Click the link to Andrews site in the first post.
Andrew - turns out my AD8599 is actually a single AD8597, so I took a gamble and ordered two OPA1656, they look good specs -
https://www.ti.com/lit/ds/symlink/o...universalsearch.tsp%3FsearchTerm%3DOPA1656IDR
This should work ok - its very similar to the OPA1642 iteration that I am assembling now.
(The OPA1642 has very good 1/f noise performance - just for reference)
Here are the measurements that I completed today with OPA1642 opamps. Conclusion: It works fine with them with an excellent set of measurements.
The HPA-1 has noe been tested and measured with LM4562 (best measurements), OPA1642 (slightly worse measurements - but see subjective impressions below) and the NE5532.
For the subjective tests, I used some CD classical music (New York Philharmonic playing Leonard Bernstein's 'West Side Story'), Fourplay and on vinyl I listened to an LSO recording on the 'Firebird Suite'.
I use Audio Technica ATH AD900 'Air' headphones that have particualrly good mid and high frequency sound - the bass is a bit 'dry' and not to everyone's tastes, but overall its the best headphone I've heard other than the Stax.
I only listened to the LM4562 and the OPA1642 - the other measurements were with one channel with an NE5532 and one with the LM4562 - so I won't comment on the NE5532 at this stage until I have a unit with both channels fitted with an NE5532,
My preference is for the OPA1642. Although the distortion is well below any hearing threshold, the sound seems more natural and free flowing. The bass also seemed to me to be a bit fuller.
The HPA-1 has noe been tested and measured with LM4562 (best measurements), OPA1642 (slightly worse measurements - but see subjective impressions below) and the NE5532.
For the subjective tests, I used some CD classical music (New York Philharmonic playing Leonard Bernstein's 'West Side Story'), Fourplay and on vinyl I listened to an LSO recording on the 'Firebird Suite'.
I use Audio Technica ATH AD900 'Air' headphones that have particualrly good mid and high frequency sound - the bass is a bit 'dry' and not to everyone's tastes, but overall its the best headphone I've heard other than the Stax.
I only listened to the LM4562 and the OPA1642 - the other measurements were with one channel with an NE5532 and one with the LM4562 - so I won't comment on the NE5532 at this stage until I have a unit with both channels fitted with an NE5532,
My preference is for the OPA1642. Although the distortion is well below any hearing threshold, the sound seems more natural and free flowing. The bass also seemed to me to be a bit fuller.
Last edited:
Nice design and layout, Andrew. If my present hpa project falls short, I'll definitely come back for one.
My impressions are the same. They're both excellent, but the 1642 sound differs just as you described.
FWIW, a few months ago I tested (along with some other duals) 15-20 of each, procured at different times, for I/V converters with a diff-output DAC. Some of the 4562's were more than 10 years old. Besides low distortion, the desirable spec is low and flat CMR vs freq on both amps. Some well-regarded opamps did not fare well. Of all the ones tested, the LM4562 and OPA1642 were the most consistent unit-to-unit, with CMR hovering around -100dB and flat across the audio band, and within a dB or two channel-to-channel.
My impressions are the same. They're both excellent, but the 1642 sound differs just as you described.
FWIW, a few months ago I tested (along with some other duals) 15-20 of each, procured at different times, for I/V converters with a diff-output DAC. Some of the 4562's were more than 10 years old. Besides low distortion, the desirable spec is low and flat CMR vs freq on both amps. Some well-regarded opamps did not fare well. Of all the ones tested, the LM4562 and OPA1642 were the most consistent unit-to-unit, with CMR hovering around -100dB and flat across the audio band, and within a dB or two channel-to-channel.
NB - please note that in the pdf presentation in post #31 above, some of the '20 dB Attenuator is OFF' labels got misplaced. A corrected version of the presentation is here:-
http://hifisonix.com/wordpress/wp-content/uploads/2021/03/X-Altra-HPA-1-with-OPA1642-Opamps-1.pdf
http://hifisonix.com/wordpress/wp-content/uploads/2021/03/X-Altra-HPA-1-with-OPA1642-Opamps-1.pdf
I just tested 3 versions of my balanced/SE to balanced and SE buffer with OPA1656, OPA1642, and LM4562 side by side. The OPA1656 and 1642 both sound superb. Very engaging and bass was very nice. The LM4562 sounded good but I think was just not as engaging. A bit “dry”. However, if you did not compare side by side, the LM4562 sounds fine and one would not know what one was missing. It’s a subtle effect. Maybe the difference between BJT and JFET/MOSFET technology?
The main reason I did this was that I had specd my buffer to use OPA1656, which are becoming increasingly hard to find in stock. Whereas the 1642 is plentiful and a little less expensive.
Attachments
That's well-put.
Maybe. It just bugs me when I hear a clear difference that I can't quantify by any measurement.
This subject has been thrashed to death on this forum over the years. On a well run DBT, most folks will not be able to tell the difference between the 4562 and the 1642 - certainly I don’t think I could.
However, in casual listening, people prefer one over the other and we should just accept that.
However, in casual listening, people prefer one over the other and we should just accept that.