output stage for dac

[pagan]

Leo
I'll see what posts i can find

agree, never put dc servo feed back to input. no good.

The pcm1794 is hard to implement.
That was one of Jocko Homo's comments also well worth it when done right.
There is some clues referenced to in "Simple I/V Converter".
but Jocko only gave clues, for others to nut out.
rboer did about the most work on it.

I havn't had a current output dac to play with yet so i didn't investigate to much.

Probably will be the next dac though.

[leo]

Thanks Allan, from what I've been hearing from this dac so far, it is capable of producing excellent sound, I have the clock,regulation etc sorted out, its just the output stage that needs working on

I've not tried the Opus but have had access to play about with a couple of the newer Arcam cdp's which use the same chip but its not really fair to make a judgement going by those

BTW I did try Jocko's simple I/V stage with a different current output dac , I've not used it with the PCM yet

[pagan]

leo said:

BTW I did try Jocko's simple I/V stage with a different current output dac , I've not used it with the PCM yet

Rboer's version?

[leo]

pagan said:

leo said:

BTW I did try Jocko's simple I/V stage with a different current output dac , I've not used it with the PCM yet

Rboer's version?

It was the simple version, I'll find the link and post it

[pagan]

Leo
I was reading some stuff on the pcm1794's and it was about the dac likes to drive a low impedance.
referring to lower than 1k
Rboers dac input impedance is about 3k3?

[Deleted]

Allan
TI application notes show it with an opamp feedback resistor of 750 (or 820) ohms and 2200pf in parallel.
Alex

[pagan]

Alex
Whats the output impedance 20-20k?

[Deleted]

Allan
Doesn't appear to be stated. The attached may be of interest, as you can then work out what happens with those previously mentioned FB resistors.
Alex

APPLICATION CIRCUIT
The design of the application circuit is very important in order to actually realize the high S/N ratio of which the
PCM1794 is capable. This is because noise and distortion that are generated in an application circuit are not
negligible.
In the circuit of Figure 24, the output level is 2 V RMS, and 127 dB S/N is achieved. The circuit of Figure 25 can realize
the highest performance. In this case the output level is set to 4.5 V RMS and 129 dB S/N is achieved (stereo mode).
In monaural mode, if the output of the L-channel and R-channel is used as a balanced output, 132 dB S/N is achieved
(see Figure 26).
I/V Section
The current of the PCM1794 on each of the output pins (IOUTL+, IOUTL–, IOUTR+, IOUTR–) is 7.8 mA p-p at 0 dB (full
scale). The voltage output level of the I/V converter (Vi) is given by following equation:
Vi = 7.8 mA p–p × Rf (Rf : feedback resistance of I/V converter)
An NE5534 operational amplifier is recommended for the I/V circuit to obtain the specified performance. Dynamic
performance such as the gain bandwidth, settling time, and slew rate of the operational amplifier affects the audio
dynamic performance of the I/V section.
Differential Section
The PCM1794 voltage outputs are followed by differential amplifier stages, which sum the differential signals for each
channel, creating a single-ended I/V op-amp output. In addition, the differential amplifiers provide a low-pass filter
function.
The operational amplifier recommended for the differential circuit is the Linear Technology LT1028, because its input
noise is low.

[leo]

I like the look of this one, may try building a pair for each channels +/- output and then feed into the discrete differential to line out.
I think the JLH would be good for this circuit
www.pedjarogic.com/1541a/pdf/discrete_I-V.pdf

[Deleted]

Leo
I don't know how stable the 1.8mA (approx) from the FET would be with temperature. There may be better ways to null the DC offset.
Alex

[leo]

Yes, that nulling will drift a bit using a fet but it was mainly the active I/V stage I liked
I'll redraw the circuit how I'd like to try
I'm not going back to op-amps now, they sound too cruddy to my ears

[pagan]

leo said:

Yes, that nulling will drift a bit using a fet but it was mainly the active I/V stage I liked
I'll redraw the circuit how I'd like to try
I'm not going back to op-amps now, they sound too cruddy to my ears

Leo

Once you go discrete...............

it's hard to go back..

[leo]

I've got some free time today so making the most of it

[ryssen]

Wondering how large the outputsignal is in the schema in post 1?

[leo]

Just alter the pair of lower resistors on the differential input transistors to change the amplification

I no longer have the dac I used this output stage with

[leo]

Its very quiet on the forum lately, nobody else done any projects or whatever they want to share ?

[Deleted]

Jul 18, 2010 9:43:31 GMT leo said:

Its very quiet on the forum lately, nobody else done any projects or whatever they want to share ?

Leo
Hopefully, Jeff and Will will soon have further to report with their SC DACs.
Will has been a bit preoccupied with JLHs, and should soon be able to fit another one of them to his SC DAC now that he has changed to the simpler filtering ,and LME49710HAs.Jeff intends fitting another PSU PCB as work load permits.
Alex

[Deleted]

now that he has changed to the simpler filtering

Alex,

Could you clarify please, I'm not up to speed with this, unless you mean the "crickets", in which case they are next-up, time permitting.

Thanks,

Chris

[Deleted]

Jul 18, 2010 10:10:27 GMT @cjarchez said:

now that he has changed to the simpler filtering

Alex,

Could you clarify please, I'm not up to speed with this, unless you mean the "crickets", in which case they are next-up, time permitting.

Thanks,

Chris

Hi Chris
I mean the change of components on the Output PCB to the simpler circuit used by the PCM1792. It involved changing a few components, and even removing a few capacitors also. The changes were given in the "Tweaks" thread.
Alex

[Deleted]

Oh!

Thanks Alex, I'm nearly there on that one.

Chris

[pagan]

Jan 23, 2008 0:58:04 GMT leo said:

I decided to knock this output stage to try with my diy based PCM1794

I just need to tweak the gain by altering the 330R resistors as its a bit high output.

Results took me by surprise, where the hell did that bass come from!!!
Any comments on what I should try or change? sorry the schematic is a bit rough, it was quickly done also a bit small but is the largest I can post

This can also be used for the DSD1796 in the SC DAC.

Leo
Did you try the input CFP?

[Deleted]

This can also be used for the DSD1796 in the SC DAC.

Allan


I can think of at least 4 good reasons why not.
1. Lack of a PCB.
2.FETs should be matched for best results.
3.Output coupling capacitor, not a low impedance direct coupled output.
4. Insufficient filtering could cause problems with some amplifiers.
Anyway, it would need to be damn good to beat the current I.C. combination, especially with LME49710HA at the output.
Alex

[leo]

You could even configure IVYIII with the SC dac if you was bored OPA1632 is extremely good for a dac with differential output .

I certainly agree with the idea of going with the pcm1792 output stage, it looks loads better than the that config used with the stock SC dac, I've not heard it myself but imagine the improvement you get with altering to the PCM1792 values would be quite noticeable sonically

[leo]

Jul 18, 2010 11:29:57 GMT pagan said:

Jan 23, 2008 0:58:04 GMT leo said:

I decided to knock this output stage to try with my diy based PCM1794

I just need to tweak the gain by altering the 330R resistors as its a bit high output.

Results took me by surprise, where the hell did that bass come from!!!
Any comments on what I should try or change? sorry the schematic is a bit rough, it was quickly done also a bit small but is the largest I can post

This can also be used for the DSD1796 in the SC DAC.

Leo
Did you try the input CFP?

Hi Alan,

No, I'd got rid of that dac before the CFP idea was posted, I might knock another one of these stages up and try it with the Sabre, just for the laugh of course If I do I'll try the CFP, I did try the CFP with the Monica3 dacs output stage and noticed it brought an improvement in high frequency detail
We can use transistors or jfets in this stage, anybody building one should compare both

[Will]

Jul 18, 2010 9:55:27 GMT @sandyk said:

Will has been a bit preoccupied with JLHs...

That should read 'preoccupied with the chuffing, sodding JLH's CL section' Nevermind, it'll all come good in the end.

One of the other things that I've been playing with is this design, which I hope to test at the end of August/September, known as Will's Super-Duper Sabre Dac



Born out of the need for an easily had I2S signal.