Dual OPA1688 Parallel hand-etched amp now boxed up with DC-DC +/-18v dual rail power supply from SHPA project.
Attachments
Last edited:
Thanks, guys! Casework, even with pre-drilled/cut cases like this still take an enormous amount of time. I had to drill the hole for the headphone jack in the front as this was a Class D speaker amp case. Plus all the holes on the baseplate. Then a cover plate for the 5.5mm barrel jack to replace IEC jack. I can’t imagine how much more time for a custom case.
This amp is surprisingly good sounding - I am almost surprised how good it sounds. It is very quiet. No noise whatsoever (hiss or hum) even with volume all the way up when source is not on. Having +/-18v rails really gives the OPA1688 some headroom for great dynamics.
The big discrete DC coupled SE Class A amps next to it sounds better of course, more power and more dynamics. But when I get the 8x parallel OPA1622 built I am not going to be so sure this will still be true.
I am going to have to get back to work on the 1622 project and order the BOM (thank you to Zman01 for helping me with this).
This amp is surprisingly good sounding - I am almost surprised how good it sounds. It is very quiet. No noise whatsoever (hiss or hum) even with volume all the way up when source is not on. Having +/-18v rails really gives the OPA1688 some headroom for great dynamics.
The big discrete DC coupled SE Class A amps next to it sounds better of course, more power and more dynamics. But when I get the 8x parallel OPA1622 built I am not going to be so sure this will still be true.
I am going to have to get back to work on the 1622 project and order the BOM (thank you to Zman01 for helping me with this).
Last edited:
How is your OPA1622 project going?
Will be interesting.
OPA1622 is a good opamp.
But it has fairly high input bias current.
I would recommend some input bias compensation to avoid DC-offset.
OPA1611, OPA1612 has much lower input bias current.
The JFET OPA1641, OPA1642 has no input bias problem.
I prefer JFET or CMOS operational amplifiers.
They have practically no input bias current.
Will be interesting.
OPA1622 is a good opamp.
But it has fairly high input bias current.
I would recommend some input bias compensation to avoid DC-offset.
OPA1611, OPA1612 has much lower input bias current.
The JFET OPA1641, OPA1642 has no input bias problem.
I prefer JFET or CMOS operational amplifiers.
They have practically no input bias current.
Hi Lineup,
I am a bit busy with work and a couple of other power amp projects don’t the 8x 1622 is on hold for the moment. I should at least order the parts so that if time presents itself I can start. I could even start with just a 2x or 4x version and add more as needed. Easy enough to warm the board from below and apply heat from a hot air pencil from above for more amp chips, as needed. I did not look at input bias current and perhaps you are right that a JFET input voltage gain stage would have been good. I had no issues with 1688 so assumed this one would be similar. But looking at data sheets now see 10pA vs 1uA so yes, a big difference. Thanks for the heads up.
I am a bit busy with work and a couple of other power amp projects don’t the 8x 1622 is on hold for the moment. I should at least order the parts so that if time presents itself I can start. I could even start with just a 2x or 4x version and add more as needed. Easy enough to warm the board from below and apply heat from a hot air pencil from above for more amp chips, as needed. I did not look at input bias current and perhaps you are right that a JFET input voltage gain stage would have been good. I had no issues with 1688 so assumed this one would be similar. But looking at data sheets now see 10pA vs 1uA so yes, a big difference. Thanks for the heads up.
Last edited:
Last edited:
The OPA1622 VSON10’s are in. Man, they are small (that’s a strip of 10). They are significantly less expensive from Arrow vs Mouser. Here is the board that they will go on, and I suppoe the corresponding tiny little VSON10 pads match up.
Attachments
Last edited:
I soldered one set of OPA1622's for a 2x parallel amp just to test and make sure it works before installing the remaining chips. Works great - at just 2x, it is more than enough for my 96dB OB-1 headphones. Using a linear 25w trafo and cap multiplier to get +/18v rails. Build uses Susumu and Dale metal thin film resistors on all key locations. JPS64's excellent layout is perfect, the amp fired up the first time, no issues. I was actually surprised it worked given that my soldering of the VSON10's was a bit sketchy. Those things really self-align surprisingly well via surface tension.
Attachments
Last edited:
I did not realize that the OPA1622 is the industry’s highest performing opamp?!
That’s quite a claim to be the top in terms of low distortion and high power.
https://www.prnewswire.com/news-rel...ce-audio-operational-amplifier-300192153.html
That’s quite a claim to be the top in terms of low distortion and high power.
https://www.prnewswire.com/news-rel...ce-audio-operational-amplifier-300192153.html
I don't understand the 220 ohm "summation" resistors you mention. There is no summation, there is no need for resistors at the input of a follower (they add a tiny bit of noise, that's all).
I think the Zobel network should be more like 22n/47R, otherwise its roll-off is too close to the audio band and affect the frequency response.
I think the Zobel network should be more like 22n/47R, otherwise its roll-off is too close to the audio band and affect the frequency response.
Measurements of Muscle Amp
The amp is powered by an Antek 25VA 15v transformer and a Prasi/MJ cap multiplier (Darlington topology) with 2200uF of bulk capacitance. I setup the amp to drive a 47ohm 2w resistor using a 2ppm 1khz oscillator. Voltage across dummy load is fed across the differential inputs of the Focusrite 2i4. I replaced the 100nF film cap in the Zobel with 22nF AVX MKT 100v film cap.
Setup:
Here is measured distortion profile for 1.0Vpp into 47ohms (0.00061% THD):
2.0Vpp into 47ohms (0.00082% THD):
and 4.0Vpp into 47ohms (0.0017% THD):
Settings for the FFT can be seen in the inset setup page.
Listening to the amp now with test tracks. Sounds very nice - neutral, clean and very quiet. Fluke 101 shows 0.0mV in mVAC. Output DC offset is -9mV on both channels. There is no scratchy noises when the volume pot is adjusted - smooth.
One final check, I connected it to a 4ohm speaker (dual PA130-8 XKi) and it sounds great, Qnty 8 of the OPA1622's in parallel should be able to drive a 4ohm speaker load since one can drive 32ohms.
The amp is powered by an Antek 25VA 15v transformer and a Prasi/MJ cap multiplier (Darlington topology) with 2200uF of bulk capacitance. I setup the amp to drive a 47ohm 2w resistor using a 2ppm 1khz oscillator. Voltage across dummy load is fed across the differential inputs of the Focusrite 2i4. I replaced the 100nF film cap in the Zobel with 22nF AVX MKT 100v film cap.
Setup:
Here is measured distortion profile for 1.0Vpp into 47ohms (0.00061% THD):
2.0Vpp into 47ohms (0.00082% THD):
and 4.0Vpp into 47ohms (0.0017% THD):
Settings for the FFT can be seen in the inset setup page.
Listening to the amp now with test tracks. Sounds very nice - neutral, clean and very quiet. Fluke 101 shows 0.0mV in mVAC. Output DC offset is -9mV on both channels. There is no scratchy noises when the volume pot is adjusted - smooth.
One final check, I connected it to a 4ohm speaker (dual PA130-8 XKi) and it sounds great, Qnty 8 of the OPA1622's in parallel should be able to drive a 4ohm speaker load since one can drive 32ohms.
Attachments
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.