Yes, they form a virtual ground for the ground and enable pins.
I remove two resistors at the top and put 0805 or 1206 capacitors instead (10-22uF) and add extra resistor at the bottom from 10k one to ground to form a voltage divider for soft start, 68k on +/-9V rails.
Currently have 3 of those in my DAP
> I think you're underestimating the real cost of product development.
You are telling someone with 35 years semicon equipment experience, both front- & backend ?
> Have you actually looked at the specs for the OPA1622? Or played with the evaluation kit?
Have you read post 225, 247, 255, 270, 297 ?
Cheers,
Patrick
You are telling someone with 35 years semicon equipment experience, both front- & backend ?
> Have you actually looked at the specs for the OPA1622? Or played with the evaluation kit?
Have you read post 225, 247, 255, 270, 297 ?
Cheers,
Patrick
good idea! so you have them connected to DAP ground now,Yes, they form a virtual ground for the ground and enable pins.
I remove two resistors at the top and put 0805 or 1206 capacitors instead (10-22uF) and add extra resistor at the bottom from 10k one to ground to form a voltage divider for soft start, 68k on +/-9V rails.
Currently have 3 of those in my DAP
did you notice any change from that?
these things are great, for around $8 including opa1622, adapter, gold plated copper pins (possibly brass, they are non-magnetic at least) plus soldering that most would not be able to do at home
Yes, they are awesome as little building blocks, I probably bought around half a dozen of these by now. And yes, connected to ground.
No, I did not notice change as I never listened to them in their "stock" configuration.
Will post photo later, although it's very guerrilla DIY
Looks good, so if this is a Digital Audio Player
There, 3 OPA1622 in AK4497 based DAP.
the input would be audio files from a storage device and this will do all processing to analogue?
Ah so there is another board for SD card, display, controls, makes sense now.
I finished connecting a couple OPA1622 chips to ground and did some comparison to the remaining ones, its certain the voltage divider is not helping performance. I can understand now why the OPA1622 is a good choice for driving headphones directly.
Absolutely!
Especially low-impedance ones, high supply current + low output impedance make it one of the best in class, if not the best.
I have a pair of DIY single-BA driver IEMs, based on Knowles RAB-32257, they sound absolutely atrocious with any source where output protection resistors are utilised, but with OPA1622 they transform into very clean, even, detailed and even have decent bass.
Unfortunately manufacturers could be somewhat conservative and offer aged but tested (and sometimes cheaper) solutions, I wish there was less lag in adoption of something like OPA1622, luckily I can DIY retrofit it into designs that allow for it
Thanks for OPA1622 @johnc124, much appreciated!
I found TI is preparing OPA828 release, a successor of OPA627 and OPA827.
It mustn't be cheap price!
OPA828 36 V, High-Precision, Low-Noise, Low Bias Current, JFET-Input 50 MHz OP AMP | TI.com
It mustn't be cheap price!
OPA828 36 V, High-Precision, Low-Noise, Low Bias Current, JFET-Input 50 MHz OP AMP | TI.com
I wonder if this is John C's swan song he was talking about? Looks like a very nice part.
4.3 nV/√Hz at 1 kHz is very bad for an audio op amp. Moreover FET based op amps are not good for audio.
Last edited:
Your first assertion can easily be argued as what is acceptable noise depends on the particular application. You second assertion is stated as a truism with so substantiation; in fact many excellent audio circuits have been built with FET-input op-amps. It's not clear what you mean by "FET-based".
Nope! I was somewhat closely involved with the OPA828 development in its early stages but I wasn't the lead for that product.
For now, you'll just have to be patient...
Isn't your beloved NE5532 5 nV/rtHz? And given the bias requirements of a BJT-input opamp vs FET-based, the current noise (and ultimate input referred noise) of your NE5532 is going to be much higher (for any reasonable feedback resistor).
4.3 nV/rtHz for a FET-input opamp is down near the lowest noise for a monolithic opamp. Yes, there's plenty of lower e_n BJT-input, but you have to mind your p's and q's in terms of current noise, too.
Ever hear of current noise and source impedance? Looks like Daniel beat me to it.
This part would much better after a potentiometer, for example.
Last edited: