Collating Favourite OPAs To Replace X-Dac V3 5532

[XTRProf]

Hi all,

I know there are many flavour of the month OPAs. However, I'm sure there are a few that really stir you to have fond memory of them and will unheasitatingly recommend them to all. I'm in the process of replacing the 5532 in the X-Dac V3. So shoot your direct replacement of 5532 recommendations here and the reasons why those OPAs hit you like a brick. Of course, not something that fainted you but something that keep you awake into the wee hours of the night gasping at the beauty and unbelievable sound of music coming out from your system with these OPAs used. Well, something like tube rolling and there are favourites.

Thanks all for your insights.

[leo]

Depends which position they are used in, one type may perform in say the I/V better than the output buffer etc

I'd fit some sockets, order a few different samples and compare them yourself to see which you like best
I could recommend one type and somebody else may not like it

Alex has already posted his favourites used in the X-dac V3, remember that a few types are only available in surface mount packages so would need to be soldered onto a SOIC to DIL adapter

[XTRProf]

Hi Leo,

Thanks for your insights. I'm only trying to gather the favourites from all perpsectives. So whether I like it or not when fitted doesn't come mind and I don't hold grudges against anybody for recommending such and such as this is what forum is all about and that is to share. This is a hobby of trial and error entering into subjectivity. Being objective will be best but ........... ;D Of course, I have Alex favourites in mind. BTW, the 6 8 pins OPAs I'm talking about is as shown in Alex picture here:





A better description will be here:









Also, Alex had given a block diagram of one of the audio channel if I'm not wrong here:







So shoout your favourites OPAs here. I'm sure there are many lurking around here at rockgrotto very interested as well. In the spirit of Christmas, let us share and give "presents".



Thanks.

[XTRProf]

Oh, ya ........, after the marathon 3 hours operation of the X-Dac V3 into the wee hours of yesternight to solder 6 DIP8 sockets into the dac, I have this to show ............



Updated 2nd January 2008. Now clearer.


Sorry, for the blur image as this was taken with a lot of HP cam shake after that 3 marathon hours on the job. I promise I will upload a better picture when I open up the X-Dac V3 again to OPA roll. Now the dac is dead with no OPAs in in.

Ya, all 5532 were KIA with broken legs and I can tell you this is horrendous work even for a skill solderer like me. This is because the PCBA is double sided with solder on top and bottom which made the job very very tedious without harming the PCBA traces. The good thing is very little damage was done to the PCBA traces at the expense of all the 5532 KIAs. But now I can leave that nightmare work behind me to have some fun with OPA rolling with some help from this forum members.

TIA for all contributions.


WARNING: Only do what I had done if only you are skill and confident enough or else you will be left with many broken traces on the PCBA and a resultant dead X-Dac V3.

[leo]

I like THS4032 if using classic op-amp I/V, also for differential to single ended operation like whats done in the first stage of the V3.
Unfortunately this op-amp isn't a straight swap, you MUST use good regulation and layout, if not it tends to oscillate sounding bright and harsh.

LM6272 is one of my favourites too but like the above it needs to used with care, I like this one better than the newer LM4562, also if you use upsampling the sound can have too much excitement in the highs so again its all down to implementation

Theres the popular OPA2132, quite smooth and good for taming brightness, these are also easier to implement, I personally didn't like them in my dac but tbh theres not many fet input op-amps I do like, AD8066 are quite good, better bass than the AD8620

Some people like OPA2107 and AD826 in dacs/cdp's

Be very careful with current feedback op-amps, if caps are used across the feedback resistors it can cause brown trousers

[Deleted]

"Ya, all 5532 were KIA with broken legs and I can tell you this is horrendous work even for a skill solderer like me. This is because the PCBA is double sided with solder on top and bottom which made the job very very tedious without harming the PCBA traces"

XTRProf
I am glad you mentioned how careful you need to be. Another
way is to use a sharp pair of side cutters and cut the leads of the old NE5532 ICs close to their bodies, then carefully remove the rest of the pins.
All that tedious work does leave you a little shaky doiesn't it ? ;D
If you have difficulty locating 2n7 polypropylene caps, it may be easier to use a 1n2 + 1n5 combination, with one of them under the PCB. One of my friends did this, and results are virtually indistiguishable. BTW, I didn't draw out the circuit, another Sydney based friend did this while modding his X-DAC V3.
AFAIK it is correct. He also transferred the whole X-DAC and new dual regulated PSU to a 1 unit rack case.
Alex

[leo]

Evox PFR are nice small polypropylene caps that work well in filtering etc, RS used to sell them, RS site is currently down or I'd try and find you the part numbers

[XTRProf]

leo said:

I like THS4032 if using classic op-amp I/V, also for differential to single ended operation like whats done in the first stage of the V3.
Unfortunately this op-amp isn't a straight swap, you MUST use good regulation and layout, if not it tends to oscillate sounding bright and harsh.

LM6272 is one of my favourites too but like the above it needs to used with care, I like this one better than the newer LM4562, also if you use upsampling the sound can have too much excitement in the highs so again its all down to implementation

Theres the popular OPA2132, quite smooth and good for taming brightness, these are also easier to implement, I personally didn't like them in my dac but tbh theres not many fet input op-amps I do like, AD8066 are quite good, better bass than the AD8620

Some people like OPA2107 and AD826 in dacs/cdp's

Be very careful with current feedback op-amps, if caps are used across the feedback resistors it can cause brown trousers

Gee, thanks, Leo. Now I have more leads to look at datasheets and to experiment. Superduber, man!

[XTRProf]

sandyk said:

"

All that tedious work does leave you a little shaky doiesn't it ? ;D

You bet! See how shaky my picture taken was when compared to previous. So all would be wannabes in this, please do heed Alex and my advice as we don't want to see your X-Dac V3 traceless and dead.

[leo]

xtrprof said:

leo said:

I like THS4032 if using classic op-amp I/V, also for differential to single ended operation like whats done in the first stage of the V3.
Unfortunately this op-amp isn't a straight swap, you MUST use good regulation and layout, if not it tends to oscillate sounding bright and harsh.

LM6272 is one of my favourites too but like the above it needs to used with care, I like this one better than the newer LM4562, also if you use upsampling the sound can have too much excitement in the highs so again its all down to implementation

Theres the popular OPA2132, quite smooth and good for taming brightness, these are also easier to implement, I personally didn't like them in my dac but tbh theres not many fet input op-amps I do like, AD8066 are quite good, better bass than the AD8620

Some people like OPA2107 and AD826 in dacs/cdp's

Be very careful with current feedback op-amps, if caps are used across the feedback resistors it can cause brown trousers

Gee, thanks, Leo. Now I have more leads to look at datasheets and to experiment. Superduber, man!

No probs, just experiment and compare, getting the right balance to match the rest of your set up may take time.
Sorting out the regulation and decoupling should also give you a wider choice to try out, its surprising how easy it is to knock the balance right out

[Deleted]

Don't forget that without nearby regulation, as with those extra 78L15 and 79L15, some of the wider bandwidth devices may be unstable.
Sandyk

[XTRProf]

All all gurus,

Just a noob question, what do those high bandwidth devices oscillate with poor power regulation. I though if there is enough juice like current and voltage, whether regulated or unregulated, for those devices, the devices will have sufficient to "eat" and not "beg" thru oscillation for more juice?

Also, is the AD8066 a unity gain device or not as I just have brief look at the specs and I said to myself wow in most except whether unity gain or not. Wow, in terms of 145MHz bandwidth, slew rate and slightly lower noise floor. Is the AD8066 an all gain stability device?

Please enlightened, gurus.

[Deleted]

XTRProf
Unfortunately just brief answers. All high speed devices need a good low impedance supply to remain stable. With devices like the AD8066, the PSU supply pins are also regarded as inputs, so they need careful bypassing. There are several Application Notes on this subject, but this old PC that I am currently using doesn't have them filed.
AD8066 is a variable gain device, and as you can imagine, it also makes an excellent video amplifier when properly used. It is unity gain stable in the non invertying mode. As far as I am aware it is also unity gain stable in inverting mode as well. It is of course, stable at higher gains. Because it is also capable of video use in a normal gain of 2 configuration, i.e. 75R output series resistor, and 75R termination at the input, it is also quite a reasonable buffer for normal audio use too.
Alex

[XTRProf]

sandyk said:

XTRProf
Unfortunately just brief answers. All high speed devices need a good low impedance supply to remain stable. With devices like the AD8066, the PSU supply pins are also regarded as inputs, so they need careful bypassing. There are several Application Notes on this subject, but this old PC that I am currently using doesn't have them filed.
AD8066 is a variable gain device, and as you can imagine, it also makes an excellent video amplifier when properly used. It is unity gain stable in the non invertying mode. As far as I am aware it is also unity gain stable in inverting mode as well. It is of course, stable at higher gains. Because it is also capable of video use in a normal gain of 2 configuration, i.e. 75R output series resistor, and 75R termination at the input, it is also quite a reasonable buffer for normal audio use too.
Alex

Hi Alex,

Oh, this is already good enough for all noobs like me. If you are to write a PhD dissertation on the subject, I think we will be in cloud 9 to the power of infinity literally. ;D Anyway, a dissertation will be better use for Australia's University of Melbourne, University of New South Wales, etc and MIT, right? I will try to find the application notes to understand it more technically. This is getting more interesting as the days go by as the technical side in me is forcing me to go from here to infinity to ebb the dark cloud for answers. Now I understand, because of the high bandwith demand even the bloody pin (or wire) becomes critical. Btw, what is the critical bandwidth to be considered as having to go the extra mile to take care of the pins and wires consideration as well thru proper regulation? 50 Mhz? I can imply below 30Mhz, from the OPA specs and recommended usage, in the X-Dac V3 to be of no concerned.

As for the AD8066, I saw some graphs in the specs showing G=1 to 5 (I think), so this roughly corrolates to implications of what I understand as variable gain stable from G=1 to 5. So with your affirmation, the AD8066 can be deemed as unity gain stable although they didn't state that in their write up. Ya, at video frequencies, most will use the industry std of 75 ohms unbalanced and 110 ohms AES or XLR/balanced std.

So the AD8066 can only be used as an audio buffer or differential conversion to unbalanced single ended? What about being used upstream as well? What I mean, can we also used them at all the 6 places in the X-Dac V3?

TIA for all ray of light karmas.

[XTRProf]

Ya, forgotten to wish all here at rockgrotto HAPPY ............... NEW .............. YEAR ............ 2008. We are now 1 year closer to the ............ Opps, shouldn't say such bad and bizzare thingy on the first day of new year. Well, let you guys and gals figure out the missing link.

[Deleted]

" What I mean, can we also used them at all the 6 places in the X-Dac V3?"

Ever heard of too much of a good thing ? ;D
Sorry, still too busy to say more.
Perhaps someone else will care to elaborate ?
SandyK

[leo]

AD8066's like the AD8020, LM4562 etc are very detailed, if you have too many in the chain it can be overkill making things sound upfront robbing the emotion.
I'd use the fast detailed ones for the I/V as these tend to be the most critical and then try various flavours for the others, just be sure to sort out the regulation or it will more than likely oscillate causing you problems not forgetting sound shite!

[XTRProf]

sandyk said:

" What I mean, can we also used them at all the 6 places in the X-Dac V3?"

Ever heard of too much of a good thing ? ;D

Yup, like eating too much turkey during Christmas and New Year days, our stomach will turn turkey as well. ;D

[XTRProf]

leo said:

AD8066's like the AD8020, LM4562 etc are very detailed, if you have too many in the chain it can be overkill making things sound upfront robbing the emotion.
I'd use the fast detailed ones for the I/V as these tend to be the most critical and then try various flavours for the others, just be sure to sort out the regulation or it will more than likely oscillate causing you problems not forgetting sound shite!

Total agreement. I will never never have something that's without emotion. Then it's machine music. I like something wraith like that will make my goose pimples work. At least I know that I'm still human and not man machine like what Kraftwerk used to sing here:





That's what Hifi and AV is all about. I will used the more clinical ones upfront and, perhaps, the BB OPA 2354 at the buffer. That will be more or less correct. BTW, I have the 2354 onboard now





and it sound overall damn good and superior without having even to delve in a mind boggling AB. Now the dac is firmly in Stereophile Class A until I can get the air, imaging, presence and prat sorted out better thru more OPAs trial. The 5532 is better in that aspect but much poorer overall than 2354. The 2354 got the volume, soundstage width and emotion very right whereas the 5532 can sound machine music in comparison. This 2354 is very very good for vocal, classical and some jazz music but stop short of rock elements which it tends to slow down the tempo and the prat.

Next stop will be 4562. ETA this week.

[XTRProf]

Leo,

After looking at your avatar cat pic, that cat of yours reminded me of a similar one with the same colour and pattern that I had 3 to 4 years ago. I had named it TOM but it had eloped with perhaps a feline partner ..................

[leo]

Don't think I've ever tried OPA2354, its probably the only op-amp out there I haven't tried
If I had a V3 then I'd join in the fun posting the results, this dac is similar to my Diy PCM1794 one

Thats our cat, his name is Busoms, I'll let you work out why we called him that

[XTRProf]

leo said:

Don't think I've ever tried OPA2354, its probably the only op-amp out there I haven't tried
If I had a V3 then I'd join in the fun posting the results, this dac is similar to my Diy PCM1794 one

Thats our cat, his name is Busoms, I'll let you work out why we called him that

Wow, you have dac made with 1794? Actually, what's the difference between 1792, as used in the X-Dac V3, and the 1794. I know 1701 or 1702 is just PCM 16bits 44Khz decoder whereas the 1792 is with 24 bits 192Khz and DSD decoding as well if we know how to use the latter.

Your name of Busoms for your cat really crack me. That name imply this literally




Oops ............ Sorry ladies, that pic is not for real as it's just a breast plate as taken from here


www.costumes4purim.com/Merchant2/merchant.mvc?page=PC/PROD/Bosom_breast_plate_316601FR


Fyi, you can buy it. So I don't think I had offended anyone here as it's probably synthetic. LOL. Hmmm, why do they need that for in the first place?


So a really wild guess is this he cat of yours is a female feline loving kind. "Womaniser or Feliniser" type I had named my former cat TOM is in line with the Tom and Jerry cartoons of yesteryears. That cat of mine was really naughty and fun loving type. So TOM would be most appropriate.

[leo]

PCM1794 is one of TI's top dacs, it has quite a high current output compared to others so you can easily use passive I/V keeping the value of the resistors lower whilst still getting a fairly high output.
Some of the other dacs in the range require to be used with a micro-controller for internal register configuration that needs programing

Regarding the feline, he acts like a silly tit ,couldn't really call him tit so decided on Busoms

[XTRProf]

After having the 4562 in the dac for about a week, I should say I love these chips for pratty and fast tempo music. Vocal, classical and jazz also not too bad but I would say it's not as good as the 2354 in this aspects. Overall, most music sounds ok with the 4562. Things I like about the 4562 are:

1) Presence and air.
2) Details.
3) Just right for pratty and fast tempo music. The 2354 has no match here.
4) Vocal, jazz and classical still no too bad.



Conclusion: It's an overall chip good at all types of music. I love it better over the 2354.

[Deleted]

XTRProf
Now that you have come this far, why not go to the next stage and supply it with regulated +-15 or 16VDC ?
Don't forget that the analogue section still has a mediocre unregulated supply.
Alex