Experience with this DIY DAC ?

I performed the op-amp mod (bypassing the second op-amp and mute circuit) and noticed a big jump in musicality, ease and presence. Very satisfying especially for such a simple mod.



Just one question though-

Has anyone attempted any power supply mods on one of these DAC's yet?

I was thinking that replacing the LM317 5V supply that feeds the DAC chip with some sort of super regulator might be beneficial with regards to dynamics & soundstage perhaps?
 
I performed the op-amp mod (bypassing the second op-amp and mute circuit) and noticed a big jump in musicality, ease and presence. Very satisfying especially for such a simple mod.



Just one question though-

Has anyone attempted any power supply mods on one of these DAC's yet?

I was thinking that replacing the LM317 5V supply that feeds the DAC chip with some sort of super regulator might be beneficial with regards to dynamics & soundstage perhaps?

I added JLH ripple eater boards to the 5V supply to the upsampler and the DAC and found a bit more more detail and spaciousness, which it had a lot of anyway. There wasn't a major boost to dynamics, those CS dacs have a smooth sound.
 
Good,

The attenuator will be a "lightspeed" type. I do have the much smaller HK DAC though.

Thanks
Peter

If it is a cs4398 board no problems except cooking up a balanced B1 circuit. I haven't seen one yet. cs4397 doesnt have enough output voltage without some gain.
I still question your motives somewhat, you will have two coupling caps and two semiconductors in each leg of the circuit, plus you will have to pick up the neutral off the board for your ground reference for the B1s. That is not minimalistic to me, compared to using one transducer (trafo) for the entire circuit.
 
Balanced B1 is available

If it is a cs4398 board no problems except cooking up a balanced B1 circuit. I haven't seen one yet. cs4397 doesnt have enough output voltage without some gain.
I still question your motives somewhat, you will have two coupling caps and two semiconductors in each leg of the circuit, plus you will have to pick up the neutral off the board for your ground reference for the B1s. That is not minimalistic to me, compared to using one transducer (trafo) for the entire circuit.

The chipset is the same! About the balanced lightspeed, the model i will use is called the PoorSerbianMan Lightspeed (Cooked by Zen Mod) from this forum. It will attenuate + take care of all the guess work out of the combination source <> amp.
I am a complete newbee so maybe I am going wrong ?

Cheers
Peter
 
Last edited:
The chipset is the same! About the balanced lightspeed, the model i will use is called the PoorSerbianMan Lightspeed (Cooked by Zen Mod) from this forum. It will attenuate + take care of all the guess work out of the combination source <> amp.
I am a complete newbee so maybe I am going wrong ?

Cheers
Peter

You're not going wrong, just in a different direction than I am going. One thing that you must deal with is the DC of 1.4V on the outputs of the cs4398. If you are feeding a pure differential circuit such as a trafo you can disregard it because it is common to both plus and minus outputs and no DC current will flow through the trafo. If you use a buffer on each output leg they all have to be referenced to ground, so that will put 1.4VDC across the inputs of all the buffers, and that will screw up your bias. If you could come up with a differentially balanced buffer circuit you could eliminate the caps at the inputs, but ZenMod's buffers will require them. Try to get his opinion on your idea.
 
Any chance you could please link me to these boards idlehands? I'm interested in improving the power circuitry of this DAC.

I got DIYed PCBs from a user on Rockgrotto headphone forum, they have a lot of info there with schematics and experience with the boards. That guy stopped making the boards and they commissioned some professionally made ones, the first run of which has sold out. Hopefully they will do more.

http://rockgrotto.proboards.com/index.cgi?board=m

What they do is take a DC supply and reduce the noise and impedence without a voltage drop, but this involves a capacitor multiplier that simulates a huge cap which needs a current limiter to deal with the power up current. 3 pin regs have a current limiter built in, so adding a JLH board after the reg but before the thing you wish to power is very simply way to get the best from low power devices. High voltage and high current versions can be made but sourcing fast transistors becomes a limiting factor.

Transistor speed matters for the power transistors, with Toshiba 2SC5171 and 2SA1930 worth the little extra over BD319 and BD140.
 
Yes, 3/4 and 9/10 are connected. Actually came that way. Chassis is plastic. I have the DAC board grounded to a screw in the chassis. I assume you mean the DPDT switch. It's a stand-in for a Grayhill rotary switch. I swapped the rotary for the dpdt in case that was the issue. However, I get the same results either way so yes, I believe it is wired correctly.

I'll check out the RCA isolation. The ground loop I'm sure I'll track down. The faint audio is where I'm stumped. I will clean the jumpers on the board and do some more investigating tonight. I'll also order some jumpers so I can pull of the daughterboard and see what happens if I skip the upsampling. As well as test the coaxial input.

thanks for your help.

ok so i lied 🙄 looks like a few days away from home, enjoying the california coast did me some good. after sitting down for just a few minutes with the DAC i realized that 3/4 were NOT connected. So simple.

Now everything is groovy😀:spin:😛

I built MTM towers, a tripath amp and now this DAC in 6 weeks. My mind just got too crowded. Just goes to show how important it is to take a step back sometimes, enjoy something else and clear the head...

thanks to everyone on this forum for info and patience.
 
I started with 270s on the primaries then tried 500 and 1k, I liked 500 the best, with a 500pf cap across the primaries. Secondaries straight to the output with a 1Kohm/.001uf RC filter across the jacks. Works for me, YMMV

Sry I didn't see this Bill. Thank you for your input 🙂. So many replies in this thread, but always nice that the thread is alive with interested people and opinions 🙂.

Which type of caps have you used.. something exotic or?
 
You can use Monacor LTR110 trannies. They are dirt cheap and they performs near as good as Sowters. I tested them and liked them very much, look there:

http://lh4.ggpht.com/_o-j1QGIOpHQ/Sk1eb0gzXhI/AAAAAAAAACo/IETxAcAcAec/s800/analog.jpg

Przetworniki cyfrowo-analogowe

I also tested "only one opamp thing" and I was not impressed. Synthetic, plastic sound, and the feeling of "being there" has completely gone. Not recommended.
You better try tube output stage like this:

Hi McGyver

I like your tube setup
would you please tell where I can get online set of pcb and ecc88 tube
If you have to choice between Lundahl, Sowter and Monacor what would you preffer
 
I finished wiring up the DAC and A-20 last night. Turn on and everything is good. I got no hum. Dead quiet. I don't put any filter/load to the pri/secodary of the A-20 and it sound really good specially for the bass.

I guess it's true what other builders said about A-20, bass is huge and tight. My CA DAC Magic with discrete opamp sound good with big bass but this DAC with A-20 is more refine and natural. Not much improvement in the hi freq though but it clears a bit.

I want to experiment some filter/load on the pri/seconday of the A-20 based on the Jensen and other suggestion previously. Can any body tell me which side is the most critical for the filter to impliment...the primary or the secondary.....anybody ???
 
Behringer SRC-2496 and Tamuras GA80080

Correct value for a 50khz low pass filter with 1K resistors is .0016 uF

Recently, I tried a Behringer dac (SRC2496) with a AKM AK4393 chip and i'm still amazed by the performance in upsampling mode.

Transformer mod is scheduled soon with this dac. I think that the CS4398 chinese dac will be something in the past soon.

Hi Guys,
First, let me say: fantastic thread!

Now, to the subject. (a little off-topic i guess)
Legarem, i also wanna try the Lampizator's suggestion. So I just ordered a Behringer src-2496 and a pair of Tamuras GA80080.
Do you recommend the same configuration as for the Tamuras and CS4398? 1k resistor and 0.0016 uF caps? If so, where can i find these caps. I can't find anywhere caps with these values.
I don't have the tools (square signal generator, oscilloscope) to test for the values, that's why i'm asking.

Thanks a lot!
 
Hi McGyver

I like your tube setup
would you please tell where I can get online set of pcb and ecc88 tube
If you have to choice between Lundahl, Sowter and Monacor what would you preffer

PCBs are custom made in Poland. You cannot at all get it online.
ECC88 or substitute you can buy on ebay or one of online shops.

I'm just working on the new analogue stage according to completely different conception, but still using trannies and tubes. I'll post it when I'll finish but now I have no time for it.
 
PCBs are custom made in Poland. You cannot at all get it online.
ECC88 or substitute you can buy on ebay or one of online shops.

I'm just working on the new analogue stage according to completely different conception, but still using trannies and tubes. I'll post it when I'll finish but now I have no time for it.

Hi every body! I'm watching this thread quite some time but this is my first post here 🙂 I have ordered the small dac version and plan to put it in pioneer player. I want to make my whole system go balanced so I'll try it with the monacor trannies and possibly tubes.

Two questions for McGyver:
On your website I have seen the photos of small dac striped spdif rca and wire connected to player board. Is that the way you connect the dac to player transport, via spdif? I can not connect dac via i2s since pionner uses another standard so I'm looking how to do it.
Second q is what do you think how would this tube buffer x 2 do in mentioned balanced setup: 12AX7B HIFI AUDIO TUBE PREAMPLIFIER KIT Board bei eBay.de: Amplifiers (endet 13.10.09 07:19:09 MESZ)