• State Resolved
  • Locked Locked
  • Replies 9 replies
  • Subscribers 29 subscribers
  • Views 802 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

PCM1792A: PCAM17292 implementation question and design review

Briquel Florian
Intellectual 280 points
Part Number: PCM1792A
Other Parts Discussed in Thread: PGA4311, LM27762, PCM1792, OPA1632, OPA1612

Hello,

I'm designing an high end power audio amp and the design step is DAC. Then I have some questions regarding the use of PCM1792A.
I feed my power stage with diff audio signal about +/-0.75V (0.5Vrms).

I plan using a LM27762 from a main power rail just below 5.5V to feed DAC, Opamps and PGA4311. Do the LM27762 the job?
I plan using EN function of LM27762 to shut down analog rails when device is standby.

First can I leave powered ON the pin 9 VDD of PCM1792A (+3.3V) when other analog pins are shut down ? Or pin 9 and other analog pin must be powered on/off together?
Second, can I leave disconected PCM1792A current output when I select other line source with relays?
Third if you have any remarks, I'm very interested in (particulary using very low resistor value about 200R to make I/V conversion in order to reduce noise but how about distorsion?)

Many thanks for help!

  • 0
    TI__Guru 52663 points

    Hi Briquel,

    1. I think the LM27762 is a pretty good fit.  If ultra-low noise is design goal, you could consider an additional LDO for noise filtering after the charge pump.

    2.  You can leave the VDD pin at 3.3V when the rest of the device is unpowered.

    3. You can leave the IOUT pins disconnected when you switch sources.  

    4. I actually expect you will get better noise performance by increasing the resistor in the trans-impedance stage.  While it may not seem intuitive, maximizing the gain in the first stage and then attenuating in the second stage will give better performance.  You just need to worry about your supply limits.

    See this article:

    Otherwise, I would recommend using COG/NP0 type capacitors and thin-film resistors for best distortion performance.

    Thanks,

    Paul

  • 0 in reply to Paul_Frost
    Intellectual 280 points

    Hi Paul, nice to read some good news about VDD pin 9 of PCM1792A and for current output which can be left disconnected.

    Just one remark, didn't fully understand your remark about aditional ldo after LM27762, I understand LM22762 is composed of one low noise ldo for positive rail and one negative low noise ldo for the negative one (then after the charge pump of the component). So even the negative rail of lm27762 is already low noise, you confirm?

    Thanks for advice about resistor value, because it's only a game of value I could just try one time I build prototype. For capacitor I plan use MKP, and off course thin film resistor.

    Thanks for support!

  • 0 in reply to Briquel Florian
    TI__Guru 52663 points

    Hi Briquel,

    I suppose the LM27762 is pretty low-noise :)

    Looking at the datasheet I now see that there are LDOs integrated in the device, so you are pin pretty good shape.

    Thanks,

    Paul

  • 0 in reply to Paul_Frost
    Intellectual 280 points

    Your pleasure :) LM27762 seems to be a quite nice modern device for designing audio.

  • 0
    Intellectual 280 points

    Hello,

    I've additional question about the design.

    reminder; my design is based on PCM1792 with OPA1632 to make I/V conversion for diff output. resistor value of I/V converter are in my design set to 200R. Power rail are set to +/-5V

    I connect VCOM port of OPA1632 to gnd, but I read on the web this is not correct because the PCM1792 output current is centered on -6.2mA so with 200R resistor this make an voltage offset about 1.24V and this rise thd level significally. So in order to avoid thd rise I have to feed vcom port with a + 1.24V voltage source made with opamp for example. As I also use the OPA1632 to make a single endeed to diff signal converter, in this config I should set VCOM pin to gnd, so it's quite annoying.

    TI, do you confirm that VCOM port of OPA1632 must be set to bias current of PCM1792 * resistor value of I/V to get low thd ?

    Again thanks for support.

  • 0 in reply to Briquel Florian
    TI__Guru 52663 points

    Hi Briquel,

    The VOCM (output common-mode) pin is for the OPA1632's output common-mode, not the input. The output of the PCM I/V stage will have a common-mode of -1.24V, but it will not be changing so you will probably not see common-mode distortion issues.  Connecting VOCM to GND should be good.

    Thanks,

    Paul

  • 0 in reply to Paul_Frost
    Intellectual 280 points

    Hi again Paul,

    Thanks for answer, I think go with VCOM connected to gnd. My question is relative to some discussion on diyaudio about the fact that if pcm1792 will have less thd if there is almost 0v on the current output pin, but seems to be strange because pcm1792 power supply is only positive and then it's normal there is about vcc/2 voltage on the pins when runing.

    Sorry for the strange question.

    Regards, have a nice week end!

  • 0 in reply to Briquel Florian
    TI__Guru 52663 points

    Hi Briquel,

    I looked over your schematic again and now I have some more comments.  When I answered your last question I was thinking that your design had a trans-impedance (I/V) stage before the OPA1632.  That is not the case.

    This means that if the common mode of the output is 0V, then you could see about 6.2mA×200Ω=1.24V as the DC common-mode on the IOUT pins.  You would see 6.2±3.9mA × 200Ω as the dynamic voltage on those pins.  We do not explicitly say this is a problem, I know some of the people over at DIYaudio do it a lot.  If you increase the 200Ω resistors then the common mode would increase even more and could be a bigger problem.  The IOUT pins require some minimum supply headroom which could be compromised by the larger resistor.  It is for this reason that you see hobbyist projects like the DDDAC use a supply >5V, though I do not recommend that either.

    It might be wise to leave an unpopulated trans-impedance stage between the DAC and the OPA1632 just in case you want to compare the performance of both.

    Thanks,

    Paul

  • 0 in reply to Paul_Frost
    Intellectual 280 points

    Paul,

    Hmm my plan is to to go on industrialisation step with this design/project, so my I will follow the engineer reasonable choice despite the fact I would love to have a SuSy opamp to make I/V.

    Reasonnable path is:

    1 there is a potential issue using OPA1632 for transimpedance stage of pcm1792 so I move to a safe/tested design

    2 Even if the approach of a classic transimpedance stage followed by the opa1632 should works, rather prefer to keep it stupid simple and just use a classic transimpedance stage, for exemple with opa1612

    3 I keep opa1632 in my design but for the classic analog input, and I use just two opa1632 to make the single ended to diff converter (two because it's stereo amp), i use it with big resistor value like 2KR/4KR in order to keep my amp analog input impedance rather low, similation give me noise and thd rather low even with such big resistor values

    Maybe there is a reliable way using opa1632 as transimpedance stage for pcm1792a but I didn't know it, on diyaudio there is design with voltage value on VCOM port about 6.2mAxresistor of I/V but time and knowledge start to miss.