-signal components. Power ground is noisy with lots of ground bounce from the LM3886 surge currents, and can capacitively couple into the small-signal ground and pollute it.

 

[clarkynoel]

Hi Siva,
Thanks for reply. I'm happy to see your comment

The current problem with my board is buzz / noise when there is no jack connected to music source / player / computer. I heard sometimes a bit louder, especially at night. Don't know why. AC noise maybe?

But when there is jack connected to input of this board, the noise is disappear. (only jack cable connected to source and there is no signal / music played). when playing music, I can not hear the noise.

I use LM3886T not TF, and I don't think this noise cause by LM3886T. The PGND and HGND is unconnected to anything.

I still try to eliminate the noise. maybe if someone can join to evaluate this board could help me to identify which part is wrong

Edit :

One final essential note: every My_Ref I've ever built have hum with inputs disconnected, it's hum free only when inputs are connected to a pre and/or source.

is my board normal ?

 

[ClaveFremen]

Several month ago, I have asked permission with Dario about make new PCB design based on My_Ref Fremen schematic and Dario allow me to do that.

Just for clarity:

I would like to ask your permission, if I make my own board based your schematic, is it OK for you?

If you want to design your own PCB it's not a problem for me.

The My_Ref design is in the public domain for DIY use and my shunt for the LM318 is a well known one.

Simply, please, give credit to Mauro Penasa for the My_Ref and to me for the FE schematic/compensation and, please, don't call it Fremen Edition, my PCB is integral part of the design, you can say something like "based on My_Ref Fremen Edition schematic"

The main issue is about the small SMD components. for some people it's more comfortable to solder non SMD parts.

I don't think you can call the SMD use an issue!

It's actually a design choice, like yours of using TH parts.

Also that design choice is part of the FE performance...

Edit: One thing that would worry me is the large Power Ground plane, which covers almost the entire solder side and extends below the small-signal ground plane at the input coupling cap and related small-signal components. Power ground is noisy with lots of ground bounce from the LM3886 surge currents, and can capacitively couple into the small-signal ground and pollute it.

Absolutely, it's for this reaon I've removed the ground plane around the LM318... and it's not the only problem... almost all of the 'tricks' used on FE boards are not present and decoupling of both opamps is subpar.

IMHO the PCB needs to be, at least partially, redesigned.

 

[gadut]

Hi Siva,
Thanks for reply. I'm happy to see your comment

The current problem with my board is buzz / noise when there is no jack connected to music source / player / computer. I heard sometimes a bit louder, especially at night. Don't know why. AC noise maybe?

But when there is jack connected to input of this board, the noise is disappear. (only jack cable connected to source and there is no signal / music played). when playing music, I can not hear the noise.

I use LM3886T not TF, and I don't think this noise cause by LM3886T. The PGND and HGND is unconnected to anything.

I still try to eliminate the noise. maybe if someone can join to evaluate this board could help me to identify which part is wrong

Edit :


is my board normal ?

I also got the same noise issue as yours, T chip version. I suspect the relay circuit give the contribution since I don't find any noise when I disconnect all relay parts, you can try it too.