Hi all. Recently a quantity of leftover Chinese stuff was dumped on the market here and there was 2015 - 2019 produced audio stuff in the lot. I got hold of a few SMSL M8A DACs at a nice price for friends. Beautiful and very small solid aluminium milled casing, XMOS Xu308 and a ES9028Q2M with 3 x OPA1612 for filtering and one of those marvels of modern technology an SMPS wall wart. Except for the switcher nothing can go wrong I thought. After connecting the others and myself found the bass to be very light and the general sound character was so so. Even my old Subbu V3 was way better. I hadn't bothered with recent ESS DACs and was looking after every corner for the horrendous ESS hump and all the nasty stuff people tell about ESS DAC chips that are supposedly acoustical hell. Scary stuff and no TDA chip nearby to cure the ΔΣ pain.
How come? Layout looks OK and part quality seems good too. Most reviews in English were very positive about this very DAC. Did we receive duds or fake ones? Not likely as they came from the distributor and were in original packaging. Are we spoiled and used to extreme good audio?! Also not likely. Are the reviews questionable? Mmmm...they are!
We found the device not delivering what we expected so I examined all details. I replaced an opamp, replaced the 3 pieces 470 µF 16V Panasonic FJ (old series) for FR etc. Still it did not perform. Is it that darn switcher SMSL uses for creating +/- 12V?! No. Then I looked a bit deeper and the bug & error hunting paid off. The regulator for the 1.25V for the ES9028Q2M DAC core was not so stable sometimes going to 0.9V. It is a LM317M in SOT-223 and it gets its 3.3V input voltage from a 3.3V pre-regulator.
Wait a minute... LM317M putting out 1.25V so ADJ pin to GND and 3.3V input voltage?!? Right there lays the issue. A recent LM317M needs about 2.5V differential voltage at 500 mA and this is an absolute minimum. All datasheet values (depends on brand) are specified at 3V. Many will now tell me that the dropout voltage depends on the output current and it sure does. But ... the core current apparently is not a stable value and it sometimes fluctuates to even a few hundred mA. Then the dropout voltage is between 1.7 and 2.5V strongly depending on brand. Let's say it is running more than 200 mA then dropout voltage is already 2.1V with the LM317M as used in my DAC. 2.1 + 1.25 = 3.35V... while it has a stable 3.3V offered at its input. Won't work out OK.
I replaced it for a AMS1117-ADJ with only 1.3V dropout voltage at its maximum 800 mA. Bingo!!! The DAC at once performed like it should with solid bass and very good sound stage. I'll save you the "veil lifted" and "blacker than black" talk as in the reviews but the device is totally different from the old situation. An amazingly large difference.
Nothing found on the web about this bug so I'll publish it here so you can mod your M8A too. So get your M8A and do the less than 1 Euro costing modification. You won't recognize it the second you will use it again. Highly recommended.
On the picture it is the SOT-223 regulator left of the "M8A" marking. It is already the AMS1117, I forgot to take a picture when replacing the LM317M.
How come? Layout looks OK and part quality seems good too. Most reviews in English were very positive about this very DAC. Did we receive duds or fake ones? Not likely as they came from the distributor and were in original packaging. Are we spoiled and used to extreme good audio?! Also not likely. Are the reviews questionable? Mmmm...they are!
We found the device not delivering what we expected so I examined all details. I replaced an opamp, replaced the 3 pieces 470 µF 16V Panasonic FJ (old series) for FR etc. Still it did not perform. Is it that darn switcher SMSL uses for creating +/- 12V?! No. Then I looked a bit deeper and the bug & error hunting paid off. The regulator for the 1.25V for the ES9028Q2M DAC core was not so stable sometimes going to 0.9V. It is a LM317M in SOT-223 and it gets its 3.3V input voltage from a 3.3V pre-regulator.
Wait a minute... LM317M putting out 1.25V so ADJ pin to GND and 3.3V input voltage?!? Right there lays the issue. A recent LM317M needs about 2.5V differential voltage at 500 mA and this is an absolute minimum. All datasheet values (depends on brand) are specified at 3V. Many will now tell me that the dropout voltage depends on the output current and it sure does. But ... the core current apparently is not a stable value and it sometimes fluctuates to even a few hundred mA. Then the dropout voltage is between 1.7 and 2.5V strongly depending on brand. Let's say it is running more than 200 mA then dropout voltage is already 2.1V with the LM317M as used in my DAC. 2.1 + 1.25 = 3.35V... while it has a stable 3.3V offered at its input. Won't work out OK.
I replaced it for a AMS1117-ADJ with only 1.3V dropout voltage at its maximum 800 mA. Bingo!!! The DAC at once performed like it should with solid bass and very good sound stage. I'll save you the "veil lifted" and "blacker than black" talk as in the reviews but the device is totally different from the old situation. An amazingly large difference.
Nothing found on the web about this bug so I'll publish it here so you can mod your M8A too. So get your M8A and do the less than 1 Euro costing modification. You won't recognize it the second you will use it again. Highly recommended.
On the picture it is the SOT-223 regulator left of the "M8A" marking. It is already the AMS1117, I forgot to take a picture when replacing the LM317M.
Attachments
Last edited:
How old is that product? I own since recent 2 pcs of SU-6 DACs. I'm getting worried. They sound really good with plenty bass - but if they are on the edge... perhaps they could be even better....
//
//
For the M8A with ES9028Q2M 2017 - 2018 I think and a third version also called M8A with ES9038Q2M followed. The Revision 1 PCB of the one in the picture is made in December 2016. Don't worry just check it. All the SMSL DACs with DC input I now know are based on more or less the same design (the +/- 12V switcher for example) but there is not always the need for a 1.25V regulator with other DAC chips.
Just have a look how things are done in that SU-6.
Just have a look how things are done in that SU-6.
Last edited:
OK, to avoid talking into air (or to myself) so last post: I now modified 3 x M8A all with the same error and all 3 show the same stunning difference after replacing the regulator so I take this as a series error. The probably means all M8A with ES9028Q2M are affected. AFAIK the M8A with ES9038 does not show this behaviour.
