RIAA stage problem ...

[jiiteepee]

Hi!

I have a fw audio interface with RIAA stage in-build.
Problem is that the RIAA EQ it's not OK.

Here's a graph showing the eq curve this stage produces (sorry 'out the weak resolution):



Since 'error' is in both channels, that would tend to indicate a bad design, or more likely, the wrong filter components were installed during manufacture?

Is it just treble roll-off 'stage' (2122Hz) missing or, is the bass turnover 'stage' (~500Hz) missing as well?


Juha

[bob91343]

Before condemning the circuit, make sure you aren't overdriving it. These circuits work down to microvolts so it doesn't take much to overdrive and get erroneous response curves.

[john curl]

Looks like no hf rolloff.

[jiiteepee]

Originally Posted by bob91343

Before condemning the circuit, make sure you aren't overdriving it. These circuits work down to microvolts so it doesn't take much to overdrive and get erroneous response curves.

Originally I noticed the problem through listening and not through measure.

My cartridge Output voltage is: 2.0mV, 1kHz , 5cm/s to zero peak, lateral velocity (5.6mV, 1kHz, 10cm/s, zero to peak, 45° velocity (DIN45500)).
If the cartridge output voltage is too high for ESI phono input then, shouldn't there be hearable distortion in playback? There isn't.
Here is a sample recorded through this ESI device and another for comparison sake:

ESI DuaFire 16/44.1 02_DS_WDYTYG_DuaFire.wav ( 5.08MB )
Terratec iVinyl 16/44.1 04_DS_WDYTYG_iVinyl.wav ( 5.08MB )

BTW, when I measured this device using RMAA software, I got warning of high (>6%) distortion so, it's possible that curve in graph isn't fully valid. I also measured another RIAA pre-amp and got same warning but the RIAA eq curve looked follow the standard.


Juha

[jiiteepee]

Originally Posted by john curl

Looks like no hf rolloff.

That's my thought as well but, IMO, by the picture, there's some roll-off berween frequency range 500Hz-5kHz. If the 2122Hz roll-off is missing totally then shouldn't the curve be turning more upwards after ~1kHz?


Juha

[SY]

Originally Posted by jiiteepee

If the 2122Hz roll-off is missing totally then shouldn't the curve be turning more upwards after ~1kHz?

No, probably not. The zero at ~500Hz will just cause the response to flatten out, which it sorta does. I think John's right here, either the pole at 2122Hz is missing or it's in the wrong place (possibly a cap or resistor value off by a decade), ASSUMING the response curve you're showing is correct. If there really is high distortion, all bets are off and you need to concentrate on fixing that issue first.

[jiiteepee]

Originally Posted by SY

No, probably not. The zero at ~500Hz will just cause the response to flatten out, which it sorta does. I think John's right here, either the pole at 2122Hz is missing or it's in the wrong place (possibly a cap or resistor value off by a decade), ASSUMING the response curve you're showing is correct. If there really is high distortion, all bets are off and you need to concentrate on fixing that issue first.

Distortion was in measure only (I suppose it came because of wrong method I used (RMAA loopback test where E-MU 0404 USB output fed the ESI phono input --> too high output level after RIAA stage)).

So, I need to open the case (again) and try to
- find the values of resistors/caps used in RIAA stage and
- solve the stage schematic

Juha

BTW, here's maybe a better picture showing the 'error' in curve (30s sweep from 5 to 22.05kHz):

[jiiteepee]

Missing edit button.

Tracing the circuit is not an easy task for me but, hey, here are the pictures showing stage compnents. I quess the RIAA stages are just after the phono input ports?

1st picture shows the connectors on rear panel.
2nd picture shows the uper side of the board.
3rd picture shows the other side of the board.


1. http://img39.imageshack.us/img39/5894/esiinpsw.jpg
2. http://img826.imageshack.us/img826/7079/06082010566.jpg
3. http://img251.imageshack.us/img251/5931/06082010570.jpg

Values of most of those tiny components can be seen on those attached pictures. Is it possible to tell by the pictures if there are something on those components (values/placing/...) which might be the reason for this issue in question (I don't have equipment other than an old multimeter which I could use for to measure resistors, and I think if I take some that tiny parts off from there I can't get them back again with my 500W soldering iron )?

(I took those two pictures using mobilephone camera so, the quality of takes isn't best possible).

Juha

[davidsrsb]

It looks like the 75us pole is there, but closely followed by a zero at about 50us. Is the non-inverting opamp circuit hitting its minimum unity gain at far too low a frequency? This would happen if the resistor from -ve input to ground is too large

[jiiteepee]

Originally Posted by davidsrsb

It looks like the 75us pole is there, but closely followed by a zero at about 50us. Is the non-inverting opamp circuit hitting its minimum unity gain at far too low a frequency? This would happen if the resistor from -ve input to ground is too large

Hmm... do you mean
- the JRC4580 op-amp near the line/phono switch (picture 2) or another op-amp found some other position (there are few of those onboard)?
- case of resistot R1 in picture ?




Juha