NAD C658 measurements and teardown Q

[kukynas]

1 hour ago, Puresound said:

Hi Dan,

I have found your very interesting project and I'm also owner of a C658 and have been doing some of the same thought of modifiying the PSU, OPA, clock.

Now I have some questions to your detailed teardown of the C658.

- Why do they have so 12 pin connector between the SMPS and mainboard? It seems like there is only 14V DC and ground!

- The XP Power SMPS (looks like ECM model) that you have chosen are very compact but looking at the R&N then it is 1% pk2pk which is normal for SMPS or have I missed something about those SMPSs?

- Do you have a picture where you connected your 5V DCs when having removed varius buck converters?

- It seems like there many NE5532 OPA (I counted 11) why so many? Have you replaced them all with OPA1612?

- Are there any low quality caps in the output signal path?

- I was thinking of removing the crystal and install a reference clock instead. What is your opnion abou that?
 

 

Hi,

 

- no idea and yes, only 14VDC + ground

- yes, R&N is normal like for any other SMPS, the only reason why I chosen this one instead of Delta with 5x lower R&N is size and regulation, you need to get down to 14V from your typical 15V output via output trim and to be able to squeeze two of them on 1 PCB, one for 14V and the other for 5V (7V output SMPS post regulated with 5V ULN reg) 

- see picture below, basically same input as with buck converter via 8 pin connector as I wanted to be able to disconnect it during further upgrades + I didn't wanna cut any traces if I would have to go back and solder back the buck converter 

- I/V stage + buffer stage (balanced + unbalanced) + headphone amp and not yet, planning to do next but only for I/V and balanced buffer stage, don't care about the rest

- nope

- wouldn't do it, you won't find better clock in this package with same fq and even if than more beneficial is upgrade of the reg with clean power as I did, picture below

 

as you can see from measurements the only benefit in my POV is better reg for clock, all the other upgrades didn't bring any measurable improvement so if I'm doing it again the only upgrade would be reg for clock, now coz I don't have previous measurements of the streamed signal (instead of coax) before upgrade I can't confirm and compare if the latest upgrade of the analog PSU rails helped but if yes again the benefit is very small

As mentioned in my earlier post you can't change all buck converters due to fact they are driven by MCU for stand by functionality and you won't find linear reg counterpart (I tried) with 3A output so all the potential benefits by changing only one or few of them are basically useless as shown in my measurements.

lets wait for my last attempt (opamps) and see what (if) contributes the most to the overall result

 

 

 

 

[Puresound]

On 12/02/2021 at 1:36 PM, kukynas said:

a new day (or maybe 3 months) has come and we have some progress...

 

finally was able to obtain decent signal generator to be able to measure streaming performance (instead of coax as with all previous tests) so below some figures, overall measurements of streaming module reveal lower low noise distortion and jitter side bands, on the other hand slightly higher harmonics, I can't perform entire set of measurements but this should give us enough clue of what to expect, just a note, all my measurements are undithered for those interested

In general, streaming files over network should provide better result than using C658 as coax fed DAC

in the meantime I performed stage 2 upgrade of the analog PSU rails but haven't seen significant improvements, something is still limiting factor but I expect to see some improvements after facelift of the I/V and buffer stage, fingers x

 

 

with dithered signal, higher noise floor lower harmonics

 

 

noise floor of the unit no signal playing

 

Hi Dan,

 

Interesting that it did not help to upgrade analog PSU rails. It is the LM317s and LM337s? How does NAD convert the 14V to +/- x Volt for OPAs? I have seen in M12 that they use LM2611 but I can't spot it around the LM3x7s

[Puresound]

On 14/02/2021 at 12:41 PM, kukynas said:

 

Hi,

 

- no idea and yes, only 14VDC + ground

- yes, R&N is normal like for any other SMPS, the only reason why I chosen this one instead of Delta with 5x lower R&N is size and regulation, you need to get down to 14V from your typical 15V output via output trim and to be able to squeeze two of them on 1 PCB, one for 14V and the other for 5V (7V output SMPS post regulated with 5V ULN reg) 

- see picture below, basically same input as with buck converter via 8 pin connector as I wanted to be able to disconnect it during further upgrades + I didn't wanna cut any traces if I would have to go back and solder back the buck converter 

- I/V stage + buffer stage (balanced + unbalanced) + headphone amp and not yet, planning to do next but only for I/V and balanced buffer stage, don't care about the rest

- nope

- wouldn't do it, you won't find better clock in this package with same fq and even if than more beneficial is upgrade of the reg with clean power as I did, picture below

 

as you can see from measurements the only benefit in my POV is better reg for clock, all the other upgrades didn't bring any measurable improvement so if I'm doing it again the only upgrade would be reg for clock, now coz I don't have previous measurements of the streamed signal (instead of coax) before upgrade I can't confirm and compare if the latest upgrade of the analog PSU rails helped but if yes again the benefit is very small

As mentioned in my earlier post you can't change all buck converters due to fact they are driven by MCU for stand by functionality and you won't find linear reg counterpart (I tried) with 3A output so all the potential benefits by changing only one or few of them are basically useless as shown in my measurements.

lets wait for my last attempt (opamps) and see what (if) contributes the most to the overall result

 

 

 

 

Hi Dan,

 

Thanks for sharing your pictures and experience.

I have some followup questions.

I plan to have ultra low noise voltage (like Belleson SPX or Sparkos SSXX) after the 2 SMPS to generate 14V and 5V, respectively.

I can see you took out the U807 and connected the 5V reference - what parts does it provide voltage to? What about U806 and U805?

Did you remove some component to install the reg for the clock?

I forgot the Headphone output which assume would be OPA U104, 106, 108, 109, 110? I'm not interested in that but will plan on doing the other when I hear your experience afther having changed those

 

 

[kukynas]

3 hours ago, Puresound said:

Hi Dan,

 

Interesting that it did not help to upgrade analog PSU rails. It is the LM317s and LM337s? How does NAD convert the 14V to +/- x Volt for OPAs? I have seen in M12 that they use LM2611 but I can't spot it around the LM3x7s

 

yes LM3XX , via diode and than regs down to +/-12V

 

2 hours ago, Puresound said:

Hi Dan,

 

Thanks for sharing your pictures and experience.

I have some followup questions.

I plan to have ultra low noise voltage (like Belleson SPX or Sparkos SSXX) after the 2 SMPS to generate 14V and 5V, respectively.

I can see you took out the U807 and connected the 5V reference - what parts does it provide voltage to? What about U806 and U805?

Did you remove some component to install the reg for the clock?

I forgot the Headphone output which assume would be OPA U104, 106, 108, 109, 110? I'm not interested in that but will plan on doing the other when I hear your experience afther having changed those

 

 

 

- don't know Belleson but Sparkos doesn't have enough current output, you need 3A output regs, I bought mine voltage reference regs (and many other PSUs) from Kim over here  ULN-HC78A Ultra Low Noise Positive Regulator 2.5V~30V Output High Current Type | eBay

- if you have bigger budget this is SOTA low dropout voltage reg capable of 6A output ULN-LD78 Ultra Low Noise, LDO(Very Low Dropout) Regulator 1.25V ~ 30V | eBay

- to DAC chip, input board regs and most probably part of the DSP board, U806 is 5v bc for DSP board, U805 is 3.3V for add on cards and don'w know what else, there is another bc underneath the board no idea what for  

- as mentioned earlier I couldn't measure any improvement by changing U807 to ULN reg so the old buck converter going back and would recommend to keep it there, no point in spending money with no benefit, once it's back I will connect my lab linear PSU to see if I can get rid off those psu spikes, if yes then linear PSU is potentially on the table

- yes, resistor (R740 next to the oscillator) which connects the original reg to the oscillator

- yes except U110, no chance to trace that opa, no idea if it's part of the IV but just in case I changed that too...

 

as shown below with changing OPAs we get 5-10db lower harmonics but that's it, I think we are hitting the NRC chip distortion so not much more gain to be expected means no reason to spend more time nor money on it

 

PSU distortion and bottom harmonics are around -135db so hardly can contribute to the sound, main harmonics are there but in low level at nearly full scale (music content never gets to such amplitude) so I'm not so worried about them and as previously mentioned the overall presentation sounds nearly identical to my RME Adi 2 Pro fs so pretty good result I would say...

 

 

 

 

[davewantsmoore]

On 24/12/2020 at 6:41 AM, stereo coffee said:

No not tested, rather manufactured and tested, to levels of sensitivity that match lets say within 10% of the actual RMS voltage levels available in media. 

 

Why RMS?

 

On 24/12/2020 at 6:41 AM, stereo coffee said:

The majority of amplifiers sold today, and yesterday are voltage amplifiers - nothing more nothing less, they simply respond to the input voltage that is needed for delivering power in a 8 ohm resistance at a certain frequency usually 1khz ,  hence if used with media that does not match their sensitivity, the opportunity of available power delivery is very much lessened.

 

I think you write this in a strange way that will confuse many..... but I agree with you.

 

Gain structure is of upmost importance......    otherwise we will need to add amplifying (or de-amplifying) components in between our source and (ideally a single) gain stage.....  which will likely have unwanted reactance.

[Puresound]

2 hours ago, kukynas said:

 

yes LM3XX , via diode and than regs down to +/-12V

 

 

- don't know Belleson but Sparkos doesn't have enough current output, you need 3A output regs, I bought mine voltage reference regs (and many other PSUs) from Kim over here  ULN-HC78A Ultra Low Noise Positive Regulator 2.5V~30V Output High Current Type | eBay

- if you have bigger budget this is SOTA low dropout voltage reg capable of 6A output ULN-LD78 Ultra Low Noise, LDO(Very Low Dropout) Regulator 1.25V ~ 30V | eBay

- to DAC chip, input board regs and most probably part of the DSP board, U806 is 5v bc for DSP board, U805 is 3.3V for add on cards and don'w know what else, there is another bc underneath the board no idea what for  

- as mentioned earlier I couldn't measure any improvement by changing U807 to ULN reg so the old buck converter going back and would recommend to keep it there, no point in spending money with no benefit, once it's back I will connect my lab linear PSU to see if I can get rid off those psu spikes, if yes then linear PSU is potentially on the table

- yes, resistor (R740 next to the oscillator) which connects the original reg to the oscillator

- yes except U110, no chance to trace that opa, no idea if it's part of the IV but just in case I changed that too...

 

as shown below with changing OPAs we get 5-10db lower harmonics but that's it, I think we are hitting the NRC chip distortion so not much more gain to be expected means no reason to spend more time nor money on it

 

PSU distortion and bottom harmonics are around -135db so hardly can contribute to the sound, main harmonics are there but in low level at nearly full scale (music content never gets to such amplitude) so I'm not so worried about them and as previously mentioned the overall presentation sounds nearly identical to my RME Adi 2 Pro fs so pretty good result I would say...

 

 

 

 

Hi,

 

I have tried to find the setup you are describing. So there is a virtual ground for creating the +/-7V like the enclosed figure?

The Belleson has 3A capabillity see link https://www.belleson.com/store/SPX-Max-Performance/SPX78-Positive-Output

Also nice Regs that you have found. I have previously looked at their clocks.

It sounds like with the OPAs, ref reg on the clock and possible LPS then it becomes a very attractive og good sounding preamp combined with all its nice fetures builtin.

Have you experimented with discrete OPAs instead of OPA 1612 - like Sparkos, Burson og NewclassD/Dexa?

 

 

[stereo coffee]

4 hours ago, davewantsmoore said:

 

Why RMS?

 

 

I think you write this in a strange way that will confuse many..... but I agree with you.

 

Gain structure is of upmost importance......    otherwise we will need to add amplifying (or de-amplifying) components in between our source and (ideally a single) gain stage.....  which will likely have unwanted reactance.

When matching sensitivity ,  0.316v RMS is the consumer equipment line level standard.   https://en.wikipedia.org/wiki/Line_level

which can also be expressed as peak to peak, or indeed dbv , or +dbu,   RMS is though a common form for quoting amplifier sensitivity. 

 

Gain structure as you say is of utmost importance,  here we see in light blue RMS of a flac file. ( Swiss band Sonar - playing flat out )

Playing this through a passive system with amplifier sensitivity ( quad 306 ) of 0.375v RMS, there is then much benefit in not adding reactance by changing level, to suit a different sensitivity.   The source then matches to the power amp, and all that is needed in between is passive resistance to attenuate. 

 

 

 

[kukynas]

7 hours ago, Puresound said:

Hi,

 

I have tried to find the setup you are describing. So there is a virtual ground for creating the +/-7V like the enclosed figure?

The Belleson has 3A capabillity see link https://www.belleson.com/store/SPX-Max-Performance/SPX78-Positive-Output

Also nice Regs that you have found. I have previously looked at their clocks.

It sounds like with the OPAs, ref reg on the clock and possible LPS then it becomes a very attractive og good sounding preamp combined with all its nice fetures builtin.

Have you experimented with discrete OPAs instead of OPA 1612 - like Sparkos, Burson og NewclassD/Dexa?

 

 

 

Hi,

don't think so, it looks each + and - has it's own traces and diode (via D800) seems to be only on "-" rail but I didn't paid too much attention to it.

sure Belleson is 3A I just never used this reg hence "I don't know", I was referring to Sparkos in terms of output requirements

I'm using his design for nearly a decade and always happy with his products

yes, definitely especially if you consider there's no other product (yet) on the market with all those features 

yes, in my previous Sabre DAC, Sparkos and Burson, sparkos were OK but nothing outstanding to be paying premium for, burson was crap, one of them melted down in that plastic tower 

[davewantsmoore]

4 hours ago, stereo coffee said:

0.316v RMS is the consumer equipment line level standard.

 

If we are testing the performance of a circuit.... we care about the performance at all levels, not just at the average.

 

It is not misleading to test a device with 2V.     If people misunderstand what the test means the that is their fault...... so the "people doing tests at ASR with 2V" are not being "misleading".

[stereo coffee]

24 minutes ago, davewantsmoore said:

 

If we are testing the performance of a circuit.... we care about the performance at all levels, not just at the average.

 

It is not misleading to test a device with 2V.     If people misunderstand what the test means the that is their fault...... so the "people doing tests at ASR with 2V" are not being "misleading".

They are misleading everyone,  IMO when they totally fail and outright refuse,  also to concurrently test,  at actual levels.

[Puresound]

3 hours ago, kukynas said:

 

Hi,

don't think so, it looks each + and - has it's own traces and diode (via D800) seems to be only on "-" rail but I didn't paid too much attention to it.

sure Belleson is 3A I just never used this reg hence "I don't know", I was referring to Sparkos in terms of output requirements

I'm using his design for nearly a decade and always happy with his products

yes, definitely especially if you consider there's no other product (yet) on the market with all those features 

yes, in my previous Sabre DAC, Sparkos and Burson, sparkos were OK but nothing outstanding to be paying premium for, burson was crap, one of them melted down in that plastic tower 

Hi,

Thanks again.

Sorry I can't figure out the setup with one diode and with a LM337 handling the "-" rail?

There are 2pcs of LM317 and LM337 right? why having 2 sets of Dual voltage PSU?

Yes I have not been able to find another Preamp with MQA, Dirac, direct support for streaming service and HDMI input/output so I can get video streaming audio sound into the system as well without having a seperate system for that.

As I haven't open the box yet. what is the space between mainboard and DSP board on top?

I have not heard so much good about Burson either as the sound is not natural

[kukynas]

11 minutes ago, Puresound said:

Hi,

Thanks again.

Sorry I can't figure out the setup with one diode and with a LM337 handling the "-" rail?

There are 2pcs of LM317 and LM337 right? why having 2 sets of Dual voltage PSU?

Yes I have not been able to find another Preamp with MQA, Dirac, direct support for streaming service and HDMI input/output so I can get video streaming audio sound into the system as well without having a seperate system for that.

As I haven't open the box yet. what is the space between mainboard and DSP board on top?

I have not heard so much good about Burson either as the sound is not natural

 

I will check the circuit and mark the traces of the - rail.

One set is for opamps (+/-12v) and the other set for NRC chip (+/- 7.5v)  maybe even for headphone opamps but didn't check that so not sure

I think space is around 30mm but can measure it properly once I'm back home from work

yep and sounded distorted, there's really no need to pay extra money for luxury/boutique opamps if you can get cheap, reliable and low distortion one from TI which is all you need in something doing D/A conversion

[kukynas]

4 hours ago, stereo coffee said:

They are misleading everyone,  IMO when they totally fail and outright refuse,  also to concurrently test,  at actual levels.

 

I don't know why it is so hard to understand the logic, ASR, Stereophile, all other audio publications including manufacturers of products all just follows industry standard set by orgs like AES, EBU, ITU etc. which specifies at which level the measurements supposed to be performed, if you don't like it go and complain to those organization and ask them to change it 

It's like complaining that car manufactures in EU (don't know the rest) providing fuel consumption at 110km/h misleading consumers coz it's old standard and highway limit for last 3 decades is 130km/h, should I complain about it to VW and all the other car manufactures that they are misleading customers as there is either 90km/h limit for country roads or 130km/h limit for highway so their 110km/h limit is completely useless? no, coz they are just following standard set by car industry to be compliant 

[Puresound]

4 hours ago, kukynas said:

 

I will check the circuit and mark the traces of the - rail.

One set is for opamps (+/-12v) and the other set for NRC chip (+/- 7.5v)  maybe even for headphone opamps but didn't check that so not sure

I think space is around 30mm but can measure it properly once I'm back home from work

yep and sounded distorted, there's really no need to pay extra money for luxury/boutique opamps if you can get cheap, reliable and low distortion one from TI which is all you need in something doing D/A conversion

That would be great. I just wonder how to get from +14V to +/-12 V without any "conversion/chip" .

I hope you are right about the space between the boards as this should be sufficient.

So you don't think there is sound effect from the luxury OPA. So which OPA chips do you think performs best?

[stereo coffee]

3 hours ago, kukynas said:

 

I don't know why it is so hard to understand the logic, ASR, Stereophile, all other audio publications including manufacturers of products all just follows industry standard set by orgs like AES, EBU, ITU etc. which specifies at which level the measurements supposed to be performed, if you don't like it go and complain to those organization and ask them to change it 

It's like complaining that car manufactures in EU (don't know the rest) providing fuel consumption at 110km/h misleading consumers coz it's old standard and highway limit for last 3 decades is 130km/h, should I complain about it to VW and all the other car manufactures that they are misleading customers as there is either 90km/h limit for country roads or 130km/h limit for highway so their 110km/h limit is completely useless? no, coz they are just following standard set by car industry to be compliant 

It is illogical to measure equipment at 12.69 x  levels,  ( 4 V RMS ASR typical    /  0.315 RMS line level standard )  greater than the mediums being played by equipment,  actually attain in normal use.

 

Drawing comparison to your reply, like testing a VW at  1395.9 kph, ( 110kph x 12.69  ) and claiming,  that is just following the standard of compliance. 

 

 

[kukynas]

2 hours ago, Puresound said:

That would be great. I just wonder how to get from +14V to +/-12 V without any "conversion/chip" .

I hope you are right about the space between the boards as this should be sufficient.

So you don't think there is sound effect from the luxury OPA. So which OPA chips do you think performs best?

- you've been right, it's S10U45S reverse rectifier fed by LM26670 (from the other side of the board, that magic bc I wasn't sure it's meaning) 

- 27mm 

- unless you are looking for some specific voicing (no idea which) to your sound I don't think you need luxury OPA, currently installed is OPA1612, previously been using LM4562 in other projects with same success, both low noise, low distortion, high PSRR and CMRR opamps   

[kukynas]

6 minutes ago, stereo coffee said:

It is illogical to measure equipment at 12.69 x  levels,  ( 4 V RMS ASR typical    /  0.315 RMS line level standard )  greater than the mediums being played by equipment,  actually attain in normal use.

 

Drawing comparison to your reply, like testing a VW at  1395.9 kph, ( 110kph x 12.69  ) and claiming,  that is just following the standard of compliance. 

 

 

 

they all follow the standards set by industry, no matter if logical or not or to your or mine liking but that doesn't mean they (reviewers or publications) are misleading people or customers 

yes, we can complain that current standards aren't useful or up to date but that's completely different story directed to different folks, those who sets the standards  

[stereo coffee]

2 minutes ago, kukynas said:

 

they all follow the standards set by industry, no matter if logical or not or to your or mine liking but that doesn't mean they (reviewers or publications) are misleading people or customers 

yes, we can complain that current standards aren't useful or up to date but that's completely different story directed to different folks, those who sets the standards  

What is the line level RMS voltage standard set by industry for consumer use  ?  ... you can see it here.  , and you can measure any given CD using Audacity as I previously showed above  , that shows mediums attain no higher RMS  level.  ( seen in light blue ) , and mediums are rigidly  compliant to the requirements of the line level standard.  

 

Measuring 12.69 x greater, there is no justification or proper reason for it.  Rather it falls,  unless addressed by countering its stance,  into a disinformation strategy, which sadly until addressed, is misleading to anyone wanting information about how equipment actually performs 

 

 

 

 

 

 

 

[kukynas]

18 minutes ago, stereo coffee said:

What is the line level RMS voltage standard set by industry for consumer use  ?  ... you can see it here.  , and you can measure any given CD using Audacity as I previously showed above  , that shows mediums attain no higher RMS  level.  ( seen in light blue ) , and mediums are rigidly  compliant to the requirements of the line level standard.  

 

Measuring 12.69 x greater, there is no justification or proper reason for it.  Rather it falls,  unless addressed by countering its stance,  into a disinformation strategy, which sadly until addressed, is misleading to anyone wanting information about how equipment actually performs 

 

 

 

 

 

 

 

 

geeez, again your you line level output? has nothing to do with testing condition, have you ever heard about sustainability threshold used during testing? have you ever been or worked for company producing equipment or product? how they evaluate if the product sustain not only nominal (standard) value but also peak value what ever it is? do you think i.e. car tire is tested to sustain eg only 30 psi or several times higher to make sure it'll survive under every possible condition? or engine or high pressure bottle or or or.

that has nothing to do with your nominal (typical) values but values set for testing purposes by industry organisation and that's what we are trying to explain you, you can't compare your nominal (typical) value and it doesn't matter if it's in RMS or PTP or what ever but value used for testing purposes to evaluate DUT

I'm not going to comment it any further, please do not reply, if you wanna continue please open your own thread where people can talk about it, this thread is dedicated to NAD C658 not why is this or that standard wrong.

Thank you...

[Puresound]

16 hours ago, kukynas said:

- you've been right, it's S10U45S reverse rectifier fed by LM26670 (from the other side of the board, that magic bc I wasn't sure it's meaning) 

- 27mm 

- unless you are looking for some specific voicing (no idea which) to your sound I don't think you need luxury OPA, currently installed is OPA1612, previously been using LM4562 in other projects with same success, both low noise, low distortion, high PSRR and CMRR opamps   

Hi

 

Amazing how you can "decode" the board, so the LM2670 BC chip is underneath the D800 on the back of the board? So it just creates a virtual "-". Are you planning to bypass this and connect your own dual voltage LPS or "-" voltage LPS to see if it helps on the noise/improves sound?

Btw what does the NRC abreviation mean?

I have also had good experience with the LM4562 from the IVY III BALANCED LINESTAGE from TPA

[kukynas]

yes, it seems to be the case by the look of it when I follow the traces...
Nope, just one 14V main LPS via 12pin connector and keep the rest as is, don’t expecting any improvement to the distortion/sound just the PSU noise/spikes you can see prior to main 1khz tone

by NRC I meant NJW1194 volume chip as identified on the first page

I had those in the buffer stage of class D amplifier and was very happy with them

[kukynas]

so, bc is soldered back and original SMPS connected, I did test the distortion with the linear lab PSU before and couldn't see any difference so SMPS stays, no point of spending more money on it if there's no benefit, tomorrow will put it all back into its original shape and I can say I'm done with it...

 

 

did also final test with -10db signal which is still above of what music would normally represent and here's the result, all harmonics went down, third more than 20db, it's fair to say we shouldn't be too much worried about its distortion 

I would still recommend to stay away from spdif input and to stay below -2db on the dial due to sharp rise of distortion past this point, if anybody needs higher output (more volume) I would suggest more powerful amplifier instead

 

 

Howgh...

 

 

 

 

.

[Jventer]

@kukynas

Thanks a lot for the fantastic posts and all the work you did!

 

[Puresound]

12 hours ago, kukynas said:

so, bc is soldered back and original SMPS connected, I did test the distortion with the linear lab PSU before and couldn't see any difference so SMPS stays, no point of spending more money on it if there's no benefit, tomorrow will put it all back into its original shape and I can say I'm done with it...

 

 

did also final test with -10db signal which is still above of what music would normally represent and here's the result, all harmonics went down, third more than 20db, it's fair to say we shouldn't be too much worried about its distortion 

I would still recommend to stay away from spdif input and to stay below -2db on the dial due to sharp rise of distortion past this point, if anybody needs higher output (more volume) I would suggest more powerful amplifier instead

 

 

Howgh...

 

 

 

 

.

Hi Dan,

 

Very strange that there aren't any improvement but thanks for a huge job.

I still think there could be some component blocking besides the NRJ for getting lower distotion and noise.

Did you keep the OPA1612 in?

Did you hear a difference when the LPS was in compared to you current setup?

Did you measure the PSU noise at the OPAs and clock and DAC?

[kukynas]

Hi,

potentially yes, I think those switching bc’s leaking some noise into the power rails / ground but without having a chance to change all of them it’s hard to judge, it’s multilayer board so impossible to trace and measure everything...

yes, OPAs1612 are there and I’ll keep them there...

didn’t hear any difference between lab psu and SMPS...

nope, dsp board has to be attached to be able to measure the DAC board power rails due to fact dsp board is controlling on/off state of the psu rails and I can’t get to the dac board as there’s not enough space to reliably measure surface parts without potentially shorting their pins, even if I could my scope does’t go below 20uV which is still more than what those uln regs are capable of + those opas have pretty high PSRR and CMRR so might not be relevant