Universal buffer/headamp based on OPA1622

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Founder of XSA-Labs
Joined 2012
Paid Member
I was too quick to announce no pops :guilty: While playing with the boxed headamp I just realized that I hear a faint pop when turning it ON without the signal applied. Nothing ear punching but it is there. I guess before I only tried it with the signal applied, and indeed when signal is applied the pop is totally masked by the music. Interestingly, the pop is there even when the volume pot is at its lowest position. After a bit of thinking I guess I found an explanation. The headamp has around 1mVDC output offset, so I guess that is what I hear when I turn it ON: 0mV to 1mV step. On my first board I used OPA2140 which resulted in literally zero output offset voltage and I had no pop there as far as I remember. I asked a fellow diyer who has that board for testing to check it again to be sure. In this headamp I used OPA1642 which can have up to 3.5mVDC input offset voltage. I am probably lucky since output offset voltage of this headamp even with the gain of two is only around 1mVDC (my DMM has low accuracy in this range, so I would not take 1mVDC literally). I'll report here if the output offset origin of the pop is confirmed or not. Stay tuned!

If someone has better explanation for the origin of this pop please let me know.

Regards,
Oleg

Super build Oleg! I have a similar PSU design with DC-DC and TPS7AXXXX but I also add a Juma cap Mx to slowly ramp the voltage up over 10sec and that eliminates pop. It adds 3v dropout though but just size DC-DC bigger voltage.

https://www.diyaudio.com/forums/gro...led-class-hpa-sub-ppm-thd-18.html#post5507714

Then I installed that PSU into my desktop parallel OPA1688 buffer amp here:
Cheap as Chips OPA1688 Low-THD Muscle Amp

703206d1536815169-cheap-chips-opa1688-low-thd-muscle-amp-opa1688-2x-parallel-cased-01-jpg


Beautiful work though. I love how compact yours is.
 
I've followed your thread with interest, XRK. Great implementation!

As for the slow rump up of the supply to avoid switch on artifacts, I plan to use OPA1622 mute pin in the next iteration. At present the head-amp has a short turn on delay but it is insufficient for OPA1642 input bias to stabilize, thus a small offset voltage is still present at its output and creates a small pop in the headphones. I will add a simple circuitry to extend the un-mute to about 2s after power up and to instantly turn it off upon power off. I have an idea how to do it but unfortunately have no time to implement it at the moment.
 
I also plan to add similar size USB DAC board to make a complete USB powered solution. It can reside in the same small chassis and allow two mods of operation: fully analog and using digital input. So far I plan to use PCM2705CDB as the DAC but if someone can recommend similar but better specs solution I'm all ears:)
 
Member
Joined 2015
Paid Member
I've followed your thread with interest, XRK. Great implementation!

As for the slow rump up of the supply to avoid switch on artifacts, I plan to use OPA1622 mute pin in the next iteration. At present the head-amp has a short turn on delay but it is insufficient for OPA1642 input bias to stabilize, thus a small offset voltage is still present at its output and creates a small pop in the headphones. I will add a simple circuitry to extend the un-mute to about 2s after power up and to instantly turn it off upon power off. I have an idea how to do it but unfortunately have no time to implement it at the moment.


Hi Oleg,


I've implemented in my AmpCasq project the mute circuitry to behave the same way you described it. It may give you some ideas. It is build around a switch, a RC and a low drop diode.
Thread is there : AmpCasq: an OPA1622 integrated headamp project.


Nice idea to build a compact USB DAC/headamp : perfect for desktop use


Chris
 
Thanks Chris!

I noticed the RC network back than when you presented your circuit and only now I noticed that you discharge the cap to GND via 10 Ohm resistor and the switch. Simple solution! I plan to use an AND gate feeding raw power to one of its inputs and regulated power through an RC delay network to the other. This way RC network will control the turn-on delay and raw power presence the turn-off moment.

Oleg
 
Founder of XSA-Labs
Joined 2012
Paid Member
I've followed your thread with interest, XRK. Great implementation!

As for the slow rump up of the supply to avoid switch on artifacts, I plan to use OPA1622 mute pin in the next iteration. At present the head-amp has a short turn on delay but it is insufficient for OPA1642 input bias to stabilize, thus a small offset voltage is still present at its output and creates a small pop in the headphones. I will add a simple circuitry to extend the un-mute to about 2s after power up and to instantly turn it off upon power off. I have an idea how to do it but unfortunately have no time to implement it at the moment.

My parallel 8x OPA1622 is also using the mute pin connected to an RC delay as well. I have that PCB in hand but need to order qnty 9 OPA1622’s and the parts for the custom dual rail DC-DC supply.
 
Can I suggest to use a I2S isolated from Texas instruments?
What about cm6632a on different board and dac+amp on other?

Isolated I2S or better USB bus all together is my plan. I would also prefer to keep USB receiver and DAC together on the same PCB since sending I2S signal through a pair of connectors and a wire with hard to control characteristic impedance is asking for trouble which otherwise can be avoided. I am not an expert (yet :) ) in high speed circuit design but I do understand the effects involved and would not risk compromising performance due to lack of practical experience.
 
Founder of XSA-Labs
Joined 2012
Paid Member
I just did a little more reading on the HE-6, InnerFidelity tested it and said measured sensitivity is actually only 77.5dB. But apparently the HE-6 sounds fantastic when driven with a big amp for improved headroom. I have an HE-400i (much more sensitive) but same technology. A planar diaphragm with conductive traces stretched across a bunch of Nd magnets.

Headphones & portable audio - HIFIMAN.com
 
Member
Joined 2015
Paid Member
Isolated I2S or better USB bus all together is my plan. I would also prefer to keep USB receiver and DAC together on the same PCB since sending I2S signal through a pair of connectors and a wire with hard to control characteristic impedance is asking for trouble which otherwise can be avoided. I am not an expert (yet :) ) in high speed circuit design but I do understand the effects involved and would not risk compromising performance due to lack of practical experience.


TI's PCM2704 even if not high end ('only' 98dB DR and 48kHz max fs) integrate the USB interface and the DAC, it is available in SSOP package, allowing you to design a very compact board with OPA1622 headamp or line output.


Typical implementation could look like this : USB audio DAC with PCM2704
 
Last edited:
Hi Oleg,
had a look at your amp and it is beuatiful! Really sleek symmetrical pcb design that will give both channels exactly the same conditions. Like it!

Btw, I saw a few postings back that you had a member offering dScope measurements. Did you ever get it measured?
 
The measurements were done by another diyaudio member back then. I asked him to post his measurements and his evaluation of the head-amp instead of my (possibly wishful) interpretation but he never got the time for this, unfortunately. I still have his measurements data in the e-mails though.

Regards,
Oleg
 
Hi,

I finally got the time to post a complete set of measurements of my OPA1622 based head-amp presented in post #246 more than a year ago. With a kind permission from Joseph K, who has done all these measurements, I'll post his comments without alteration but shortened to the essential part where appropriate.

Oleg
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.