Using the AD844 as an I/V

[georgehifi]

Played around a liitle more this morning to try to reduce the staircase noise comming from the PCM1704 to the input of the AD844 (remember mine is straight in), so I put a 1000pf to ground on pin 2 (input) (ala Pedja's TDA1541 circuit) to see what effect it had on the noise, by my calculations with the 1k output impedance of the PCM1704 then a 1000pf to ground this should filter -3db @160khz? But it did nothing, nada, zip to the staircase noise noise or amplitude. I didn't want to go larger with the cap, as I feared taking out the pcm1704.

Cheers George

[zenelectro]

Originally Posted by georgehifi

Played around a liitle more this morning to try to reduce the staircase noise comming from the PCM1704 to the input of the AD844 (remember mine is straight in), so I put a 1000pf to ground on pin 2 (input) (ala Pedja's TDA1541 circuit) to see what effect it had on the noise, by my calculations with the 1k output impedance of the PCM1704 then a 1000pf to ground this should filter -3db @160khz? But it did nothing, nada, zip to the staircase noise noise or amplitude. I didn't want to go larger with the cap, as I feared taking out the pcm1704.

Cheers George

Yep. But of course it is not the 1k OP Z of DAC that you use to size the cap but the -IP Z- of the IV. As such the cap will end up being huge,
it will cause HF noise multiplication (of the IV not DAC) through the IV, and
decrease it's linearity.

No free lunch George. I've done all that stuff and much more. If you really
want a LPF -before- the I-V, try an air core inductor between dac and
IV then a cap to gnd directly at DAC OP. Given physical constraints of the
size of inductor, the LPF frequency will be very high. To get this right,
prolly need to simulate it.

Why do you need to filter the DAC beyond simple OP LPF? What is it you are
hearing?

Z

[georgehifi]

Originally Posted by zenelectro

Why do you need to filter the DAC beyond simple OP LPF? What is it you are
hearing?

Z

It's just truly amazing as it is, but I cannot help myself I have to experiment, and I thought if I could lower the staircase noise before the AD844, it could lessen the duties that the I/V stage has to perform, instead of first getting filtered at TZ output.
I wonder then why Pedja put a 1.5nF to ground on the TDA1541 AD844 circuit.

Cheers George

[AudioLapDance]

Originally Posted by zenelectro

Yep. But of course it is not the 1k OP Z of DAC that you use to size the cap but the -IP Z- of the IV. As such the cap will end up being huge,
it will cause HF noise multiplication (of the IV not DAC) through the IV, and
decrease it's linearity.
...

Could you say more, sir?

Is the 1541 case different from the 1704 case?

Cheers,
Jeff

[abraxalito]

Originally Posted by dirkwright

Here's what a more professional RFI resistant box looks like. This is from Par-Metal (I am only a satisfied customer) and contains a SV oscillator (not quite finished yet).

Not so fast It matters not one iota how resistant the box is if you don't filter the cables coming in and out of it. The cables are the Trojan Horse here.

@zen - are you seriously recommending an air-core between the DAC and the I/V stage? Do not go there Use passive I/V then use the inductor (lots, and lots of inductors) between the I/V resistor and any active stage.

[qusp]

I prefer these, been drooling over the zerocases for ages, but have decided i'm getting a couple this year, covers EMI and shock/vibration mounting as well, also getting a couple of their deepdrawn mini enclosures/cans, maybe mumetal for the clock, just for OTT value.

[dirkwright]

Originally Posted by abraxalito

Not so fast It matters not one iota how resistant the box is if you don't filter the cables coming in and out of it. The cables are the Trojan Horse here.

I assumed normal filtering for all incoming and outgoing cables. What I said is still true: all those pretty black anodized boxes are basically worthless for shielding.

[dirkwright]

Originally Posted by georgehifi

It's just truly amazing as it is, but I cannot help myself I have to experiment, and I thought if I could lower the staircase noise before the AD844, it could lessen the duties that the I/V stage has to perform, instead of first getting filtered at TZ output.
I wonder then why Pedja put a 1.5nF to ground on the TDA1541 AD844 circuit.

Cheers George

I think that's a fine idea.

I've played with a 100nF to ground, followed by a 10 ohm resistor, and then into a standard IV opamp IV converter. It works in simulation, don't know about real life. Inductors add really strange phase shifts as I recall.

[georgehifi]

Originally Posted by dirkwright

I think that's a fine idea.

I've played with a 100nF to ground, followed by a 10 ohm resistor, and then into a standard IV opamp IV converter. It works in simulation, don't know about real life. Inductors add really strange phase shifts as I recall.

That's a big cap to ground that the PCM1704 will see, I'm too scared to try that, it could cook the 1704.

Cheers George

[jcx]

thinking wrong - Iout can happily go straight into gnd copper whithout hurting the chip