Have a look at my new TDA1545 DAC (lots of details on page)

[peufeu]

Hello !

I have made a non-OS dac with a TDA1545, featuring an innovative IV stage and jitter reduction techniques. Have a look :

http://peufeu.free.fr/audio/extremist_dac/

Any comments are appreciated

Regards
Pierre

[Pedja]

Will look for sure, just wanted to say: finally!

Pedja

[peufeu]

what do you mean by 'finally' ?

[Pedja]

It was about year ago…

http://pub134.ezboard.com/ffakeidsfr...icID=237.topic

I was patient, didn’t I?

Pedja

[Bricolo]

Very good work!
I'm even more impressed by your work on the website. Extremely complete.

http://peufeu.free.fr/audio/ can now be considered as a reference website for DAC builders

[tbla]

[jan.didden]

Originally posted by peufeu
Hello !

I have made a non-OS dac with a TDA1545, featuring an innovative IV stage and jitter reduction techniques. Have a look :

http://peufeu.free.fr/audio/extremist_dac/

Any comments are appreciated

Regards
Pierre

Hello Pierre, nice site!

Also nice design. Interesting I/V stage, very subtle! I still wrestle with the idea that the opamp is not in the signal loop. The opamp has a varying output signal, right? If not, there would be no point to have it in the circuit. Now, this varying output signal can only come from the inverting input signal. So, there IS an input signal. If I understand correctly, the task of the opamp is to keep the DAC output (which is at its inverting input) at zero volts. Suppose the inverting input wants to go pos. This is then corrected by the opamp. The opamp reacts on the error voltage. Looking at it this way, this opamp stage is a normal 100% feedback error loop. Hmm.

Jan Didden

[peufeu]

You're right ! Time goes so fast. I have so many other things to do, I almost didn't do anything audio-related for six months. But I'm back to it now.

[Bricolo]

One thing you could try is the pass labs D1's I/V stage

[peufeu]

Originally posted by janneman
(snip) ... Looking at it this way, this opamp stage is a normal 100% feedback error loop. Hmm. (snip)

First, the DAC output voltage must be kept at 3.3V not 0 volts, but that is not important regarding your remark.

Imagine there is no opamp. In that case we would feed the JFET grid from a constant reference voltage, which would be 3.3V + a fixed amount corresponding to the Vgs voltage of this JFET wiht this specific current. Thus, the JFET (acting as a follower) would keep its source at around 3.3V. This has two main drawbacks :

• The reference voltage would have to be adjusted according to the specific JFET used (this means a trimpot)
• The source voltage would vary a little with the DAC output current, because the source resistance is not that low.

Enter the opamp. Thanks to Feedback, it will apply a correct voltage on the JFET grid to get the right voltage on the DAC output. No more trimming and voltage variations.

Then, how is it out of the signal path ? Simple : the opamp output does not couple to the output signal. Voltage variations on the JFET grid do not transfer to the output signal except through the JFET's internal parasitical capacitance, which is low. Thus, the opamp is out of the signal path...

This is by contrast to the traditional IV using an inverting opamp, where the opamp transmits the full musical signal.

OK ?