The opa1632 is widely used as I/V for the pcm1794/8. I intend to use it in a DAC I'm building.
The problem is that the VOCM sets the voltage at the outputs of the opamp, not at the inputs. If VOCM is set to ground, the dac outputs will sit somewhere near2VDC (depending on I/V resistors). But it seems those DAC prefer a 0V output.
Here's what I'm considering:
R1/R2 are in fact a multi-turn trimmer to adjust voltage exactly.
The zener would be replaced by a lm336-5.
Does it look allright ?
The problem is that the VOCM sets the voltage at the outputs of the opamp, not at the inputs. If VOCM is set to ground, the dac outputs will sit somewhere near2VDC (depending on I/V resistors). But it seems those DAC prefer a 0V output.
Here's what I'm considering:
R1/R2 are in fact a multi-turn trimmer to adjust voltage exactly.
The zener would be replaced by a lm336-5.
Does it look allright ?
Attachments
If VOCM is set to ground, then the outputs of the opa1632 are held at ground. Now, the PCM1798 for example has an offset of 3.5ma on its outputs. Using 510R resistors for I/V, we end up with an offset of 1.785V at the inputs of the opa1632.
This is just the other way around when compared to the usual i/v opamp, where the offset will be found at the outputs of the opamp.
An example of the opa1632 being used as I/V is the IVY from twisted pear: http://www.twistedpearaudio.com/linestages/ivy.aspx
Apparently, the pcm179* tolerate the offset but I'd like to see if removing it it improves things.
This is just the other way around when compared to the usual i/v opamp, where the offset will be found at the outputs of the opamp.
An example of the opa1632 being used as I/V is the IVY from twisted pear: http://www.twistedpearaudio.com/linestages/ivy.aspx
Apparently, the pcm179* tolerate the offset but I'd like to see if removing it it improves things.
Have you actually measured the distortion of the system using an OPA1632 with VCOM grounded? It's pretty bad. Doing what you want, by shifting VCOM up with 3V will lift the outputs with 3V as well. So, you end up having to use a Single ended stage anyway to sum the whole lot together and get rid of the offset.
I've hit my head against the wall on this one so much that I ended up sticking to the single opamp per phase config as in the datasheet. The part that makes my head spin is that there seems to be people here that've had success using OPA1632's and THS4131's. I've tried the config at TP and to be honest, I couldn't get it to work. Or, at least, it worked but the distortion performance was not very good... Also, these chips got super hot, and no, they weren't oscillating. All I can assume is that it's not the right application for them.
But, I'd like to hear what others have done, with measurements to confirm the performance
I've hit my head against the wall on this one so much that I ended up sticking to the single opamp per phase config as in the datasheet. The part that makes my head spin is that there seems to be people here that've had success using OPA1632's and THS4131's. I've tried the config at TP and to be honest, I couldn't get it to work. Or, at least, it worked but the distortion performance was not very good... Also, these chips got super hot, and no, they weren't oscillating. All I can assume is that it's not the right application for them.
But, I'd like to hear what others have done, with measurements to confirm the performance
gbyleveldt said:
Well actually, I'm perfectly happy with that. It will allow me to bias a DS1802. I'll get rid of the offset with a balanced receiver afterwards (I need only single ended outputs). Btw, aren't you supposed to summ the whole thing anyway to cancel distortion ? For balanced outputs, TI suggests the use of two chips.
It's good to know that someone did measurements. I'll try to get access to a good soundcard to do some when I'm done. I'll let you know.
I'm curious about the chips getting hot. What were the supply voltage ? They'll be drawing around 20ma. At +/-15V, it would be a lot to burn for a soic. I'be using +/-10V supplies.
PS: according to the datasheet, I've to take the vcom (and thus the output) to -1.78VDC (with 510R I/V resistors) to get 0V at the inputs.
00940 said:
Hehe, yes, but I was trying to take a shortcut as I wanted to run the setup balanced using only 1 DAC instead of 2 as per the app note.
I ended up still using 5534's as the I/V and then fed the differential signals (with the -3.3V offset) to a THS4131. This gave me the results I wanted and pretty good THD of around -108dB keeping everything as balanced. No offset issues as it was summend in the THS with VCOM sitting at gnd.
00940 said:
I ran them on +-12V, but sticking to +-10V is a safer bet. I'm still uncomfortable with the heat though...
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.