OK - yes. I tried to discuss the topic more in general. The AVCC of course becomes a modulator if one varies the ref. voltage. Maybe even a transistor?
What mechanisms works as a pure adder without modulation?
//
Noise in the reconstruction filter or in any amplifier in an analogue signal path is usually almost completely additive. Almost, because there are things like resistor 1/f noise (random resistance fluctuations) that can modulate the signal to some extent. For DAC reference noise, amplitude modulation is usually the main effect, not some small second-order effect.
OPA1612 is probably the best choice, low-noise, loop stable on MLCC >3-5uF load, and has 100,00% enough output current to feed 9039q2m.
The only thing I like to call stupid is powering the opamp from 12V(SO-8 probably overheated), when you need just 3.3V, USB 5V rail is more than enough 😉
The only thing I like to call stupid is powering the opamp from 12V(SO-8 probably overheated), when you need just 3.3V, USB 5V rail is more than enough 😉
I tried once with an opamp and it didn't work, but I don't remember if it was opa1612, opa1656 or something else. I have SMSL DL200 and I'll try again, maybe I'm wrong.
For sure wrong, even about 2.25V 😉 VS = ±2.25 V to ±18 V, or 4.5 V to 36V.
Following your idea 4.5 + 3.3 = 7.8V right?
You have to take a look at the Input range first (>=2V from power rails), and the Output range.
Following your idea 4.5 + 3.3 = 7.8V right?
You have to take a look at the Input range first (>=2V from power rails), and the Output range.
I think your calculations are wrong.
1) OPA1612 has the power supply 4.5V (2*2.25) minimum, so 5V is enough.
2) Maximum OUTPUT VOLTAGE is VCC-0.6V @ 2kOhm load.
Looking at datasheet's Figure 27 , we can expect less than1V from the rail also at higher current (20-30mA).
5-1=4 > 3.3v, so this is also OK.
IVX is right, OPA1612 is enough for Q2M.
But I have previously suggested OP for the Pro version and other DACs with higher current requirements.
Alex.
If you want 3.3 V out, I agree that a 5 V supply should be OK.
But there seems to be a problem on the input, at least according to the datasheet. The common mode voltage range only goes to (V+)-2 V, so 3 V with a 5 V supply.
With e.g., a 2.5 V reference and a bit of gain, this can of course be corrected. But it may give some additional noise.
It may work at 3.3 V in with a 5 V supply, but it is outside the specification for the OPA1612.
But there seems to be a problem on the input, at least according to the datasheet. The common mode voltage range only goes to (V+)-2 V, so 3 V with a 5 V supply.
With e.g., a 2.5 V reference and a bit of gain, this can of course be corrected. But it may give some additional noise.
It may work at 3.3 V in with a 5 V supply, but it is outside the specification for the OPA1612.
https://www.diyaudio.com/community/threads/es9038q2m-board.314935/page-133#post-5543691
the link above goes quite deep into the AVCC power supply topic 🙂
the link above goes quite deep into the AVCC power supply topic 🙂
So the best solution is exactly what is already in the DAC, provided the OPA1612 can handle the power (there are no thermal data whatsoever in its datasheet). If it should overheat, you could use a preregulator to 6 V.
2x OPA210 (or OPA2210) is even better than OPA1612. Lower 1/f noise and available in smaller package. Cheaper as well.
I measured the actual voltage out from the opa1612 to be 3.65v.
Looks to me that this approach could be used to generate multiple high quality supply's with identical voltage.
If I used one LT3045 LDO to supply the reference voltage to lets say 3 OPA1612 and use a double LT3045 to supply +6v to pin 8 of all these OPAs then theoretically I get 6 high quality 3.3v supplies to feed 2 clocks, 2xAVCC, 1x DVCC and whatever else.
Any objections to this approach?
Looks to me that this approach could be used to generate multiple high quality supply's with identical voltage.
If I used one LT3045 LDO to supply the reference voltage to lets say 3 OPA1612 and use a double LT3045 to supply +6v to pin 8 of all these OPAs then theoretically I get 6 high quality 3.3v supplies to feed 2 clocks, 2xAVCC, 1x DVCC and whatever else.
Any objections to this approach?
Don't know what your goals are but I doubt hacking the device will result in improvement. LT3045 alone does not guarantee anything and separate LDO board is not the best use of LT3045. Another thing is that AVCC is the critical supply on ESS dacs. I wouldn't touch other supplies unless you are sure there is something wrong with them.
You certainly don't understand me. This approach works really well and that's why I want to try it out on my other dacs.
There are more places in a dac where clean 3.3 volt is needed. If all it takes is to supply one single clean reference voltage to pins 3 and 5 and clean 6-12v to pin 8 to generate 2 x 3.3v then why not use three opamps supplied by one reference 3.3v to get 6 x 3.3v. This reference voltage could be generated by one single LT3045 LDO and the supply voltages to all the pins 8 could as well be fed by one single LT3045.
There are more places in a dac where clean 3.3 volt is needed. If all it takes is to supply one single clean reference voltage to pins 3 and 5 and clean 6-12v to pin 8 to generate 2 x 3.3v then why not use three opamps supplied by one reference 3.3v to get 6 x 3.3v. This reference voltage could be generated by one single LT3045 LDO and the supply voltages to all the pins 8 could as well be fed by one single LT3045.
You could try it. However, AVCC (Vref) is the most sensitive and critical. IME clocks are next most sensitive since there are analog oscillators inside clock modules. VCCA and or DVCC should probably be separate, but they are not as sensitive as the first two above. VCCA includes the clocking functionality in the dac chip. DVCC is for digital stuff.
Thanks Mark
I will certainly use separate supplies for any 3.3v input.
I think I gonna use Dip adapters with OPA2210 attached close to the next cap.
Thanks bohrok2610 for the tip.
I will certainly use separate supplies for any 3.3v input.
I think I gonna use Dip adapters with OPA2210 attached close to the next cap.
Thanks bohrok2610 for the tip.
Its also recommended to keep the regulators very close to the loads. Should be on the same ground plane. If its necessary to make a small regulator board maybe it can be connected as an extension to the dac board ground plane. OTOH, interconnection wires between the dac board and loads involving RF can cause problems. IIRC even AVCC has some RF loading according @KSTR .
Regarding sockets, they add inductance. Maybe better to use SMT parts and swap them out using solder if experiments need to be run.
Regarding sockets, they add inductance. Maybe better to use SMT parts and swap them out using solder if experiments need to be run.