Help interpeting CRO traces please?

[rodneytoady]

Hi. Sorry, this won't be the easiest read - but I've tried to provide all the information. I've just built a very simple op-amp based 6dB/octave LF response-shaping filter for audio use. The turnover point is about 30Hz. The image attached called "simple_schematic_1" is the working circuit. The 9.4 kOhm resistor in the output part of the circuit originally got there by accident, a result of other circuits I was messing around with. This is fine. It does what I want. Attached image "sharp.jpg" shows the CRO screen input and output traces (output is below, showing a little gain because of resistors 4.7 kOhm and 10 kOhm in the feedback circuit of the first op-amp).

I need to get rid of the spurious 9.4 kOhm resistor on the output to add other circuits (a twin-T notch filter, in fact). I tested what happens when I bypass the resistor with a 2 cm length of wire (a little bit of resistor lead) and suddenly the circuit gets noisy (please see attached images "simple_schematic_2.gif" and "fuzzy.jpg"). It doesn't look too bad in the images because photographing a CRO screen with a cheap digital camera is not easy for me - ALL the traces appear unsharp but were quite clean and sharp in real life, except the lower trace in this one). I also got an image of the fuzzy output trace highly magnified at 1 uS per division time scale. ("enlarged.jpg"). Again, this was a bit hard to photograph well, but it looks like high-frequency oscillation, perhaps?. RF interference perhaps?

I'm completely at a loss how to interpret what's going on here. Why is the circuit quiet with the (apparently) spurious 9.4 kOhm output resistor and noisy when it's bypassed with a short piece of straight-wire? I don't feel confident moving on until I understand why this is. Can anyone offer a suggestion? Thanks, Rodney.

 

[MikeMl]

That thing on the output end is a 100K pot with the wiper grounded, and the right end just hanging open?
Where is the pot set when you jumper around the resistor?

If it is set too close to the left end, the pot resistance (to ground) could be so low as to overload the output current capability of the opamp.

 

[rodneytoady]

more info

Actually, it's facing into another identical op-amp, more or less infinite impedance, so I thought an unterminated output was a fair representation. Interesting point about the pot impedance, but I understood that it presented a 100 kOhm load no matter where the volume was set. Have I drawn that part of the schematic wrong? I'm in Australia, so it's just about to go night time here, apologies if my next reply is tardy. Rodney.

 

[ericgibbs]

Actually, it's facing into another identical op-amp, more or less infinite impedance, so I thought an unterminated output was a fair representation. Interesting point about the pot impedance, but I understood that it presented a 100 kOhm load no matter where the volume was set. Have I drawn that part of the schematic wrong? I'm in Australia, so it's just about to go night time here, apologies if my next reply is tardy. Rodney.

G'day Rodney.
Can you post a full circuit diagram also showing the twin T section and following OPA.?

As you describe the final circuit it would appear that the 100Kpot and the 9.4k are redundant anyway.??

 

[rodneytoady]

here's what I did instead

Hi Eric. I'm not too good with the drawing program I used to do the circuits, so it would be fairly onerous for me to do a complete circuit diagram (it's part of an electronic crossover). Instead, I did this; I un-soldered the output wire so it really is an unterminated circuit exactly as drawn. I also removed the pot from the circuit. So we have a circuit running into infinite impedance with nothing except the spurious resistor (or not) between the output of the second op-amp and the CRO test probe. So now the circuit looks like the new diagram attached. Repeated the tests, same result. There's something weird going on when I bypass that resistor. Could that tiny part of the circuit oscillate? No caps or chokes, can't see how. Yawn...I'm off to bed now, will cheeck keenly over coffee tomorrow AM. Rodney

 

[rodneytoady]

Ahh...could the CRO itself be influencing the result? I mean in one case there's a resistor in between op-amp output and the CRO test probe, in the next test, there is not. ??? Too techy for me. Rodney again.

 

[ericgibbs]

Ahh...could the CRO itself be influencing the result? I mean in one case there's a resistor in between op-amp output and the CRO test probe, in the next test, there is not. ??? Too techy for me. Rodney again.

hi R,
catch you tomorrow,,, that last circuit dosnt show the twin T filter network.

Does your scope probe have a *10 switch..??

 

[rodneytoady]

full circuit diagram

Hi again,

Here's a full diagram of the proposed circuit from input to output. Note that I have stopped building at the red line because I'm baffled by this noise problem, described in my first post. Cheers, Rodney.

 

[rodneytoady]

Oh yes, the CRO probe DOES have a *10 switch.

 

[audioguru]

Please double-check the part number of the IC. An LM5534 was an old core-memory driver, not an opamp.
Maybe you have NE5534 opamps?
The NE5534 has a wide frequency response because it has a small-value internal frequency compensation capacitor. With the low gains in your circuit an external capacitor is recommended to be added between pin 5 and pin 8 as shown in its datasheet to prevent oscillation.

I betcha you made your circuit on a breadboard and the capacitive-coupling between the input wire and the output wire causes oscillation at a high frequency.
Maybe you forgot to use supply bypass capacitors and positive feedback comes through the power supplies.

 

[rodneytoady]

Yep! NE5534AM op amps. I stand corrected. Your suggestions sound exactly correct. I will try that out (but this may take a while - I've got family responsibilities today). Thanks. Rodney.