AD844 60mhz 2000V/?s for I/V stage duties after the dac convertor.

[niss_man]

@georgehifi

Hi George

I have just been starting to design a dac (1st attempt) using AD1862 r2r dac chips , PMD100 filter and CS8414 receiver. Are you able to check out the output iv/buffer in my schematic and tell me if there is anything wrong with it as I know you have had great experience with this particular buffer? I have seen a capacitor on pin 2 of ad844 to ground (input from dac) and have not got it in my diagram and don't know whether it is needed, Also I don't know whether I have all parts for op627 buffer correct and whether values are ok. The AD1862 outputs 1ma max.

 

Cheers

 

Simon

 

 

 

Update....

Edited December 19, 2016 by niss_man
extra detail. No trawling

[georgehifi]

Quote

I have seen a capacitor on pin 2 of ad844 to ground (input from dac) and have not got it in my diagram and don't know whether it is needed, Also I don't know whether I have all parts for op627 buffer correct and whether values are ok.

 

All looks fine, I never used any cap on pin 2 of the AD844 triple stack, the only thing I can see is you need to ground (to analogue ground) pin 3 of all the AD844's.

When this is done you could power up. (Ooops you did ground pin 3 didn't look to the left).

 

Once powered check what dc offset you have at the TZ pin if more than a few mV's you should use the offset null pins 1 and 8 only on one of the AD844's use a 20kohm pot not 20ohm (it's a data sheet misprint) then zero the offset at TZ.

 

Then for good measure you can use the dc offset pins 1and 5 on the OPA627 buffer to null out any further dc offset on the output rca's.

 

If you have a scope look for any oscillations at TZ and the output rca's.

 

BTW I have never used the AD844 with a AD1862 dac, only with PCM1702 or 1704's.

Mick Melony of Supratek may have used it with an AD1862 as he was very active on my DiyAudio thread about the AD844 triple stack, using it with other dacs, like the TDA1541, as well and giving good praise, I think he's under Statman there.  

 

Cheers George

 

http://www.analog.com/media/en/technical-documentation/data-sheets/AD844.pdf

 

http://www.ti.com/lit/ds/symlink/opa627.pdf

 

Edited December 16, 2016 by georgehifi

[niss_man]

Thanks George

 

The AD1862 circuit has a trim circuit included on pins vsa, trim and adj. Do you think this would be good enough to use as the triming for dc offset or would it need the other trimming circuitry as you have suggested?

 

 

 

[Addicted to music]

FYI and just my opinion.

 

I think the OPA 627 is far over rated, over priced because of DIY forums.   

My preference is for the updated version: OPA 827 a SOIC version, I find this version a better update and better in SQ than the 627.   Even the OPA 1612 is cheaper nad is better value.  There are other ADI and LME that are IMO way better than any 627 or in that fact better than any OPA. 

[georgehifi]

5 minutes ago, niss_man said:

Thanks George

 

The AD1862 circuit has a trim circuit included on pins vsa, trim and adj. Do you think this would be good enough to use as the triming for dc offset or would it need the other trimming circuitry as you have suggested?

 

yes by all means, if the 1862 dac has dc offset trim for it's output, start there, but the AD844 may have it's own as well, as the OPA627 may have.

 

Always best to be dc coupled from dac convertor output all the way to the speakers if you can do it, as all coupling caps sound different because they all colour the music in some way, BEST CAP IS NO CAP!!!!

 

Cheers George  

[niss_man]

Ok. I have added 20k pot to ad844 as below and populate the board with it if the offset adjustment at ad1862 isn't enough and leave it out if the ad1862 adjustment does the trick.

 

[niss_man]

19 minutes ago, Addicted to music said:

FYI and just my opinion.

 

I think the OPA 627 is far over rated, over priced because of DIY forums.   

My preference is for the updated version: OPA 827 a SOIC version, I find this version a better update and better in SQ than the 627.   Even the OPA 1612 is cheaper nad is better value.  There are other ADI and LME that are IMO way better than any 627 or in that fact better than any OPA. 

My opa627 will be socketed, so if I want to opamp change I can. I like the lm4562 personally. That is a dual opamp though. It seems LME49710 is it's single equivalent.

[georgehifi]

Quote

6 minutes ago, niss_man said:

Ok. I have added 20k pot to ad844 as below and populate the board with it if the offset adjustment at ad1862 isn't enough and leave it out if the ad1862 adjustment does the trick.

 

 

Just use the 1862 to nulls it's own, and use the 844 to null it's own, and also the OP627 to do itself. Better a little on all than a lot on one.

I found the 627 the best of all, for socket use, the BUF03 was the most dynamic but it was a touch dirty sounding, and the 627 came closest dynamically to it, it sounds just a touch dark, but I'd rather that than a touch bright.

 

BTW only get the OPA627 for RS or Element14, as anywhere else your bound to get fakes, the 627 is the most fake'd opamp out there.

 

Cheers George  

[Addicted to music]

23 hours ago, niss_man said:

My opa627 will be socketed, so if I want to opamp change I can. I like the lm4562 personally. That is a dual opamp though. It seems LME49710 is it's single equivalent.

 

You can use an adaptor via Browndog for SOIC. 

The adaptors also comes in single and 2 version of double. 

The OPA 827 is a modern version of the 627, .  I got the OPA 827 free via a sample program from Texas Instrament by creating an account, and then selecting the "sample" icon.  There are a limited numbers they send to you, and usually arrive at your door in 4days, and trackable; all free.

All OPA are safe for any audio as they are so heavily compensated unlike the LME or ADI, therefore very stable. 

[Nada]

On 17/12/2016 at 9:48 AM, Addicted to music said:

....  other ADI and LME that are IMO way better than any 627 ....

 

What do you recommend for a socketted OPA627 direct drop in replacement then?

[georgehifi]

13 minutes ago, Nada said:

 

What do you recommend for a socketted OPA627 direct drop in replacement then?

From what I've listened to the OPA627 is the best, (I have not heard the 827) for DIL socket. But as I said before there are so many fakes of these, you must make sure it's genuine, as it got some mediocre reviews because of the fakes, some just being 50c TLO71 re-baged to be OPA627 and the listener/user is none the wiser. Art Vandelay told me he got caught out with fakes, and that was from WES Components (Wagner), who are reputable, so their buying division got suckered with the fakes.

 

Cheers George  

Edited December 18, 2016 by georgehifi

[Addicted to music]

32 minutes ago, Nada said:

 

What do you recommend for a socketted OPA627 direct drop in replacement then?

 

The LME 497210/OPA 627 would be a good candidate.  That's compensated too.  You find that many manufacturers are using the LME. An extremely good representation.  Not as dynamic as the OPA 627, but then who cares, the LME are very resolving, microdetails are in spades where with the 627 you need to listen hard to find those same details and sometimes it's not there!  The very reason you asked and the answer is directed with the point of stability.  Not all opamps are the same.  If you go ADI such as the AD826/826, LME49990 or the extremely detailed AD4627-1BRZ you are going to need a CRO to watch for stability.   

The other thing with op amp rolling I've also notice IMO that with jfet inputs the representation is grainy compared to to BJT input.  I've also notice this on discrete components too and you'll find in Headfi there is the same mention of this findings observation.

And if you're wondering all my opamps are direct from E14 or TI as samples.  I've never bought any on eBay.  The LME are direct from a specific supplier who had were agents for National Electronics b4 it was bought out by TI.

Edited December 18, 2016 by Addicted to music

[georgehifi]

Just remember, if you don't know how yours is configured for gain, only use opamps that are "unity gain stable" you can find this in the pdf data sheets on all opamps, just use the pdf search function for "unity gain stable" as the pdf's can tend to be 20> pages long

 

Cheers George

Edited December 18, 2016 by georgehifi

[Addicted to music]

8 minutes ago, georgehifi said:

Just remember, if you don't know how yours is configured for gain, only use opamps that are "unity gain stable" you can find this in the pdf data sheets on all opamps, just use the pdf search function for "unity gain stable" as the pdf's can tend to be 20> pages long

 

Cheers George

 

And that is the issue with opamp rolling, the data sheet must be followed like the dot on top of the I!  

Not just unity gain is so important, the supply voltages are all different.  They very, when I was active in a Headfi, some were rolling the opamp in the wrong supplied voltage.  Others were using a OPA 637 in a unity gain circuit!  Also the supply bypassed capacitors must be used as instructed so not all will be a assumed blind roll in the dark!

[niss_man]

I think you meant lme49710 not 497210 adicteded to music. Both lme49710 and opa627 are unity gain stable and have up to +/- 17v supply ability. I'll get some lme's from rs. I do enjoy their Lm4562 brother and will be good to compare to the opa627's.

[georgehifi]

BTW Anyone who wants to try their own triple stacked AD844 I/V stage and various buffers after it.

 

I started a DIY for it over at DiyAudio, quite popular with some heavy hitters chiming in, be prepared it's large over 1800post 95 pages and 1/4 million views.

 

It gets into the much better sound of double then triple stacking that I did of the AD844 around half way through.

Our Mick Melony (Supratek) (Statman on the site) even tried quad stack 844's on his TDA1541 converter.

 http://www.diyaudio.com/forums/digital-line-level/227677-using-ad844-i-v.html

 

Cheers George

 

 

Edited December 18, 2016 by georgehifi

[niss_man]

I was originally looking at that thread but found it heavy going as you have to trawl through hundreds of posts just to find the useful stuff. If I post the 2 different types of Ad844 conversion, one with buffer and one without are you able to attach those diagrams to your first post to help others out?

[georgehifi]

8 minutes ago, niss_man said:

I was originally looking at that thread but found it heavy going as you have to trawl through hundreds of posts just to find the useful stuff. If I post the 2 different types of Ad844 conversion, one with buffer and one without are you able to attach those diagrams to your first post to help others out?

 

I'll see what I have on file, if not you'll have to trawl, as I tried many different things, with the same PCM1704 dac based (Cary 303/200) hdcd CDP.

 

The best I found was triple stack AD844 (not using it's own buffer output) taking the signal from the TZ of the AD844 to a separate (genuine) OPA627 buffer. I'll post up if I find something.

 

Cheers George   

[georgehifi]

Found this for you.

 

Cheers George

[niss_man]

I was thinking more on the lines of below. It has extra detail for guys like myself to find useful. If you are happy with it @georgehifi you could post it on the other forum as an (edit to your 1st post) addition.

Cheers

 

SImon

 

[georgehifi]

By all means Simon post it up on DiyAudio after this guys post who also used multiple stack AD844 with OPA627 output buffer.

http://www.diyaudio.com/forums/digital-line-level/227677-using-ad844-i-v-96.html#post4922426

 

Cheers George

Edited December 19, 2016 by georgehifi

[Nada]

George - your thread on DIYaudio is great but overwhelming. Why dont you think about a quick edit to your first post to make theDIY  thread usuable? It would be great to write a synopsis and include nis_man diagram.

[georgehifi]

18 hours ago, niss_man said:

I was thinking more on the lines of below. It has extra detail for guys like myself to find useful. If you are happy with it @georgehifi you could post it on the other forum as an (edit to your 1st post) addition.

Cheers

 

SImon

 

 

Just realized, that the AD1862 also put out around 1mA, and you used 2 x stacked AD844's. You should up this to 3 stack, as there was a marked improvement again with 3 of them with the PCM1704, as it will lower the input impedance as see by the 1864 to around 10ohms which is better for a current output dac and gives it more current headroom as well.

I'll post this also on you post at DiyAudio.

Cheers George

[Nada]

@georgehifi

 

The AD1862 datasheet says it can swing  "OUTPUT CURRENT Bipolar Range ±1 mA" at 2.1kOhms output impedance.  Thats 2mA peak to peak isnt it?

 

The PCM1704 does ±1.2 mA ie 20% more but has only half the output impedance at 1kOhm and you like 3 stacked AD844?

 

What does that indicate in terms of stacking AD844 for:

 

1. a single AD1862 versus

 

2.  stacked double AD1862 (per channel) putting out 2mAp-p @ I guess an effective 1.05kOhms shared output impedance?
 

Edited December 21, 2016 by Nada

[georgehifi]

Stacking is the AD844's

 

The PCM1704 is also 1.2mA p to p, it can current saturate a single 844, two was much better and three sounded the best, 4 did nothing over three, also the input impedance of a single 844 is around 40ohms not bad, 3 stack brings it down to 13ohms much better loading for a current output dac, yet not too low to kill it off too much and have to make it back again with the TZ resistor to ground, which is 2.7kohm (for gain) and 470pf across it to make a low pass filter at -3db at 125khz.

 

As for a TDA1541 Mick from Supratek had to use 4 or even 6 stacked 844's because the TD1541 gave out so much I believe 4mA at 0dbf

 

Cheers George  

Edited December 22, 2016 by georgehifi