ES9038Q2M Board

In either case, it clearly seems like your I/V isn't the culprit in the rather mediocre performance but the 9038 board, seeing as Victor is still getting better all-in performance (~-115 dB on the 3rd harmonic at 0 dBfs).

Still nice knowing these guys perform better in current mode. Also, the measurements are appreciated!
 
Member
Joined 2009
Paid Member
I think that ESS generally recommends current mode, without being specific about the differences.

The main source of noise in the original circuit seems to be the relatively high resistor values in the filter/buffer. So the dynamic range of the original design gets nowhere near the performance possible with the DAC chip.

Perhaps I should try the current mode with e.g. the LME49720 to see if it will be better than the LME49990 for this.
 
It might be, but at this level, it's pretty much impossible to tell outside of empirically. Paradoxically with your 1.26 K feedback network, I'm getting the 49720 as the lower noise option, albeit very slightly (3.56 vs 3.66). (If I made a bonehead about the current noise contribution, do correct me)

I don't have the respective current outputs of the 9018k2m vs the 9038q2m either. I'm assuming they're the same, in which case, you have 1.5x the gain.
 
I did not get it..is current or voltage mode just pending on the output circuit following the DAC output or do I need to set it somehow in the controller for the DAC chip?


JensH, the 2 resistors I see on the picture in your report (thanks for publishing it!)are a V-devider for the the 1,65V to feed to the yellow wire?
 
Member
Joined 2009
Paid Member
Yes, current mode or voltage mode is determined by the circuit connected to the output of the DAC. Some DAC's are intended for current mode, e.g. some of the high-end converters from TI and some DAC's are intended for voltage mode, e.g. converters from AKM and Cirrus Logic. The ESS DAC's can typically be used in either current mode or voltage mode.

In current mode the output is held at a fixed voltage, typically provided by an op-amp with the inverting input directly connected to the output of the DAC (one op-amp for each polarity). The non-inverting input is connected to a reference voltage, in this case 1.65 V (half of the 3.3 V supply). The feedback will make sure that the DAC output is held at the same voltage as the reference voltage. The current through the DAC output will vary depending on the audio signal applied to the DAC.

In voltage mode a high impedance filter/buffer stage is connected to the output of the DAC. So the DAC output will have some voltage swing in this case.

Yes, I used two resistors and a tantalum capacitor to make a 1.65 V reference.
 
I own this NE5532 replacement: HDAM Double differential circuit Fully symmetrical circuit Double op amp replace NE5532 LME49720 MUS02 OPA2604-in Amplifier from Consumer Electronics on Aliexpress.com | Alibaba Group

I am waiting for the delivery of my power supply and then I want to test the board with this replacement.

@freezebox: Please can you tell me (us) which DC blocking caps are to be bridged and which resistors are to be removed. The board revision is VR1.04.

Thanks!

Hi, do you own the old or the new board?
 
Hi Jensh.
Interesting measurements.
Can you show a schematic of your output stage in current mode?
Voltage mode was original like on board?
Thank you.

Pretty sure the I/V was just the classic inverting opamp with a Rf of 1.26 kOhm, not exactly sure the capacitor size needed for the lme49990. He's running it differential, so they're just two identical circuits for each leg.

And he said that the opamp for voltage mode was a straight opamp swap, so the as-provided circuit.
 
Member
Joined 2009
Paid Member
Yes, nothing special. See attached schematic.

And the voltage mode was with the original circuit and op-amp, except for the brief test with the LM4562 instead of the NJM5532.
 

Attachments

  • DAC Buffer for ES9038Q2M_test.png
    DAC Buffer for ES9038Q2M_test.png
    44.8 KB · Views: 2,002
Member
Joined 2009
Paid Member
As I wrote here they are quite similar. The new one also use voltage mode. Some of the resistors and capacitors in the filter have been changed slightly compared to the "old" version.
I don't think the minor changes have any significant impact on the performance.

So it looks like the main decision point is whether you want to:
1. Have most of the power supply on the PCB and I2S input only (new)
or
2. You want SPDIF and Toslink inputs, but can accept to add a regulated power supply (old).
 
As I wrote here
So it looks like the main decision point is whether you want to:
1. Have most of the power supply on the PCB and I2S input only (new)
or
2. You want SPDIF and Toslink inputs, but can accept to add a regulated power supply (old).

Option 2 has some other features too ...
2. You want switch-able I2S,SPDIF and Toslink inputs, Digital Volume control option, and filters (a digital "pre-amp' in essence)...but can accept to add a regulated power supply (old)