Opamps for PCM1794 I/V conversion?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hey there guys, I almost posted this in the Parts section, but figured it would get more relevant views and opinions here. I have a dual mono PCM1794 PCB I picked up a while back, and I'm getting ready to actually finish it. The only opamps that I have on hand are the OPA445, OPA547T, and a handful of OPA2132P. I have used the OPA2132P in the past for single-chip I/V with the PCM1794 and it worked well, but I wanted to order some other chips to try this time around, seeing as how I'm ordering parts anyway.

Please do not recommend off-board solutions, or alternative I/V schemes, that's not the intent at this time.



This is the style of board in question. The one I have is barebones, no DAC chips, I/V stage caps, or opamps. Perfect starting point :)

721212d1544473647-opamps-pcm1794-conversion-2018-hi-fi-parallel-pcm1794a-dac-audio-decoder-assembled-_57-jpg


It uses six single, and four dual opamps total, and I was wanting to spend 3~5$USD or so per chip.



I was thinking for duals, since they are used as the I/V stage chips here, I would try the OPA2132P that I have on hand, and maybe some LME49720/LM4562. I may grab some NE5532 just for comparison. I'm strongly leaning to simply using the LME49720 here. It has very good specs, and has a strong pedigree f great measurements, sound, and performance.

I also might have a set of four Burson V6 "Vivid" duals on the way to test out (samples for review) in this position as well :)



For singles, maybe the OPA132, OPA227, the usual I see used on these boards/builds is the NE5534, but it feels pedestrian to use these here for much other than comparison purposes. Maybe the OPA1611 on a browndog adapter? I was originally thinking LME49710, but it's unobtainium after going EOL.

Seeing as how the board uses the singles merely for differential-to-single-ended conversion, I'm not so sure a fancy-pantsy chip makes much sense here, based on prior experience, after a certain point.


Any comments would be appreciated.
 

Attachments

  • 2018-Hi-Fi-Parallel-PCM1794A-DAC-Audio-Decoder-Assembled-_57.jpg
    2018-Hi-Fi-Parallel-PCM1794A-DAC-Audio-Decoder-Assembled-_57.jpg
    94.2 KB · Views: 2,371
As delivered, it's basically four complete sets of Fig.24 circuitry from the datasheet-

attachment.php



It only differs really in the fact that there is an additional single opamp per channel to further sum the differential outputs from each PCM1794 to single-ended outputs. The capacitors are not yet populated, but the included resistors match the datasheet values. The I/V opamps (U1/U2) are a dual opamp in this PCB implementation, all others are singles.
 

Attachments

  • pcm1794.png
    pcm1794.png
    117.9 KB · Views: 1,770
There are Brown Dog adapters for two singles to DIP, to plug into a DIP socket. Singles can be DIP or SMD.

When I see a lot of film caps around opamps, I usually check to see if the film caps are for power bypassing. If so, they may not be very useful at higher frequencies where opamps can have instabilities, or maybe no problems. Something to be aware of is all.
 
Last edited:
AD811 huh? That one and the AD797 were recommended to me as well. Hmmmm.

There are Brown Dog adapters for two singles to DIP, to plug into a DIP socket. Singles can be DIP or SMD.

When I see a lot of film caps around opamps, I usually check to see if the film caps are for power bypassing. If so, they may not be very useful at higher frequencies where opamps can have instabilities, or maybe no problems. Something to be aware of is all.

I'm not against the idea of adapters to run two singles on a dual socket, just trying to think of the best way. I'm not afraid of SMD either, and that would make going for a dual with single chips nice and compact too.

Also-
As shown, the red caps on the PCB picture are the film caps (2200/2700pf) on the schematic in the I/V and differential input summing stages, the PCB has ceramics to ground at each opamp supply pin.
 
I'm not against the idea of adapters to run two singles on a dual socket, just trying to think of the best way. I'm not afraid of SMD either, and that would make going for a dual with single chips nice and compact too.

AD811JR would be a good choice on a single-to-dual-SMD adapter but note its limited to +/- 12V operation due to running very, very toasty. On Aliexpress its about $1 a chip. I'd suggest no higher than +/- 10V rails for it. It is far and away better than AD8017 in subjective sound quality as I/V even though both chips have very similar paper specs.
 
No, mans.
PCM1794 output is really up to 7 mA.
DAC output are better to keep not more than +4:+6dbU.

This will demand for I/V resistance as low as 470 Ohms.

Provided schematic show us 750 Ohm I/V + -1/2 differential amplifier behind I/V stage.

Now check could single opamp provide THD levels comparative with PCM1794 being loaded to such resistance.

I.e. first stage opamps must be buffered for obtaining good performance. Even simple follower with tens mA bias current will greatly rise performance.

Best chips from “musical” reproduction view are LM6171, AD8065, AD797 and LT1364. Of course, properly cooked.
 
For I/V, filtering and buffering/cable driving I'm only using bipolar, unity gain stable, C-load , ultra fast (video) amplifiers: the only ones giving me emotion.

Hence try: AD826, LT1208, LM6172, LT1469 (the best on the paper)

For a different soundstage and 3D focus: LME49860, LT1358, AD827

Need a try LT1364, LT1355 and LT1361

All mentioned are dual DIP8.

I cannot stand anymore Jfet opamp sound, sorry
 
Last edited:
I have the single PCM1794 version of this board.
The AD797s with the dual adapters work well on this board.
Next thing is that these op amps need good decoupling so you might want to ensure that the power legs are decoupled very closely to ground using the appropriate caps.
On this board the AD797s are very sensitive to the IV stage power supply quality.
LM317/337 were lackluster with lack of real bottom end control and power with an emphasized lumpy upper midbass and no high frequency detail and definition.
Substituting a Jung Sulzer power supply driven by NE5534 did improve things nicely but not enough.
Adding a Jung Didden Super Regulator did open things up markedly and it started to sound in the correct direction. Still a bit of lumpiness in the low end but missing the detail and transparency found in a decently modded ESS9028 and likely 9038Q2m Dac. ( not op amp rolling....but power supplies)

Finally I added a matching power supply to this board driven by 3x5534 op amp rails. A power supply being sold for $20+ USD. It did not match the Jung Didden Super Regulators but I would put it a step above the Jung Sulzer Regulator with one caveat.

The caveat is that the PCM1794 chip is driven by an LT1761 that taps off the IV stage power supply. What I did was to remove this tap by removing the SMD jumpers and completely removing the LT1761-5 on the board. Then I directly fed an op amp driven power supply to the 1794 chip. This improved things a lot better BUT, again the speed of the 5534 could be discerned as lacking definition and transparency.

Given the low current requirements of this 1794, I suspect my next step is to create an op amp buffer power supply (AD797) to drive the 5V current requirements of this 1794. This is the next stage. Using something like an AD797 configured in a non inverting buffer configuration with a low noise 5 V on the +input will likely do the trick.

So without the last suggested mod, and even with the prior mods to the 5V, the 1794 lacks behind a current ESS chip dac well modded. With a good 5V supply to the 1794. It could possibly get close.

OP amp rolling compared to the suggested mods and not supplying top end power supplies is completely underwhelming. The 5v current supply to the 1794, needs to be stable and fast reacting and stable as heck....high speed op amps are needed. Spend the efforts there....not in the choice of op amps.

Finally, I will suggest a current boosted OPA1611 over the AD797 in the differential amp leg. It appears to sound better than the AD797.

Focus on the power supplies.....fast and stable with reasonable low noise.

Visit the huge thread on modifying the 9038q2m and learn about the importance of the current supply to the DAC chip. That is the weakest leg and then onto the IV stage and power supplies. I've given up on LT304x supplies. Indeed they are low noise but they are too slow and unstable dynamically...they give a very smooth sound but they lack detail and microdynamics and imaging in the sound. Been down that road.
 
Last edited:
I am using ADA4898-2 for the I/V and LT1028 for the rest. Can't say anything more because I'd have to build a proper linear PSU for the opamps, feeding them from a +-12V SMPS. No matter what I/V opamps I use (tried NE5532, NJM2014, ADA4898-2) they get melting hot, so i suspect some oscillation is going on there. Forgot to say that I've modded the analog supply voltage for the DAC to be 7V (similar to DDDAC1794). I/V resistors are 330.
 
For I/V, filtering and buffering/cable driving I'm only using bipolar, unity gain stable, C-load , ultra fast (video) amplifiers: the only ones giving me emotion.

Hence try: AD826, LT1208, LM6172, LT1469 (the best on the paper)

For a different soundstage and 3D focus: LME49860, LT1358, AD827

Need a try LT1364, LT1355 and LT1361

All mentioned are dual DIP8.

I cannot stand anymore Jfet opamp sound, sorry

That LT1469 looks fantastic on paper, and isn't too terribly priced in small quantities (10.78USD each for less than ten) and seeing as how there are only four needed. Wonder how well behaved it is?

The AD826 is a very capable looking opamp if we go off of the datasheet. It seems to have an interesting internal schematic too, with a cascoded gain stage internally. Price looks good at $7.09USD each too. A quick search for experience with this chip show it to be well behaved in almost all applications.

LM6172 looks like a middle-of-the road type. cheap too. Hmm.

LT1358 looks very nice on paper too.


The AD826 looks very promising.
 
I have the single PCM1794 version of this board.
The AD797s with the dual adapters work well on this board.
Next thing is that these op amps need good decoupling so you might want to ensure that the power legs are decoupled very closely to ground using the appropriate caps.
On this board the AD797s are very sensitive to the IV stage power supply quality.
LM317/337 were lackluster with lack of real bottom end control and power with an emphasized lumpy upper midbass and no high frequency detail and definition.
Substituting a Jung Sulzer power supply driven by NE5534 did improve things nicely but not enough.
Adding a Jung Didden Super Regulator did open things up markedly and it started to sound in the correct direction. Still a bit of lumpiness in the low end but missing the detail and transparency found in a decently modded ESS9028 and likely 9038Q2m Dac. ( not op amp rolling....but power supplies)

Finally I added a matching power supply to this board driven by 3x5534 op amp rails. A power supply being sold for $20+ USD. It did not match the Jung Didden Super Regulators but I would put it a step above the Jung Sulzer Regulator with one caveat.

The caveat is that the PCM1794 chip is driven by an LT1761 that taps off the IV stage power supply. What I did was to remove this tap by removing the SMD jumpers and completely removing the LT1761-5 on the board. Then I directly fed an op amp driven power supply to the 1794 chip. This improved things a lot better BUT, again the speed of the 5534 could be discerned as lacking definition and transparency.

Given the low current requirements of this 1794, I suspect my next step is to create an op amp buffer power supply (AD797) to drive the 5V current requirements of this 1794. This is the next stage. Using something like an AD797 configured in a non inverting buffer configuration with a low noise 5 V on the +input will likely do the trick.

So without the last suggested mod, and even with the prior mods to the 5V, the 1794 lacks behind a current ESS chip dac well modded. With a good 5V supply to the 1794. It could possibly get close.

OP amp rolling compared to the suggested mods and not supplying top end power supplies is completely underwhelming. The 5v current supply to the 1794, needs to be stable and fast reacting and stable as heck....high speed op amps are needed. Spend the efforts there....not in the choice of op amps.

Finally, I will suggest a current boosted OPA1611 over the AD797 in the differential amp leg. It appears to sound better than the AD797.

Focus on the power supplies.....fast and stable with reasonable low noise.

Visit the huge thread on modifying the 9038q2m and learn about the importance of the current supply to the DAC chip. That is the weakest leg and then onto the IV stage and power supplies. I've given up on LT304x supplies. Indeed they are low noise but they are too slow and unstable dynamically...they give a very smooth sound but they lack detail and microdynamics and imaging in the sound. Been down that road.

Interesting!


The main hurdle on this board is figuring out what I want to do for the PSU voltage regulation. I felt that the LM317/337 wouldn't cut it here. I was thinking of shunt regulators of some sort. I'll have to do some more reading to get ideas.
 
Last edited:
The AD826 looks very promising.


I have built an entire power amplifier with 30 of them... immagine how they can sound to me....:rolleyes:

Don't dislike LM6172: it has the highest slew-rate above all and sound excellent in IV DAC converters.

My suggestion: invest 50 bucks, buy all of them from an official distributor (too many chinese fakes) plus 8 DIP sockets and make a swapping exercise for few days.
Listening satisfaction will return to you the investment. Datasheet cannot say nothing about the final soundstage and result you will get.
 
Hey there guys, I almost posted this in the Parts section, but figured it would get more relevant views and opinions here. I have a dual mono PCM1794 PCB I picked up a while back, and I'm getting ready to actually finish it. The only opamps that I have on hand are the OPA445, OPA547T, and a handful of OPA2132P. I have used the OPA2132P in the past for single-chip I/V with the PCM1794 and it worked well, but I wanted to order some other chips to try this time around, seeing as how I'm ordering parts anyway.

Please do not recommend off-board solutions, or alternative I/V schemes, that's not the intent at this time.



This is the style of board in question. The one I have is barebones, no DAC chips, I/V stage caps, or opamps. Perfect starting point :)

721212d1544473647-opamps-pcm1794-conversion-2018-hi-fi-parallel-pcm1794a-dac-audio-decoder-assembled-_57-jpg


It uses six single, and four dual opamps total, and I was wanting to spend 3~5$USD or so per chip.



I was thinking for duals, since they are used as the I/V stage chips here, I would try the OPA2132P that I have on hand, and maybe some LME49720/LM4562. I may grab some NE5532 just for comparison. I'm strongly leaning to simply using the LME49720 here. It has very good specs, and has a strong pedigree f great measurements, sound, and performance.

I also might have a set of four Burson V6 "Vivid" duals on the way to test out (samples for review) in this position as well :)



For singles, maybe the OPA132, OPA227, the usual I see used on these boards/builds is the NE5534, but it feels pedestrian to use these here for much other than comparison purposes. Maybe the OPA1611 on a browndog adapter? I was originally thinking LME49710, but it's unobtainium after going EOL.

Seeing as how the board uses the singles merely for differential-to-single-ended conversion, I'm not so sure a fancy-pantsy chip makes much sense here, based on prior experience, after a certain point.


Any comments would be appreciated.

If you need, I still have a large stock of LME49710 and LME49720, TO99 metal case, also LM4562.
Regards
 
missing the detail and transparency found in a decently modded ESS9028 and likely 9038Q2m Dac. ( not op amp rolling....but power supplies)

Mike,
Interesting. However, power supplies and opamps (and the proper decoupling of such) are only part of what is needed for best Sabre sound quality. Getting jitter to ultra low levels and tweaking registers for optimum sound quality does more things than good power supplies and clean output stages in isolation can do. Also, DSD gets away from the sound of Sabre built-in PCM interpolation filters. It all matters, and there is a lot more even you haven't heard yet.
 
I know you asked for no alternative schemes but if the board is really barebone, it would be easy to implement on it a passive I/V scheme, with just 22R resistors instead of the I/V opamp and a bit more gain at the output. It would save you 4 dual opamps and relax the PS requirements. smms73 has posted some great info on the pcm1794's thd with such solution.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.