Yes I found the 1868 cards to be a bit 'poppy' too, great cards in general though. The 1688 cards I have tend to do it less I think. I have wondered in the past if a re-cap might help - I did a YMF719 card recently and feel like the boot pop was reduced but those cards don't pop much anyway.

From my experience, some ISA cards (not just ESS) tend to do this on faster systems, mostly under Win9x though. I rarely see it in pure DOS.

Changing the "I/O recovery time" parameter in the BIOS can sometimes help. But with certain cards, it seems like I need to put them in a slower, period-correct system to fully eliminate the pops. Basically, the further I moved away from Socket 7 the more chances increased that old ISA sound cards would pop.

badmojo wrote on 2022-01-21, 04:18:

Yes I found the 1868 cards to be a bit 'poppy' too, great cards in general though. The 1688 cards I have tend to do it less I think. I have wondered in the past if a re-cap might help - I did a YMF719 card recently and feel like the boot pop was reduced but those cards don't pop much anyway.

Yes, Im' fond of these cards too. I'm considering doing a recap on at least one of them and making some "before/after" recordings.

Joseph_Joestar wrote on 2022-01-21, 07:06:

From my experience, some ISA cards (not just ESS) tend to do this on faster systems, mostly under Win9x though. I rarely see it in pure DOS.

Changing the "I/O recovery time" parameter in the BIOS can sometimes help. But with certain cards, it seems like I need to put them in a slower, period-correct system to fully eliminate the pops. Basically, the further I moved away from Socket 7 the more chances increased that old ISA sound cards would pop.

I'm running a pure DOS 6.22 installation. Thanks for the tip about I/O recovery time, I shall see what setting do I have in BIOS.
I've also tried reducing CPU cpeeds and disabling caches - no effect.
The card itself resides on an ISA riser with an SBC, I've tried changing both the riser card and the distance between the soundcard and the SBC - also no effect.
Enabling\disabling internal soundcard amplifier changes the overall sound quality and +12v presence requirement, but the popping still persist.

Overall, the popping sounds don't bother me. I'm mostly curious what's causing this and could it be mitigated.

Well, regarding popping sounds - they are also present in DosBox, so I'm pretty confident now this is a "junk" DMA transfer issue on startup. It's not ES186x fault by any means.

P.S
And I'm also feeling incredibly stupid now. Having ordered some good caps to replace the old ones only to realize that all of them are bypassed on pcb when external amplifier is disabled.
And also all of the old "TBOR" brand caps are still in-spec regarding their capacitance and ESR, which is something worth praising, considering they are almost 25 years old.

UPD:
I've discovered one thing, which make huge difference in sound between ES1868 and ES1869: coupling caps on AOUT_L and AOUT_R pins of the chip.
All of my ES1868 cards use electrolytic ones, while ES1869 use ceramic.
I've removed right channel ceramic cap and replaced it with bipolar electrolytic one, then recorded split stereo track to test.
The difference is drastic. Ceramic one sounds high pitched and "flat".
Then I tried 3.3uF film cap (yep, it's huge). It performed almost the same as 10uF electrolytic.
After soldering 100nF film, performance was just a little bit better than 100nF factory ceramic.
I believe 100nF value of the default coupling ceramic cap is just too small for the impedance of line-in recording\headphones.
Also, bigger value cap suppresses popping sound volume a bit.

UPD2:
Replaced both 100nF ceramic caps with 10uF X5R ceramic ones (both are 0805 smd). I can confirm, that it is indeed a cap value that matters most.
When using line-out jumpers (disabling external amp), these coupling caps and the speakers\line-in impedance form a high-pass filter.