Experience with this DIY DAC ?

You could connect them on the DAC board with the ground where the 2 caps meet (104) like on this photo.

181363d1279897837-experience-diy-dac-daughter-pcb.jpg

Hi Ryssen,
can you post a picture of the back of this board?

Thanks
Andrea
 
Tried to get the adapters to work but had mixed results.

With Ryssen's help I managed to get sound from the first adapter but only in one channel. Fearing i may have damaged the chip I tried to remove it but wrecked the pads on the PCB. Removing 28 pin smd chips is a nightmare... However the sound from the one channel was very good!

The second adapter, which originally i had populated with op-amps and decoupling components, produced no sound even with components removed. So may also have a damaged chip.

I have continuity from the chip pins to outputs and continuity to ground.

I have spare PCM1798 chips and will order some more boards to try. I built the boards like this:

IMG_2058.jpg


IMG_2057.jpg
 
Are you sure there was contakt on all the soldering points,the ground and so on?

Hi Ryssen

As it happens a couple of contacts seemed questionable. I will try to reflow the joints later.


I received this interesting email from Bill Whitlock of Jensen Transformers regarding suitable transformers for current output DAC's:

Hello Rich,

Thanks for writing. I get lots of inquiries about transformer outputs for
various DAC chips - and I know this idea is very popular among audiophiles -
but I have never seen anyone ever publish performance data (linearity,
noise, etc.) after such a modification. The chip manufacturer never intended
such use, otherwise specifications like "output voltage compliance" would be
on the data sheet. I've inquired of application engineers about this
"uncharted territory" use of these DACs and received the following typical
response from Bob Gonzalez of Texas Instruments in January 2010:

"The PCM1794 (and its replacement, the PCM1794A) was designed to feed into
the summing junction of an opamp powered with dual supplies and having its
positive input at common as shown in figure 24 of the PCM1794A datasheet.
Whereas the PCM may operate with a non-zero compliance voltage it has not
been characterized for that configuration and therefore no compliance spec
appears in the datasheet. I hope this helps. If you have further questions
please email or call me at 972-644-5580."

For example, for the DACs you mention, full-scale output current is 7.8 mA
pk-pk. Therefore, a 10 Ω load on this output would produce only 78 mV pk-pk
or 27.6 mV rms full-scale signal. Since the impedance ratio of any
transformer is the square of its turns (voltage) ratio, a turns ratio of 1:4
would make a 10 Ω load at the DAC output look like over 160 Ω at the
transformer output and still raise the full-scale output voltage to only
61.7 mV rms. The only way to get more output voltage is to raise the load
impedance seen by the DAC. For example, a load of 507.5 Ω at the DAC output
would produce 3.96 V pk-pk or 1.4 V rms at full-scale (still a bit short of
the industry norm of about 2 V rms). If the DAC chip could tolerate this
kind of output voltage swing (i.e., "compliance"), no transformer step-up
would be required at all. Then a simple 1:1 transformer could be used to
convert the differential output to single-ended, if desired. But, since the
DAC manufacturer doesn't specify "output voltage compliance" at all, I
believe there's a high risk that it would work badly, if at all.

I hope you will understand that I'd rather not have one of our transformers
put at risk of being called "the problem" when such applications work badly.
We have a 35-year reputation as makers of the world's highest performance
audio transformers, but even the finest transformers can produce
disappointing results when used improperly. In any case, we do not make a
suitable 1:4 step-up transformer for this application (we make a 4:1
step-down transformer, but it is a gross oversimplification to believe that
all one has to do is interchange primary and secondary to make an equally
high-performance step-up transformer).
 
Did you download the chapter on audio transformers from the Jensen website. It will help quite a bit.

A perfect analogy are the step up trafos used with MC carts.

He mentions a 10 ohm load for the 1794, and that is probably a good ballpark figure for a working circuit.

Don't get confused by the trafo impedance ratings. A 600 ohm line trafo does not have a 600 ohm unloaded impedance, it is more like 100k ohms. The 600 ohm rating is the max load it will drive and still meet specs.

If you put 600 across the secondaries, the primary will also be 600. If you put 5k across the secondaries, the primary will also be 5k.

The impedance reflected from winding to winding will always be the square of the turns ratio X the load.

For instance, if you use 10 ohms across the primary of a 1/10 trafo the secondary will be 1k. The voltage will be multiplied by the turns ratio so the output voltage would be .780v P-P. If you could double the 10 ohms to 20 ohms then you would have 1.56v P-P.

The problem is, the dacs werent designed to generate voltage. They want to work into a dead short. That is the problem, what kind of load can you get away with. The dac designers are no help so everyone is on their own, but it is easy to see why you would want a 1/18-20 trafo ratio so you can keep the load as close to 0 ohms as you can.
 
The OEPs aren't any better than the Monacors but they will sound different. Neither of them give you all the specs you need to compare them to quality trafos, and I'm sure it's for a reason.

It's not surprising the opa627s sound better than the Monacors. I would keep searching for some good trafos on Ebay. The A-20s are not actually that good, you can do much better for less.
 
I have heard 5 different brands of decent quality output trafos with the cs4398. They all sounded different in one way or another, and in many cases are not subtle differences.

A published spec of 30-30khz is meaningless unless a reference to level and distortion at its lowest specified frequency is also listed.

Keep in mind that if you buy a good used pair of trafos and they are not to your liking you can easily resell them. That is not true when you buy cheap stuff at retail price.

Lundhals are problematic for the unknowledgable, generally because of the termination requirements.
Units like the A20s are very old technology but still are pleasant sounding. Another old unit to try would be the Altec/Peerless 15356, they pop up on Ebay all the time.
Cinemag makes good, inexpensive trafos, $40 and up.
Sowter has a few good inexpensive trafos
Jensen has models of all prices and specs.
I am currently using the Jensen JT 11-DM, I bought them off of Ebay for $26 a piece. They are the second best that they make and are my best, by far.
The Tamura GA80080 are also very good, Ebay bought for cheap.

The list is endless. In general, stay away from anything listed as an input trafo unless you have the patience and test equipment to get them set up right. Input and output trafos are not interchangable.

I hope I helped a little. Please go to the Jensen website and read the writeup about audio transformers by Bill Whitlock, it's great.
 
Bill,
you helped me really very much.
Thank you for your post ;)
Now I will start to take a look at ebay...
Do you know if the PCM1798 dac board can use the same trafos of cs4398 (1:1)?
I have 2 Dac board, mine is with cs4398 and another that will arrives soon with a pcm1798 dac..
Thank you again,
Dan
 
Behringer SRC2496 and UT A-20

Hi All,
My cheap Chinese Ebay DAC has died and I am now considering a Behringer SRC2496. They are about $200 from Amazon. I have my UTC A-20 transformers and am wondering if anyone has any experience with this kind of combination. Do you have any advice?

Lastly, what is a good way to connecting the transformers to the DAC output? I have seen several approaches. I am no engineer but know how to solder and experiment.

Thanks and I hope all is well with you.

Ciao - Mush