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.
Just learned how to do a single-shot measurements using all four channels of my DS1054Z. The transient is there. When unmuting there is a sudden DC offset on the output of around -30mVDC which decays to zero within approx. half a second. Interestingly that the mute releases the OPA1622's output when positive rail reaches +4VDC level which was exactly the design target. Despite the unmute delay there is still a transient that I can not explain. If someone has an idea what can cause such behaviour please comment!

The legend for the waveforms:
CH1 (yellow): +5VDC input.
CH2 (light blue): headamp output.
CH3 (pink): negative regulated supply rail
CH4 (dark blue): positive regulated supply rail

Regards,
Oleg
 

Attachments

  • TurnOnTransient1.png
    TurnOnTransient1.png
    41.8 KB · Views: 390
  • TurnOnTransient2.png
    TurnOnTransient2.png
    41.1 KB · Views: 386
The turn-on transient comes from the OPA1642 (see attached). The offset comes from the OPA1642 output before the OPA1622 unmutes. Interestingly, the OPA1622 does well when unmuting and just passes the DC offset further to the load. Next I'll try to replace the OPA1642 with the new part and see if it helps. If it doesn't, I'll probably try it with the OPA2140 instead.

The legend for the waveforms:
CH1 (yellow): +5VDC input.
CH2 (light blue): OPA1642 output.
CH3 (pink): negative regulated supply rail
CH4 (dark blue): positive regulated supply rail

Regards,
Oleg
 

Attachments

  • TurnOnTransient.OPA1642out.png
    TurnOnTransient.OPA1642out.png
    42 KB · Views: 379
Today I finally received the OPA2140 and tried it immediately in the headamp. The result is the same, the small pop is still there. I conclude that this family of op-amps simply has such a turn-on transient... or there is something else going on which I don't understand.

In principle it is possible to delay the unmute of OPA1622 further by changing D3 (LED) to a resistor and using a capacitor instead of R15. Since the internal impedance of the enable network of the OPA1622 is around 700k (see p.16 of the datasheet for details), resistor can probably be as high as 1M which in combination with 1u capacitor should give a long enough unmute delay.

Regards,
Oleg
 
Just tried replacing the unmute circuit (LED + resistor) by an RC network using 1M resistor and 1uF ceramic capacitor. The turn-on pop was nearly gone (unmute delay was roughly 0.5 sec or less) but a much louder turn-off transient noise appeared. So I reverted everything back to original. I guess the only solution in this case is using a more sophisticated muting circuit controlling the relay at the output of the headamp. Since the turn-on pop is not so disturbing I do not plan to change this design. If somebody is interested in having a PCB, I have some left. PM for details.

Regards,
Oleg
 
Thinking a bit more about possibilities to get rid of the pop I think XEN SHPP board can help here. The headamp board provides the taps for taking the power to the protection board offered by Patrick (EUVL) and the connecting from the headamp output to the output jack is anyways made with the jumpers, so can be easily routed through the protection board.

Also, if somebody knows about similar solution, or just simple circuit for power up unmute delay and instant mute upon power down, I would be happy to design a daughter board for my headamp which would fit its pinout for easy installation.

Regards,
Oleg
 
The GB is long closed and shipped.
You will have to start a new one at the same thread.
There is a minimum of 25 PCBs before we shall ship.

I admire your endless effort with the OPA1622.
For me it is not worth the while.
Maybe one day you should try some discrete Class A circuit, just to compare (subjectively).


Patrick
 
Today I finally received the OPA2140 and tried it immediately in the headamp. The result is the same, the small pop is still there. I conclude that this family of op-amps simply has such a turn-on transient... or there is something else going on which I don't understand.

In principle it is possible to delay the unmute of OPA1622 further by changing D3 (LED) to a resistor and using a capacitor instead of R15. Since the internal impedance of the enable network of the OPA1622 is around 700k (see p.16 of the datasheet for details), resistor can probably be as high as 1M which in combination with 1u capacitor should give a long enough unmute delay.

Regards,
Oleg
Have you read John Caldwell's and others paper on developing the opa1622?
Which of these 4 opamps for headphone amp/preamp?
Part5 is about start up and shut down pops.
 
Last edited:
Thanks for looking at my project, Patrick! The entire idea of using the OPA1622 was its "convenience" as an all-in-one solution which soon proved to be not so trivial to implement. Since there are tons of projects on this forum dealing with all kinds of amps topologies I took this route for myself to avoid interfering with the efforts of others. Since I am not skilled enough to come up with a better discrete design than already exists I use a pragmatic approach and use the best measurable spec op-amp solution.

Regards,
Oleg
 
Have you read John Caldwell's and others paper on developing the opa1622?
Which of these 4 opamps for headphone amp/preamp?

Yes Andrew, I did. The difficult part in using the OPA1622 is its enormous bias currents and sensitivity to input impedance and RF interferebce. The IC was meant to be used after the voltage output DAC in portable applications (the reason for small size of the IC). Trying to interface the OPA1622 with the volume pot, input RF filter and AC coupling cap proved troublesome which lead to adding a buffer/gain stage using the FET input op-amp... And so things started to add up... endlessly:) But I like the process of designing the whole thing and the sound I hear in the end.., except this tiny pop since I'm a perfectionist and it steps a bit on my nerves:)
 
I admire your endless effort with the OPA1622.
For me it is not worth the while.
Maybe one day you should try some discrete Class A circuit, just to compare (subjectively).


Patrick

Patrick, I don't know if a Cavalli amp satisfies your acceptance threshold..:)
But we compared and simply no contest, subjectively.
And let's forget about measurement-wise..
 
Last edited:
Yes Andrew, I did. The difficult part in using the OPA1622 is its enormous bias currents and sensitivity to input impedance and RF interferebce. The IC was meant to be used after the voltage output DAC in portable applications (the reason for small size of the IC). Trying to interface the OPA1622 with the volume pot, input RF filter and AC coupling cap proved troublesome which lead to adding a buffer/gain stage using the FET input op-amp... And so things started to add up... endlessly:) But I like the process of designing the whole thing and the sound I hear in the end.., except this tiny pop since I'm a perfectionist and it steps a bit on my nerves:)

I still debate about the decisions we made in the input stage of OPA1622, even though the chip has been very successful. Some compromises had to be made to allow for a wider input voltage range and low supply operation, and they resulted in higher input bias current. Improving EMI sensitivity on the other hand normally involves a trade-off with input voltage noise. There's always the next chip to make!

In the future I would also like to incorporate some of the technology we developed on the new OPA1692: OPA1692 SoundPlus Low-Power, Low-Noise, High-Performance Dual Bipolar-Input Audio Op Amp | TI.com into other products. We had to get clever to reduce distortion with only 650uA of power supply current and 5-6 MHz of bandwidth without totally giving up all stability with capacitive loads.
 
Wow! Thanks for the teaser, John! The OPA1692 looks really interesting. I am looking forward for the new parts you are designing. I hope to finish my current projects by the time you release them:)

The MSOP package version of the OPA1692 will release this week. SOIC will follow-on early next year. It was designed for low power applications, but the distortion cancellation technique we developed is applicable across many other applications. This coming year I'll be releasing some fun new parts! :D
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.