RIAA pre amp oscilates.

[nigelwright7557]

I have just had a pcb made for an RIAA pre amp.
The circuit has 100nf and 47uf decoupling but oscilates at 250KHz. The circuit doesnt seem much different to other examples so I dont know why it oscillates. The op-amp is a NE5532.

[Conrad Hoffman]

If it oscillates, you've got positive feedback via some path. Without seeing the layout its hard to tell, but I'd start with the physical locations of the ground of R6 and C9. If C9 can "wiggle" the voltage at R6 slightly due to trace resistance and daisy-chaining, that'll do it. I also have had odd things happen when grounding inputs like pin 3 of the second opamp. I like to put some series resistance there.

[Moon Track]

There always is internal capacitance between output and input.
If you noticed in original construction there is suppressing input capacitor.
http://www.paia.com/riaa.asp

[nigelwright7557]

Originally Posted by Moon Track

There always is internal capacitance between output and input.
If you noticed in original construction there is suppressing input capacitor.
http://www.paia.com/riaa.asp

I added the 100pf capacitors but they made no difference to the oscillation.
If I short the input to ground it stop oscillating.
Looks like there is some feedback between output getting capacitively coupled to the input.
I dont have a cartridge connected so dont know if that would stop the oscillation.

[kaos]

Should have a series resistor on pin 3 of the input (100 – 510 ohms). A parallel cap across R6 is sometimes needed too (or sometimes from pin 3 to ground). What’s the point of the second stage unity gain inverter?

[nigelwright7557]

Originally Posted by kaos

Should have a series resistor on pin 3 of the input (100 – 510 ohms). A parallel cap across R6 is sometimes needed too (or sometimes from pin 3 to ground). What’s the point of the second stage unity gain inverter?

It could be coz the input is floating, I dont have a cartridge to load it.
On the scope there is 400mV on the input of what looks like oscillation feedback through capacitive tracks. Its certainly the same frequency as the oscillation. Pin 2 and 3 have tracks in parallel so this could be part of the problem.

The unity stage is for adjusting the gain if required.

[nigelwright7557]

Originally Posted by kaos

Should have a series resistor on pin 3 of the input (100 – 510 ohms). A parallel cap across R6 is sometimes needed too (or sometimes from pin 3 to ground). What’s the point of the second stage unity gain inverter?

I tried a 1k on input and it still oscillates.

Does anyone know what the loading of the cartridge is ?
It could that would load the input and stop the feedback breaking through.

If I connect a signal generator the circuit settles down but it is being loaded by 600ohms.

[grufti]

MM cartridges are often about 1k. You can check specs on the ortofon website for example.

[kevinkr]

I'd start by seeing if the amplifier is stable without the EQ caps installed, if it isn't you need to do some work to figure out what the mechanism is.

IIRC The 5532 is unity gain stable so the high levels of HF feedback should not cause a problem..

Note that if you have a lot of capacitance at the output of the first amplifier stage a small series resistor right at the output (usually after the feedback connection) may help.. If the feedback network is part of the problem a small (say 100 ohms) resistor between the op-amp output and the network and connection to the next stage can help.

You can also scale the feedback network, lower resistor values and larger caps also help wrt to current noise performance with 5532. This also would reduce the effect of stray capacitances on EQ accuracy and phase margin. IMO Your resistors should be 1/10 the current values and the caps 10X. They are way too high for a 5532 wrt to input current noise. I suspect the amplitudes involved would not cause significant problems despite the effective 150 ohm load at high frequencies. (If you are concerned scale for 600 ohms) I bet you also have significant offset at the output that would go away.

Not unlikely that layout design is part of the issue - can you take a picture of your implementation?

[kaos]

Originally Posted by nigelwright7557

I tried a 1k on input and it still oscillates.

Sounding like a ground or layout issue. The 5532 is usually pretty easy to apply. Are you using a star ground? Are the 5532s prime units, or junk box?