Not with the 844 I used the same supply rails that was for the original opamp in the CAL CL15 which was nothing special.
Cheers George
As for OPA861, you will not only need decoupling caps on both rails, but also another one across Vcc and Vee. I used perfboard for my I/V and it turns out it will oscillate without it. 0.1uf~0.22uf NP0/C0G will do just fine.
Poting,
nice to see you back here
does anyone used opa861 to work with differential output dacs (or doubled single ended dacs)
Is It correct that when both + and - signals got summed we can rid off dc offset that comes from not only from opa861 but also from dac, for example pcm1794 which Is 6.2 mA
nice to see you back here
does anyone used opa861 to work with differential output dacs (or doubled single ended dacs)
Is It correct that when both + and - signals got summed we can rid off dc offset that comes from not only from opa861 but also from dac, for example pcm1794 which Is 6.2 mA
Last edited:
hi samoloko,
I've never thought about going balanced. If DC offset is your concern, you may implement a servo to overcome this problem.
Poting
As 99.99% of this topic seem to be about anything else but ad844 as i/v...and i don't have the time to read everything , could somebody point me to ad844 behavior when used as classical i/v (virtual ground with feedback) circuit?
There's something i don't really understand: why would the people use the ad844 without a feedback with an additional buffer when ad844 with feedback has plenty of output power to drive alone very tough loads.
The datasheet says that ad844 is designed for video cables drive(50 ohms) and that it operates linearly at +-6v which i think is good enough by any standard for 1vrms output.
I myself found that i prefer to stick with 0.5...1v rms output for i/v instead of the standard 2vrms and that is for many reasons.Instead of having lower SNR because of the transimpedance i/v operation, i found out that actually i get better SNR, because i can use lower noise op amps , even much cheaper ones, after the i/v, with a certain 2x...4x gain which also provides better PSRR and CMRR than most op amps used in unity gain stages.
Strangely enough , after so many years since the 2v RMS output standard was set in cd players and v-out dac outputs i find that that was a deeply wrong decision as rail to rail op-amps weren't good enough back then, nor the unity gain stable op-amps at high frequencies.Now they are pretty decent, but expensive or designed to work with very low or too low rail supplies.
I think that most of the "high end " techniques and components use should be reevaluated if you start using the classical transimpedance vitual ground i/v stage with a 0.5...1v rms output, but most design engineers would automatically assume that, if i would use +-6v or 5v usb than that is for they should be allowed to show off in front of their chief designers with a the cheaper rail to rail c-mos op amps and here is where i deeply hate engineering skills.You give them a chance to do it right and they screw it again so that hordes of pseudo-scientists would reinvent the i/v stage with the most expensive op-amps on the planet.
There's something i don't really understand: why would the people use the ad844 without a feedback with an additional buffer when ad844 with feedback has plenty of output power to drive alone very tough loads.
The datasheet says that ad844 is designed for video cables drive(50 ohms) and that it operates linearly at +-6v which i think is good enough by any standard for 1vrms output.
I myself found that i prefer to stick with 0.5...1v rms output for i/v instead of the standard 2vrms and that is for many reasons.Instead of having lower SNR because of the transimpedance i/v operation, i found out that actually i get better SNR, because i can use lower noise op amps , even much cheaper ones, after the i/v, with a certain 2x...4x gain which also provides better PSRR and CMRR than most op amps used in unity gain stages.
Strangely enough , after so many years since the 2v RMS output standard was set in cd players and v-out dac outputs i find that that was a deeply wrong decision as rail to rail op-amps weren't good enough back then, nor the unity gain stable op-amps at high frequencies.Now they are pretty decent, but expensive or designed to work with very low or too low rail supplies.
I think that most of the "high end " techniques and components use should be reevaluated if you start using the classical transimpedance vitual ground i/v stage with a 0.5...1v rms output, but most design engineers would automatically assume that, if i would use +-6v or 5v usb than that is for they should be allowed to show off in front of their chief designers with a the cheaper rail to rail c-mos op amps and here is where i deeply hate engineering skills.You give them a chance to do it right and they screw it again so that hordes of pseudo-scientists would reinvent the i/v stage with the most expensive op-amps on the planet.
Last edited:
I think if you search for Pedja Rogics answer to that you'll find your answer.
Without feedback the 844 is more stable I believe to glitches from the dac. But you need three of them (stacked) to sound the best using a PCM1704 so there's no current starvation. And the output buffer was easily beaten for sound by an OPA627.
Cheers George
Although a friend of mine is able to swear on Pedja 844 common base circuit i simply don't believe their reasons.They all use these i/v stages with old dacs without oversampling .First strike: oversampling has its own reasons. AD844 has HUGE, i mean HUGE slew rate and it should be FREE of any problems due to those glitches if there are any...I have a TDA1541 based Sony player with no low pass filter at all and it seems that Nakamichi has a copy of that design , which again has no low pass filter at all and is using the same chip set as the SONY. They simply don't need any damn low pass filter that is how bad those glitches are...Even so...the ad844 should handle them easily and a passive filter should wipe them out easily.
I used an LM6172 after the tda1541 as virtual earth trnasimpedance i/v stage and it sings as good as possible with all his noise added.If there are any other artifacts that i don't like , i'm absolutely sure that the dac itself is producing them and that is how i understand the use of that 50 ohms input impedance of the ad844 as white noise generator while the diamond buffer is there to cover everything with harmonic content, so they are simply using a masking effect.The musical content itself may be poorly mastered and can constitute as a reason for a lot of diy decisions to mod digital equipment. That can be done in many ways.If anybody can argue with real measurements that I'm wrong i have nothing against it , but I'm mostly interested in a really honest and professional view on ad844 used with feedback.I simply don't trust anybody telling me that ad844 can't handle any glitches of an audio dac. AD844 has a full power bandwidth of 20mhz at 20v p-p.
I used an LM6172 after the tda1541 as virtual earth trnasimpedance i/v stage and it sings as good as possible with all his noise added.If there are any other artifacts that i don't like , i'm absolutely sure that the dac itself is producing them and that is how i understand the use of that 50 ohms input impedance of the ad844 as white noise generator while the diamond buffer is there to cover everything with harmonic content, so they are simply using a masking effect.The musical content itself may be poorly mastered and can constitute as a reason for a lot of diy decisions to mod digital equipment. That can be done in many ways.If anybody can argue with real measurements that I'm wrong i have nothing against it , but I'm mostly interested in a really honest and professional view on ad844 used with feedback.I simply don't trust anybody telling me that ad844 can't handle any glitches of an audio dac. AD844 has a full power bandwidth of 20mhz at 20v p-p.
Last edited:
Go way back into these pages, someone did extensive measurements and screen shots of it, used as single and multi stacked, then come back.
Cheers George
there are lots of interesting trials to read about and i could loose myself in this subject
They weren't trials, but proper measurements done with analysers ect.
It's the only way you'll answer your own queeries, by spending time and having a look.
Cheers George
This is probably a very remedial question but. I am putting together an AD844 with TDA1541a. When I adjust the offset pot (the one with the 2SK170) to minimise the offset on pin 2 of the AD844 I can only get it down to about 9mV is that ok?
Also when i minimise the offset on pin2 (as described above the offset measured on pin5 of the AD844 goes up to about 4-500mV. I can reduce the offset on pin5 by adjusting the pot to about 5-10mV but then I have about 30mV on pin2.
Do I have something completely wrong? If not what is the best compromise setting for best sound?
Thanks
Ian
Also when i minimise the offset on pin2 (as described above the offset measured on pin5 of the AD844 goes up to about 4-500mV. I can reduce the offset on pin5 by adjusting the pot to about 5-10mV but then I have about 30mV on pin2.
Do I have something completely wrong? If not what is the best compromise setting for best sound?
Thanks
Ian