PCM1704 NOS Strange Clipping

LuisMCP · 2013-03-27 11:37 pm

The DAC seems to work properly in the beginning a nice sinusoidal can be saw in scope, but 1 minute after, a clipping in negative zone is slowly growing up to a certain level.

Both channels L and R are affected, two different I/V converters was tried and same results was obtained.

Pin 17 (SERVO_DC) has different voltage in both chips. One at 0,00V and another at -2,40 volts.

Some idea what the matter?

Regards

georgehifi · 2013-03-28 1:38 am

Are pins 11 and 20 at + and - 5vdc constant? Maybe one of your 5v regulators are on the fritz when it warms up.

Cheers George

Ken Newton · 2013-03-28 11:56 am

I agree with George, the first thing to do is check whether the power supply rails are stable and without significant ripple voltage, especially, the negative rail.

LuisMCP · 2013-03-28 1:10 pm

I'm using a Twisted Pear Dual Power Supply (LCBPS).
I have checked +/-5V correctly, and I have change LM337 and electrolitics caps and all remains the same.

1-2 minutes of clean signal, then clipping in negative sinusoidal. Power off the DAC and exactly the same thing ever and ever.

Both channels fails equally, I have interchanged the PCM1704 (they are socketed to DIP 20 adapter), I have put only one PCM1704 in L, then in R channel, and allways get the same results.

I don't have another PSU for testing right now, but probably this afternoon I would get one.

merlin el mago · 2013-03-28 2:50 pm

Have you checked with scope?

LuisMCP · 2013-03-28 4:37 pm

merlin el mago said:
Have you checked with scope?

Yes, as I have mentioned it in my fist post.

merlin el mago · 2013-03-28 5:03 pm

Luis could you post a pic and schematic used?

s3tup · 2013-03-28 5:52 pm

Drop the I/V, put some 20ohm resistor to ground instead, check with scope. This way we eliminate everything but the DAC and it's digital side.
If the clipping is still there, and the supplies are absolutly fine then... well, i hope you wont see that clipping thing again

LuisMCP · 2013-03-28 6:03 pm

s3tup said:
Drop the I/V, put some 20ohm resistor to ground instead, check with scope. This way we eliminate everything but the DAC and it's digital side.
If the clipping is still there, and the supplies are absolutly fine then... well, i hope you wont see that clipping thing again

interesting. I will try this

A question... I/V I use is a common opamp v to i converter, and of course, inverts the signal. Then, If I see negative side clipped in scope and i/v inverts, DAC output is with positive side clipped, and the problem would be in positive rail of psu, not the negative rail.

s3tup · 2013-03-28 6:10 pm

Opamp's PS may fail too, so it won't supply the current required = give high virtual impedance = clip. What is the resistor value in opamp's feedback?

Anyway, lets check the DAC first

LuisMCP · 2013-03-28 7:10 pm

s3tup said:
Drop the I/V, put some 20ohm resistor to ground instead, check with scope. This way we eliminate everything but the DAC and it's digital side.
If the clipping is still there, and the supplies are absolutly fine then... well, i hope you wont see that clipping thing again

s3tup said:
Opamp's PS may fail too, so it won't supply the current required = give high virtual impedance = clip. What is the resistor value in opamp's feedback?

Anyway, lets check the DAC first

I have drop the I/V and load the DAC with 22R and NO clipping at all.
A perfect sine, triangular wave, square and dirac impulse

(24bits /96kHz) in my scope

I have implemented i/v as pcm1704 datasheet with psu of batteries, same that uses Sen I/V (that also clips although more gently)

One channel with OPA627 and other with AD8610. RF = 2k7, CF = 330pF (MKP)

Both opamps clips at the same time...

I'm planning to get LME49713 for I/V
OPA627 seems to sound best to my ears

Regards

s3tup · 2013-03-28 7:21 pm

Great, so DAC is okay.

Let's remove the CF first, then take a photo of actual circuit...
These opamps may oscillate, so their layout and PS bypassing is critical.
Check their supplies whether they are okay... Batteries aren't that consistent.

LuisMCP · 2013-03-28 9:24 pm

s3tup said:
Great, so DAC is okay.

Let's remove the CF first, then take a photo of actual circuit...
These opamps may oscillate, so their layout and PS bypassing is critical.
Check their supplies whether they are okay... Batteries aren't that consistent.

Problem solved !! :drink:

At the beginning, the I/V stage was a Sen I/V of xen Audio that I implemented for testing. This circuit uses a floating PSU with batteries because is very sensible to noise and has a very low PSSR if any.

As a good diyer

i'm very lazy to build a new psu for an opamp i/v stage, and take the same psu used for Sen I/V, forgeting that it was FLOATING psu.
With the time, GND is wandering, and makes to signal cllips very nasty.

As I could see Sen I/V also clips with this system of floating PSU.

Next days I could to compare Buffalo III vs PCM1704 NOS, but 1704 is best to my ears.

Attach some captures on my scope, including a dirac impulse (img.2) at a resolution of 24/96.