I have a question about the PCM1794a. As I read the datasheet, there is a -6.2mA offset at the output pins. How are we supposed to handle that? In the demonstration circuits shown, there doesn't appear to be anything done about it. The outputs from the DAC goes directly to the inverting pin of the opamps. Do opamps not care about a DC current offset? I'm confused.
It's all about the magic of feedback.
OK, so you're saying I think that the opamps have enough "compliance" or what ever it's called to handle a DC current offset presented to the negative input pin when it's arranged as a transimpedance amplifier with feedback, which is the way they are typically used in I/V converters. Is that right?
What about using a discrete op amp instead of the chips? If performance isn't any better, at least it would be more fun. I don't want to turn into a "chip jockey".
Thanks. I looked at the schematic and I don't understand it. It appears to be highly paralleled mosfet's in common gate configuration, is that right?
There is no loop feedback, so I am unaware as to how the DC current offset of this DAC is handled by the Legato. In my sim's, the input for a common gate/base configuration is highly sensitive to DC offset.
I think this particular DAC is pretty hard to work with, and maybe not worth the effort.
No its pretty easy to work with really.
Legato circuit is actually a bit more than that. There are parallel mosfets, but that is only to reduce their impedance in their linear region.
The Legato is a simple balanced output I/V stage. At the balanced outputs it will have common mode DC offset (same on both phases), but you will use output caps if your balanced gear can't handle that (most can without any issue). Differential offset is adjusted out with a pot and can be brought to zero. If you need SE output Legato uses a separate BAL/SE stage that also has a pot to adjust to 0 offset. One bonus is that the SE output is buffered and can drive headphones.
Not hard at all
If you don't want to adjust anything use IVY-III instead. It will work as stock, and if you like you can double the output swing. With IVY-III all outputs will swing around ground, because it does use global feedback (super symmetrical at the I/V stage) at all stages.
Legato circuit is actually a bit more than that. There are parallel mosfets, but that is only to reduce their impedance in their linear region.
The Legato is a simple balanced output I/V stage. At the balanced outputs it will have common mode DC offset (same on both phases), but you will use output caps if your balanced gear can't handle that (most can without any issue). Differential offset is adjusted out with a pot and can be brought to zero. If you need SE output Legato uses a separate BAL/SE stage that also has a pot to adjust to 0 offset. One bonus is that the SE output is buffered and can drive headphones.
Not hard at all
If you don't want to adjust anything use IVY-III instead. It will work as stock, and if you like you can double the output swing. With IVY-III all outputs will swing around ground, because it does use global feedback (super symmetrical at the I/V stage) at all stages.
I'm still flopping about like a fish out of water about this. Does anyone have distortion readings for the passive I/V stage used on the COD? Does the distortion go down if the I/V resistor is made smaller? How does the distortion compare to using an active I/V stage with virtually nil input impedance?
Also, is there much of a common mode signal output on the Iout+ and Iout- pins from this DAC? I have read that at least some DAC's have significant common mode noise in their outputs.
Can the DC blocking capacitor be used without the I/V resistor? The DAC has a DC offset of -6.2mA which can cause problems with some types of I/V stages.
Also, is there much of a common mode signal output on the Iout+ and Iout- pins from this DAC? I have read that at least some DAC's have significant common mode noise in their outputs.
Can the DC blocking capacitor be used without the I/V resistor? The DAC has a DC offset of -6.2mA which can cause problems with some types of I/V stages.
with a 90 ohm IV resistor there is about 1v of DC offset out of the DAC. you need the caps to block this DC from a following Buffer stage or use some form of compensation like a servo etc. The distortion is very low using a resistor. I would bet an active stage would have comparable results but a different harmonic structure to the distortion.
I know the active vs passive thing has been debated to death here. but I can tell you first hand a passive IV stage with a jfet buffer sounds WONDERFUL! that is of course MPO on the subject. and YMMV but it is an easy thing to experiment with and see what works for you!
Zc
I know the active vs passive thing has been debated to death here. but I can tell you first hand a passive IV stage with a jfet buffer sounds WONDERFUL! that is of course MPO on the subject. and YMMV but it is an easy thing to experiment with and see what works for you!
Zc
Thanks. I saw somewhere that someone varied the I/V resistor of some other kind of current output DAC and got much worse distortion readings with higher resistor values. Let me go search on the forum for something about this. I initially did a Google search and got a few results, but nothing good, hence my question here since I have this DAC and COD board.
I have a schematic for a passive I/V stage with jfet buffer. No coupling capacitor is required since the DC offset is so small that the CRD in the circuit compensates just fine. The I/V resistors are 127 ohms in this circuit, for about 1.3Vrms output.
I was playing around with a Zen type of active I/V stage using bipolar transistors but the input is very sensitive to DC offset. I put a blocking capacitor at the input (no resistors at all) and it seemed to work just fine with no loss of bass response.
I have read that the current output DAC's will output the same AC current regardless of the load impedance up to the maximum voltage they can put out, but I have not seen a specific distortion vs. impedance curve anywhere.
I did some searching on finding a distortion vs. I/V resistor value chart or something like that with no luck. I don't think it's a linear relationship, but that's just a hunch. I do know that if the resistor is too large then the "compliance" limit will be exceeded such that the voltage output will be clipped. People keep repeating this mantra that a current output DAC must see the lowest possible load impedance but no one puts up any proof. I did a lot of searching Google and reading various websites with no luck so far. The DAC datasheets are no help, since all they say is "do it our way to achieve our results".
So, how low is low enough? 9 ohms? the Zen (Sen, Cen, etc.) only manages 15 ohms. If you use bjt's instead of jfets you can get less than 1 ohm input impedance, according to my simulations.
Also, there is a zero field transformer input circuit that can achieve very low input impedance, depending on the transformer. The Lundahl LL1517 has a winding resistance of some 9 ohms. With the zero field circuit, the input impedance can be made to be 9 ohms and the frequency response can be extended to 1 Hz. The advantage with a transformer is ground isolation. My simulations show this circuit to have high noise though, but I think that's a software issue.
So, how low is low enough? 9 ohms? the Zen (Sen, Cen, etc.) only manages 15 ohms. If you use bjt's instead of jfets you can get less than 1 ohm input impedance, according to my simulations.
Also, there is a zero field transformer input circuit that can achieve very low input impedance, depending on the transformer. The Lundahl LL1517 has a winding resistance of some 9 ohms. With the zero field circuit, the input impedance can be made to be 9 ohms and the frequency response can be extended to 1 Hz. The advantage with a transformer is ground isolation. My simulations show this circuit to have high noise though, but I think that's a software issue.
There's some spec floating around that the DAC pins shouldn't see more than 25mV but that's probably only with the really antique DAC's.
One way to do this is to do a passive I/V with a really low value resistor, and then use a normal gain stage after that. I put the RLC filter in with the I/V resistor before the gain stage. S/N is around 120-125dB depending on what kind of gain stage is used.
One way to do this is to do a passive I/V with a really low value resistor, and then use a normal gain stage after that. I put the RLC filter in with the I/V resistor before the gain stage. S/N is around 120-125dB depending on what kind of gain stage is used.
http://www.diyaudio.com/forums/digital-source/221743-testing-pcm1794.html
It seems like 50R resistor is the way to go for passive I/V for this DAC. I'm going to try this with the Ventus as buffer.
It seems like 50R resistor is the way to go for passive I/V for this DAC. I'm going to try this with the Ventus as buffer.
Interesting... I was experimentig with higher and lower values for I/V. I have a AD1865 and found the lower values being better for my taste. I am just curious has someone tried less then 1ohm resistor for current to voltage conversation. I know you need a phono or mic input stage but it would be interesting to try....
Hi,
If I want to use two COD boards in dual mono configuration, what is the best solution to feed them with i2s input signals (from a USB to i2s converter) ? Just wire both PCM terminal blocks in parallel ?
Will there be a shift between the left and right data bits ?
You just wire them in parallel, you set one board for left and one for right. I found it best to connect each board to the source with separate runs.
Zc
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.