A NOS 192/24 DAC with the PCM1794 (and WaveIO USB input)

Rax · 2014-09-01 8:56 pm

carlsor said:
Internally, a current control diode is a JFET with the gate and source tied together. It should sound the same as an adjustable single JFET circuit. The current control diode seems difficult to implement with precisely 400ua current. It comes in 100ua, 300ua, and 500ua values varying as much as plus or minus 40%. Good luck finding singles or combinations that get close to 400ua. A lot of testing throw-aways at over $2usd each.

the same difficulty stands, though, for JFET based CCSs... Variability in build makes them hard to preconfigure. A trim pot may be the trick.

The only option that would be good out of the box is a BJT-based one (at least two proposed this far), which can be set pretty precisely with specified parts values...

nige2000 · 2014-09-01 9:31 pm

The two 300uA ones I've tested were close enough match
That could be just pot luck though

dwjames · 2014-09-01 10:18 pm

nige2000 said:
The two 300uA ones I've tested were close enough match
That could be just pot luck though

Well my 2 x e-301 measure 365uA and 369uA when tested with a 3.3v power supply.
The 2 x e-101 are 105uA and 155uA.
Hmmmmm.....
So the closest match available is just to use the e-301. I guess 365ish uA is close enough to 400uA to get away with for now right?

dwjames · 2014-09-01 11:26 pm

ok, so I just bodged them in quickly for now. Dc offset on the outputs has dropped to 2.2v.
Only heard a couple of tracks so far, but definitely like what I'm hearing. Precise and detailed, but airy, spacious, full and lush in way that sounds like it comes real easy.
Definitely worthwhile, just need to work out that way to hit that magic 400uA figure in a way that's easily done and easily reproduced..

nige2000 · 2014-09-01 11:37 pm

You using Rhopoint 120r?
Your bias could have been lower than 2.7v previous?

Think these diodes could be an option too if there matched and rload is adjusted
Wonder what higher ccs and lower output rload sounds like

dwjames · 2014-09-01 11:44 pm

nige2000 said:
You using Rhopoint 120r?
Your bias could have been lower than 2.7v previous?

Think these diodes could be an option too if there matched and rload is adjusted
Wonder what higher ccs and lower output rload sounds like

Good call, yes. Rhopoint GG102D 120R 0.1% so it will have been a bit low before.
I forgot to measure the dc offset with the 6k resistors in place. It's been a long day...

nige2000 · 2014-09-01 11:53 pm

Think 120r was 2.5v ish at the
Stock 6k r

No matter lowering the bias should have had a neg effect on sq
But that didn't happen

Chanh · 2014-09-02 6:22 am

Here it is guys! With many late nights..., 9Months son, bread&butter, and along with the unfriendly criticisms (crazy, idiots.., has more money than brain cells....). It was somewhat deflated. Nonetheless, I am now managed to complete this master piece of Doede's Engineering in association of Tent Labs years of experience in audio! Hopefully it sounds better then its look!

Rax · 2014-09-02 6:55 am

Zen Mod said:
instead of Jfet , it's possible to use this http://www.ti.com/general/docs/lit/getliterature.tsp?genericPartNumber=lm134&fileType=pdf

...maybe option #1 (post #2443) deserves a closer look at this point, having in mind the thought put in the other options and the pluses and minuses found this far.

Rejection is very good across the audio band, and as it works well from 1V it should deal easily with the conditions at hand.

dwjames · 2014-09-02 7:00 am

Wow Chan! That's quite some feat of engineering. You must feel like you work in a DAC factory

But soon you will get to hear it and I'm sure that will be a happy day indeed

dwjames · 2014-09-02 7:14 am

Ps, it's not 'crazy', it's an obsessive mental condition, but that's fine, we all have it here. You're in good company

stijn001 · 2014-09-02 8:06 am

Rax said:
...maybe option #1 (post #2443) deserves a closer look at this point, having in mind the thought put in the other options and the pluses and minuses found this far.

Rejection is very good across the audio band, and as it works well from 1V it should deal easily with the conditions at hand.

I've been glaring over the spec sheet too, looks very promising. Comes in various useful packages, It's cheap and readily available. I'm trying to calculate the RSet value. For 0.40mA at 2.4V would required the Rset to be 166Ohm, would you agree?

stijn001 · 2014-09-02 8:11 am

Chanh said:
Here it is guys! With many late nights

Looks "god all mighty!" :worship:

Don’t let the “skippers on land” (= Dutch proverb) and naysayers discourage you…

dddac · 2014-09-02 9:35 am

Hi Chanh,

nice piece of work!! look forward to your listening tests. I like to use this on my website as a special project, so it remains more visible. could you get me some more pictures to select from and eventually a nice report? thanks !

Chanh · 2014-09-02 9:43 am

Hi Doede,

All photos captured so far were from my iPhone. I will capture them proper with my D800 if the intend is using on your website?

Btw, thanks James and Stijn for the encouragement! Much appreciated!

I & Tuyen will be doing the critical listening and will report back. If opportunity permits, we will also comparing the stock single DAC vs mod with Tent shunt regs.

Watch this page!

Not to forget, I would like to thank Doede and All have contributed into this amazing DAC!

Best,
Chanh

dddac · 2014-09-02 9:47 am

great Chanh, indeed, some D800 pictures would be good. we can share by Dropbox the RAW format? I can work them in Lightroom than. we can do normal email for the details. I will set up a shared folder for us, as these files are large...

we will make sure all names and contributors will be honored

Rax · 2014-09-02 11:52 am

stijn001 said:
I've been glaring over the spec sheet too, looks very promising. Comes in various useful packages, It's cheap and readily available. I'm trying to calculate the RSet value. For 0.40mA at 2.4V would required the Rset to be 166Ohm, would you agree?

I think 165 ohms should work...

It's temperature-dependent, but the spec sheet mentions a way to cancel that. May be worth considering and I think the voltage drop allows it.

stijn001 · 2014-09-02 12:22 pm

Rax said:
I think 165 ohms should work...

It's temperature-dependent, but the spec sheet mentions a way to cancel that. May be worth considering and I think the voltage drop allows it.

Thanks,

I saw the temperature dependency, they write head sinking will reduce the thermal effect by a factor of 3. The TO-92 package is also probably most temperature stable.
Keeping the part count down to two would obviously be nice. I'm intrested in trying this first.

I've just ordered a bunch of them.

dwjames · 2014-09-02 12:54 pm

stijn001 said:
I've just ordered a bunch of them.

which exact part number did you order?
I see:
LM334Z/NOBP - TO92
LM334Z/NOPB - TEXAS INSTRUMENTS - IC, CURRENT SOURCE, ADJ, 40V, | Farnell UK
LM334M/NOBP - SOIC 8
LM334M - TEXAS INSTRUMENTS - IC, CURRENT SOURCE, 40V, 0.01A | Farnell UK
LM134H/NOPB - Metal Can
LM134H/NOPB - TEXAS INSTRUMENTS - CURRENT REGULATOR, 40V, METALCAN | Farnell UK

Rax · 2014-09-02 2:26 pm

stijn001 said:
Thanks,

I saw the temperature dependency, they write head sinking will reduce the thermal effect by a factor of 3. The TO-92 package is also probably most temperature stable.
Keeping the part count down to two would obviously be nice. I'm intrested in trying this first.

I've just ordered a bunch of them.

sounds good to me. you guys have been running the built kits - so you know better - but I assume there's not much opportunity for heat near the board so this should be pretty solid. unless anyone runs this with shunt supplies which may get some heaters in there...

A NOS 192/24 DAC with the PCM1794 (and WaveIO USB input)

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member