Forum breadcrumbs - You are here:CommunityGeneral discussionsAYA 2 list mods
You must be logged in to post to the forum.

AYA 2 list mods

PreviousPage 4 of 5Next

I’ve fitted the Grundig DEM mod , 100nf  caps used are just small yellow box polyester .

definitely a noticeable difference in sound , width is a little narrower but instruments better separated, more space around them . Bass seems warmer, I’ve only listened to a couple of tracks . It seems s little more tiring to listen to .

I think I’m going to give up with the 1200uf fc and go back to 1500uf fm . I’ll then see how it sounds with the DEM mod still fitted .

btw I mainly using usb input with onboard pcm2706 . Music is from Amazon prime HD , I’ve set output to 16bit / 44.1


Yes, my experience with sync DEM clocking was probably similar. It turned into a trade-off between the accuracy at one, and scale and a bit of musicality on the other side. Anyhow, it remains a useful means to tweak the TDA1541A sound.

PCM2706 is a previous generation of adaptive mode USB receivers, but in its days it was the best among such parts. So it can not achieve the ultimate resolution, nor does it support UAC2, but it is generally sonically neutral – not fully natural as well, but it was generally a decent part.


Quote from kingbusoms on 26 January 2024, 9:28 PM

I remember one of my TDA1541A chips (one Philips stamped ) was noisy had hissing noise . I presume that was the DEM faulty ?   Once faulty I suppose the chip is beyond repair ?   Luckily I have a S2 and a Taiwan stamped spare

Yes, the non-operational TDA1541A DEM circuit can be heard as a noise, but this noise occurs only in the presence of a signal.

And yes, once it fails, you can not do anything about it. But don't be that afraid of TDA1541A chip reliability. Just don't leave it on permanently, and let it sleep every night, and it will be thankful and probably work well for many years.

I tried to find some graphs showing the performance of some faulty TDA1541A chips, as I thought I saved some in some folder(s), but I had no success for now. Anyhow, I have some faulty chips, and I can use them to take new graphs, as time permits.

But they (were and) will be similar to what I got when I mistakenly missed to properly drive the DEM. So, the chip shown here was correct, but its DEM did not work (pin 17 was left floating).

Thanks Pedja and thank you for the graphs ! After more listening I prefer the sound without clocking , as you say its useful to try these tweaks especially cheap and easy ones .

I'd like to try a better USB to i2s , I presume I could try one of those Amanero boards or Jlsounds



Hi there, Eldam on board ,

I readed this fastly, here to re read Pedja's notes about clock speeds articles but I readed that thread too as I always have my AYA tunned to death like Jaffrie budy. (Btw, Hi Pedja and Jaffrie)


@kingbusoms , the precisions I gave are importants in my system for the result I shared. But indeed the final result is a set up in relation to what you have around, so no final better tweaking list but just tips to try.


The FC cap I strongly suggest is not the 35V but the 50V large diameter short length (there are 3 ref so size for Pan. FC 1200/50V). Is it important ? YES it is. The // cap was a FM cap iirc with a 5 mm pitch to make the whole a little snappy, but the // cap can be your fav cap too to correct tonal according your own system)

A gift more : I do not use the 3 x decoupling 100 uF caps after the regs and I just decouple the -15V pins with 0.2 uF acrylic (same than PPS, Pan is making now an acrylic close to the PPS, same same but different at ears, I migth be autobiased !lol).

From there, tune the main smoothing cap of the -/+ 5 V rails : the BOM or else, your taste, your system.

A little gift for the AYA2/4 owners (I unluckilly missed the AYA5 wagon and clock board because of life) that are lazy and not as obsessive than I or Jaffrie :

Here is a setup I like very much, but you will have to find some decoupling tests elswhere : for the two analog reservoir caps rail, please choose exactly that ref : Pan FM 2200/16  before they are NLA... definitly here a different animal in my system. It is important here to notice I use direct coupling at the output (no blocking caps).

for the output resistor after the last opa buffer go figure why : choose the same ref than the AYA bom but in 1/8 W ratting. Give some finesse in my system, certainly due to the very low thickness of the leads ! It is not advised, but I just put it in lieu w/o soldering in the vias. Then I use after a simple copper silver plain cable to go till the end of the board; cheap, good.

I also use a Ayrton Perry wirewound from Rohpoint for the I/V task and keeped the AB elswhere. The rohpoint tweak known elswhere by the TDA1541A enthusiasts but it is not as critical. I also use polystyren tin foil // to it, better than the usual styrens I found. But all of this is in the AYA 2 thread if I remember or at least in this site in the discussions. I find also the Vishay AC0 green wirewound to be good, but do know it is not Ayrton P type so is inductive . There are a non inductive version I didn't find to purchase alas... The Rohpoint is hard to find and also in the needed value (but up to you to pump up stronger your preamp).

Happy tweaking, hope it works for you too 🙂

Eldam (Damien from France)

PS : Grunding mode was if I am correct with Fs 171 K hz and Bit clock 5.6xx M hz due to the digital filter SAA7720A before the TDA1541A present in all the CD players back then.

Not sure Grunding mode is effective (for the sound)  in NOS mode i.e. 44.1 K hz. The TDA divides by 4 after the internal oscillator; it may be too slow and and also the 14 DEM decoupling caps needs different values. Not sure, I am not confortable with those relations between the I2S pins speed datas and the Dynamic Element Matching things. But my understanding is the 470 pF free running Oscp DEM cap gives exactly this 171 K hz speed to the oscillator and the 14 x 0.1 uF caps if I understood are their value chosen accordingly so.

If you use Grunding mode, perhaps it is better to upsample x4 your redbook library from the computer before to stream towards the AYA.

I found the simultaneous mode a far better improvment than playing with I2S and Osc values.


(FS is 176.4 K hz, not 171 K hz, sorry for the error)

Aha, nice to see you here Eldam. Lol. Think we 2 are are craziest here. Like to share our findings but somehow it's difficult to many to swallow. Think the key issue is AyA builders aren't as compulsive & curious as us. Oh still wrapping my head round this simultaneous thingy ya,  me using I2S direct from my SdTrans 384 .



Thanks Eldam , its great to keep the modding and discussions alive !

I will order a 1200uf 50v FC and change the 35v version .

I have 0.22uf smd PPS on bottom of board for -15v and +/-5v   .

I have tried 0.47uf BG NX HiQ and 22uf sepc for C302  and no cap .

I don't have anything fitted to C305/308 , I did try 1uf Oscon but currently removed

I have 2.2uf Silmic  for C401,402,403,404

I use Paul Hynes Shunt regs for the +5v , -15v and OPA861  all psu caps for +/-5v  and OPA861 are 1500uf FM 16v

I removed the Grundig Dem mod and refitted fitted 470pf evox PFR. I also have a 680pf I was going to try


I keep looking at the JLsounds usb to i2s , it also allows to connect to TDA1541A simultaneous mode . Disconnect link for BCK and SYS pin on TDA1541A.  Run LR_CLK to LE, Data L to DataL  ,BCLK to BCK ,Data R to Data R. Add additional 4k7 resistor pin1 to 3 of Jlsounds board  .Pin 27 of TDA1541A connected to -5v

PreviousPage 4 of 5Next
Scroll to Top