-
WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!
ESS THD ‘Hump’ Investigation
- Thread starter jackenhack
- Start date
The examples are for the ESS9018K2M. Is it the same for the 9038Q2M?
I offer to cure the hump as much as I can on my Khadas Tone board, and send it to someone for measurement.
But I need a member to do the math for me first, based on the Khadas´ schematics,
to find out what values have to be changed, if any.
Any advice on a crystal oscillator for tweaking is also welcomed. I have read in this forum that a certain 50mHz(?)
oscillator has better results but limits the upper sampling frequency...?
About the TI opamps proposed:
There is no nominal voltage given for the OPA1611/1612 but it looks their lower treshold is 4.25V.
So close to the Khadas 5V supply voltage.
According to the Khadas schematics the pins of their (unbranded) I/V Opamps RT6862D are compatible with OPA1612.
So besides the luck of being a pin-compatible replacement the OPA1612 will probaly introduce a problem in needing
their own -/+15V supply to perform better... Just a guess, I might be wrong here
But it leads to the question:
Could a weak supply voltage for I/V conversion have a negative influence on the hump?
Last edited:
They are not on the ESS website. So the manufacturer will probably someone else. But it would be interesting to measure if an
highly integrated DAC that is using the ES9038 with an ESStech audio amplifier like the ES9602 shows the hump as well...
But anyway, could anyone check the math explained by HououinKyoma on the Khadas board´s schematic?
(value of external resistors in the I/V stage)
https://www.audiosciencereview.com/forum/index.php?threads/ess-thd-‘hump’-investigation.5752/page-14#post-161200
Maybe we are lucky and there is a relatively quick fix...
Based on my very limited circuit knowledge, do you mean we need to adject R40 and R53?
IVX
Major Contributor
Sorry if offtopic, finally I made cycling of the signed 16-bit variables(int16_t second or int16_t third depends on which pin of GPIO pulled down to GND) in the range +512/-512. So the tuning looks like that: you shorting by tweezers two pads, marked "2" either "3", on the PCB, and particular variable starts cycling online(no any clicks). You are or AP script, whatever, keep watching for a minimal THD, and when the min value of THD is reached, short these two pads again to stop the cycling. In my case, the 2nd harmonic wasn't necessary to adjust becuase it is low enough. If your THD without compensation is less than .005%, the scanning range +/-100 could be ok.
IVX
Major Contributor
johan, I have no idea why there 250kHz but it is bad idea to me. AK4490 is much cleaner about out-band noise and probably low order and high Fcut LPF is ok but ES9038Q2M has a lot oh HF noise yet. I didn't notice if 2.2-4.7nF(0603 C0G) to GND and between DAC outputs affects THD performance at all, so I keep LPF 15db/oct at 90kHz.
So R0402 is correct...?
If my calculation is correct, the needed virtual ground voltage is 1.65*(806/(806+774))=0.84171V, and the current value is 3.3*3.3/(10+3.3)=0.8188V. To make it right, you can either change R40 from 3.3k to 3.424k or change R38 from 10k to 9.638k. The closest 1% resistor to 3.424k is 3.48k & for 9.638k is 9.53k. The other option is paralleling a 267k 1% resistor to 10k to make it 9.639k. Just my 2c.
@HououinKyoma
How precisely should the virtual ground match the recommended value? Within 1%? The tone board already have it within 3% and still have THD hump. And Rs (Rdac) of ES9038Q2M is specified to have 11% tolerance, so even if correct resistors value are installed individual board still can be far off your recommended value unless calibration on individual board is performed. Seriously I think this is a chip design issue from ESS's side to have such a big tolerance and also have critical design parameters based on this very loose value.
Last edited:
0402 is the size spec, r36 or r37 is the real name of resistor
For info over at MARCH audia dac1 review thread (aka Khadas TB) we can se Amir did probe LP roll off for CD rate and think it looks do a pretty good LP job especially in comparison to some other modern DAC's that looks more or less as a pass through. Guess IV stage LP filter you talk about there is of 1st order roll off and can tell it wont matter much to the final LP curve from Amir if that 1st order LP is set at 100kHz or 250kHz. Below is a traced copy of that dac1 curve with cursor set around 22,05kHz and i tried compensate curve a cascade of what could be called a 1st order LP filter Linkwitz Transform up at 250kHz and then a 1st order LP at 100kHz and on scales >100dB its really impossible to see any difference.
IVX
Major Contributor
BYRTT, I talking about out-band noise, which is really higher than AK4490, I saw there spread-spectrum noise with the peak at 10MHz or so and 3 spikes 50, 75 and 100MHz. When I try to see by my scope out-band noise for AK4490, I see nothing at all. To be honest, I feel that SE9038Q2M is highly overpriced, AK4490 for 1/3 money makes the same THD+N performance(.0003% and easy to get, I seriously doubt if SE9038Q2M can steady show .0001% without THD_COMP cheating), I see only regarding the S/N SE9038Q2M has the advantage(and for footprint size + SPDIF input) but do you really care if your DAC S/N -121dbA either -126dbA?
If you want to roll opamp, you may try opa1642 + ad8397 which comes from apogee element, I believe apogee did a lot of testing on this.
Looks like this is the way to go. This will be my first SMD opamps to roll.
brt8888 thanks for doing the math. 3.48k / 0.5% for 3.424k and 9.65k / 0.5% for 9.638k are available as metal film resistors
in the R402 footprint from digikey - delivery time is 10 weeks
Similar threads
- Poll
