BUF03's
As you said "Turn it up". I managed to build one of the AD844 stacks with 3 SOIC's. Still need to do the second one yet. The Ex is supposed to be out of here in 2 weeks. I hope to give the 3 stack a whirl shortly after.... Dave
Hi George, I really like the SQ of the BUF03. To bad they are obsolete. My preamp/headamp allows me to listen at stupid levels.
As you said "Turn it up". I managed to build one of the AD844 stacks with 3 SOIC's. Still need to do the second one yet. The Ex is supposed to be out of here in 2 weeks. I hope to give the 3 stack a whirl shortly after.... Dave
Hi George, I really like the SQ of the BUF03. To bad they are obsolete. My preamp/headamp allows me to listen at stupid levels.
I think you may be in for a real shock.
If the difference in SQ between the single and double was because of current starvation with the low output of the PCM1704 (+-1.2mA)
Then your TDA1541 with more output than that will definately want at least 3 stacked or even 4 to counter the current starvation, if that's what is was for the massive difference we got in SQ.
Maybe it just came down to the load the dac was running into being halved by stacking from 30ohm to 15ohm?
I don't know all I can say is this is the best sounding cdp or dac I have heard so far for redbook cd replay, and I get to hear some exotic stuff.
Cheers George
Last edited:
George, try a single OTA like OPA861 which can be run up to over 12mA and then add an external buffer after that - no need to do any stacking at all and the 861 will be equivalent to four (+?) stacked 844s.
Also allows you to choose the best buffer you can find.
Or use OPA-860 with its own buffer, alas the buffer is closed-loop, but uses current feedback and is somewhat better than voltage opamps used a buffers. It would make for an interesting comparison for those willing to try it - and it may possible, who knows, sound better than 5 stacked 844s.
Keep in mind -/+ 5V only, but once output is nulled, will swing 8V p/p - or 2.8V RMS - so gain needs to be set to give 2V RMS @ 0dbFS.
It works - I am listening to it right now (@ 2.2V RMS output), but not with all current absorbed by "-" input as you have to with 1704 DAC - using it differentially from a Sabre DAC.
Link
Cheers, Joe
Huh? I had two 15V rails and regulated down to 5V rails and no way I use the digital supply.
Noticed there was no official thread so I created:
http://www.diyaudio.com/forums/digital-source/240987-opa861-i-v.html#post3600147
hi, after searching for days for a suitable i/v for my 1541 dac and i think i found my suitor 
please tell me tho, gents, how would you go about setting 844s up for differential dac? i'm talking 2 tda1541 with +/- L & R signals. i'd like to convert it to unbalanced with transformers, would i be able to omit the output cap from george's initial schematic?
please tell me tho, gents, how would you go about setting 844s up for differential dac? i'm talking 2 tda1541 with +/- L & R signals. i'd like to convert it to unbalanced with transformers, would i be able to omit the output cap from george's initial schematic?
I have not done dual dif dac setup with the stacked AD844's, maybe someone else would like to chime in here.
As for the output cap, I don't use one, I am dc coupled from dac output right through to the speakers.
Unless your talking about the small pf cap across the TZ resistor? As this is just a first order HF filter cap, and yes it can be omited but you will get a bit more hf rubish on the output, which is so high it's not heard, but it could worry preceeding amps amps that do not like hf rubish on their inputs.
Cheers George
As for the output cap, I don't use one, I am dc coupled from dac output right through to the speakers.
Unless your talking about the small pf cap across the TZ resistor? As this is just a first order HF filter cap, and yes it can be omited but you will get a bit more hf rubish on the output, which is so high it's not heard, but it could worry preceeding amps amps that do not like hf rubish on their inputs.
Cheers George
Just thought I'd post this quote on using just the AD844's I/V stage with no feedback from Walt Jung up for all to read again, and this was without stacking them as we have found now is clearly better than just using a single.
Walt Jung
PS: On the 844, I once used it as an I/V, and the biggest limitation was/is the output stage. Sounds better by going around it, but then of course one is only using half the chip. Drive a clean buffer from the 844 TZ pin 5, and it is a whole different matter.
Jocko Homo's reply to this quote from Walt Jung
When used in the open-loop manner, you have basically created the I/V stage I made, but in chip form.
Cheers George
Walt Jung
PS: On the 844, I once used it as an I/V, and the biggest limitation was/is the output stage. Sounds better by going around it, but then of course one is only using half the chip. Drive a clean buffer from the 844 TZ pin 5, and it is a whole different matter.
Jocko Homo's reply to this quote from Walt Jung
When used in the open-loop manner, you have basically created the I/V stage I made, but in chip form.
Cheers George
Last edited:
DC offset is nil from the 844 itself even stacked, but if your dac gives out dc then this is what will be present at TZ pin 5 also.
Cheers George
So 2ma injection for 1541 then 4sets of stacked 844 to 1:1 trafo... sounds like a charm
All working?? Tell us how it sounds?
Wow!!! does this mean you have 4 sets of 4 stacks for 4 x TDA1541 in stereo dual dif setup??? 16 x AD844's that's got to go down as some kind of i/v record!
Cheers George
haha sorry george, huge misunderstanding.
im still doing the artwork for the pcb for it but my dac will have all salas shunt regs for 2 tda1541 and around 8 ad844. i had to stop the drawing to figure out what i/v i wanted but now i can finish.
1 dac will be fed inverted data for -L,-R output and each out will have at least 2 ad844 stacked accompanying it. so (L,R,-L,-R) x stacked ad844s. i'd like to stack more than two but not sure if salas shunts could supply that. my recent experience with stacked 1543 had me finding out that salas is only good up to 9 chips then the voltage sags.
1543 is rated at 60ma max.
my new dac isn't anything special (cept to me) but it's best of what i know all thrown in together.
i wouldn't mind doing 16 stacked 844s after all i was set to do 20 stacked 1543 in differential until i realized salas shunt wasn't meant for that. i think that setup would've been some kind of record, too
(p.s. the benefit of paralleling 1543 is not that noticeable in differential setup as compared to the same number of them in single end mode)
im still doing the artwork for the pcb for it but my dac will have all salas shunt regs for 2 tda1541 and around 8 ad844. i had to stop the drawing to figure out what i/v i wanted but now i can finish.
1 dac will be fed inverted data for -L,-R output and each out will have at least 2 ad844 stacked accompanying it. so (L,R,-L,-R) x stacked ad844s. i'd like to stack more than two but not sure if salas shunts could supply that. my recent experience with stacked 1543 had me finding out that salas is only good up to 9 chips then the voltage sags.
1543 is rated at 60ma max.
my new dac isn't anything special (cept to me) but it's best of what i know all thrown in together.
i wouldn't mind doing 16 stacked 844s
after all i was set to do 20 stacked 1543 in differential until i realized salas shunt wasn't meant for that. i think that setup would've been some kind of record, too (p.s. the benefit of paralleling 1543 is not that noticeable in differential setup as compared to the same number of them in single end mode)
Just ordered 16 AD844 from RS Component for future experiments. This is what I am going to modify. This unit is originally produced for car entertainment but it has potensial for indoor use. We will see. Will take a while. Have a number of AD797 which I had problems with using as a buffer at unity gain. I was not able to prevent it from oscillating no matter what I did. May be I can find a place for it here.
Last edited: