ODAC - general and technical discussions

[GeorgeBoudreau]

Originally Posted by Neil_J

I'm wondering why the Cbyp cap on the MIC5205-3.6YM5 (pin 4) was omitted? Does it run better without it? Or would it improve noise to add it?

Hi,

A quick history lesson:

The original design used a 3v3 ADP151 which does not need a bypass. When it was decided to push the output voltage to 2Vrms NwAvGuy moved to a 3v6 regulator and decided on the MIC5205. It had the same footprint as the ADP151, SOT-23, and when NwAvGuy tried it out he was able to reach the 2Vrms target.

Yes, the MIC5205 has a bypass for noise reduction but the lack of the capacitor was not noticed when the board underwent analysis. The ODAC noise floor is at the limit of the DAC and I am not sure if the addition of the capacitor would have improved the numbers. That being said there is plenty of room for an adventurous individual to scrap a ground pad and add the recommended 470pf capacitor. ( 0805 size would be fine).

Regards,
George Boudreau
Yoyodyne Consulting

[Neil_J]

Originally Posted by GeorgeBoudreau

Hi,

A quick history lesson:

The original design used a 3v3 ADP151 which does not need a bypass. When it was decided to push the output voltage to 2Vrms NwAvGuy moved to a 3v6 regulator and decided on the MIC5205. It had the same footprint as the ADP151, SOT-23, and when NwAvGuy tried it out he was able to reach the 2Vrms target.

Yes, the MIC5205 has a bypass for noise reduction but the lack of the capacitor was not noticed when the board underwent analysis. The ODAC noise floor is at the limit of the DAC and I am not sure if the addition of the capacitor would have improved the numbers. That being said there is plenty of room for an adventurous individual to scrap a ground pad and add the recommended 470pf capacitor. ( 0805 size would be fine).

Regards,
George Boudreau
Yoyodyne Consulting

Absolutely great reply, thanks!

Looking forward to buying one of these....

[DoktorSteve]

The noise on C101 is about 30mV p-p on my scope but that probably includes some radiated noise which is well above 1mHz. I really cannot see much below 1 MHz apart from a 20mV ripple which I think is due to the charge pump. You can get rid of this ripple with a 100uF electrolytic but I cannot hear any difference. The power up current may exceed the USB limits with this in place or it may simply be current limited by the regulator.

I wonder if anyone else has been able to increase the value of the post filter resistors R102 and R103. I increased mine to 1k (my lowest value above 220R) which I think lowers the knee frequency by a factor of 3 and reduces measured output at 20kHz by 5% ( is the 0.4 dB?). I now find the ODAC/O2 a lot easier to listen to. I have yet to wire the 2 resistor values to a switch so that I can blind test.

[DoktorSteve]

Humble pie time. I wired 1k and 220R resistors to a switch and flicked between them. I was not able to detect any difference between the positions. I could not even detect the change in volume due to the change in resistance (<.5dB) to get a clue which resistor was in use. I tried with NMJ2068, 5532 and LM4562 and I still could not tell the difference.

[ceausuc]

How much voltage can this DAC handle?
Actually what are the MAX and MIN voltage that can take to work properly?
Thank you.

[GeorgeBoudreau]

Originally Posted by ceausuc

How much voltage can this DAC handle?
Actually what are the MAX and MIN voltage that can take to work properly?
Thank you.

The MIC520x regulator is rated for a MAX +16volt input. However, input voltage is only one consideration when talking about a regulator. This is a very small device and it is not capable of dissipating much heat. The controller has its own regulator and consumes approx 100ma while the ES9023 only uses 30ma. Running with the 5volts from the USB port both regulators are well within operating limits.

George

George

[ceausuc]

Originally Posted by GeorgeBoudreau

The MIC520x regulator is rated for a MAX +16volt input. However, input voltage is only one consideration when talking about a regulator. This is a very small device and it is not capable of dissipating much heat. The controller has its own regulator and consumes approx 100ma while the ES9023 only uses 30ma. Running with the 5volts from the USB port both regulators are well within operating limits.

George

George

Thank you, I understand that.
I would like to try a different power source and I can't find easily one rated at 5V. It's easier to find 6/7V sources. Do you think would that be too much? Did you ever tested the DAC with higher voltages?

[GeorgeBoudreau]

Originally Posted by ceausuc

Thank you, I understand that.
I would like to try a different power source and I can't find easily one rated at 5V. It's easier to find 6/7V sources. Do you think would that be too much? Did you ever tested the DAC with higher voltages?

OK.. a 7 volt supply will double the power dissipation and be very near the upper limit for the U103 regulator. I cannot say how hot the chip would be but I expect it to be very very warm. The ES9023 DAC regulator, U101, will be fine at 7 volts.

If you plan on powering the ES9023 separately you will need to break the trace leading to L100 and apply your new power source between L100 and ground.

DO NOT attempt to feed 7volts to L100 without breaking the trace on the back side of the board or you will damage your PC.

::WARNING:: This has not been tried before and you could smoke your ODAC, PC or external power supply.

Have fun

[DoktorSteve]

Hi George, concerning the addition of the bypass capacitor I would suggest that people don't bother! The cap actually reduces the amount of feedback at high frequencies. I tried adding 220pF to pin 4 of U101 with the 22uF in C101 and the ripple increased to over 100mV in the MHz region (why do I never write this stuff down?) i.e. the regulator with the bypass capacitor is too slow and cannot compensate for the periodic discharge of C101 by the charge pump (and maybe oscillates?). You can see the error signal on pin 4. It is a bit unclear from the description on the Micrel document what the noise cap does but he diagrams show that it filters the feedback signal.

[captain pugwash]

My PC and Amp are about 5m apart in the same room, which is best - a 5m USB cable connected to the odac at the amp end via a short phono cable or a short USB cable with a 5m phono cable to the amp?