Another stupid opamps question

[ricothetroll]

Hi,

While testing a circuit with a non-inverter amplifier (TL074), I noticed that the saturation level was pretty low : about +/- 5V measured on the scope, with +/- 12V supply. My resistor values were 100R (R1) to ground and a 1k pot (R2) as feedback resistor, and at the non inverting input a 34k-100p RF-filter, 10uF coupling cap followed by 1M to ground. I then switched R1 for 1k and R2 for a 10k pot, and then the saturation was back to normal for a TL07x, about +/-10V.

My PSU is built around LM317 and LM337.

I don't understand why I had such results, so any clue is welcome !

Best regards.

Eric

[R.G.]

Probably current limiting.

Most modern opamps can only drive a load as low as 2K or so to full output. The opamp output has to drive the feedback resistor and the gain set resistor as well as any output loading. Make your 100R be 1K, your feedback pot 10K, change the capacitors to get the same time constants, and see if it doesn't work better.

[ricothetroll]

Hi RG,

Thanx for your answer !

I actually did replace my resistors by those values and now it works good ! I wanted to test lower values for noise issues, but I think 1k/10k is good enough anyway.

Best regards

Eric

[Mark Hammer]

Apparently TL07x op-amps also exhibit better noise specs when the resistance values are higher.

[merlinb]

Apparently TL07x op-amps also exhibit better noise specs when the resistance values are higher.

FET opamps perfom better than BJT opamps when resistances are high, but noise is always reduced as resistance is reduced, whatever kind of opamp is used.

[ricothetroll]

Apparently TL07x op-amps also exhibit better noise specs when the resistance values are higher.

That's a good thing to know when designing ! Thanx !

FET opamps perfom better than BJT opamps when resistances are high, but noise is always reduced as resistance is reduced, whatever kind of opamp is used.

Then I think 1k/10k isn't a bad compromise !

Best regards.

Eric

[Gus]

noise and jfets
http://www.qsl.net/n4xy/PDFs/Semiconductor_Application_Notes/Vishay-Silcnx_FET_LoNoise_AN106.pdf

[ricothetroll]

Thank you Gus !

[PRR]

> I wanted to test lower values for noise issues
> Apparently TL07x op-amps also exhibit better noise specs when the resistance values are higher.

No, it is that the noise Voltage of TL072 is so high, that small resistors are of no benefit.

OTOH, the noise Current of TL072 is so low that large values do not hurt as much as they might with BJT amps.

Noise voltage of TL072 is near 1.4uV.

Noise voltage of 200 ohms is near 0.2uV.

So noise voltage of 10K ohms is near 1.4uV.

If the source impedance and the to-ground NFB resistor are under 10K, their noise contribution is negligible compared to TL072. There's no point going smaller than say 5K or 3K. Merlin is right that smaller resistors reduce noise but the improvement becomes negligible while there are other consequences (TL072's THD rises below 2K total load and just won't pull more than 20mA).

If your source is under 10K and noise voltage is actually a problem. 5532 is a better pick.

Guitar impedance above 1KHz may be over 10K. A 5532's input current noise will induce large noise voltage. A TL072 has "no" input noise current, its noise voltage is lower than the 20K-100K source, a TL072 is essentially noiseless in this application.

[ricothetroll]

Thanx a lot PRR !