I have it connected to a scope, and I'm running a freq response plot using a scope.
This is a fairly simple opamp circuit, which i want to check for stability.
My problem is am IC gets warmer than i like at higher freqs, so i suspected an oscillation. Mind you i have just added an input R to gnd, since i forgot that earlier, and i hope that solves it.
Strangely i had none of this with my breadboard prototype.....
So it could still be oscillation.
This is a fairly simple opamp circuit, which i want to check for stability.
My problem is am IC gets warmer than i like at higher freqs, so i suspected an oscillation. Mind you i have just added an input R to gnd, since i forgot that earlier, and i hope that solves it.
Strangely i had none of this with my breadboard prototype.....
So it could still be oscillation.
Post a diagram of your circuit 
Here are a couple of threads on opamp oscillation. The CFB thread is a bit different but it shows what oscillation looks like and how it can show up.
http://www.diyaudio.com/forums/anal...rent-feedback-opamps-simulation-practice.html
http://www.diyaudio.com/forums/anal...u-have-checked-see-its-stable-havent-you.html
Here are a couple of threads on opamp oscillation. The CFB thread is a bit different but it shows what oscillation looks like and how it can show up.
http://www.diyaudio.com/forums/anal...rent-feedback-opamps-simulation-practice.html
http://www.diyaudio.com/forums/anal...u-have-checked-see-its-stable-havent-you.html
Drive the circuit with a square wave. If the output show ringing, it's on the verge of oscillation. It may also just break into oscillation.
If it's getting hotter than expected at high frequencies, it's probably just an effect of driving a capacitive load. Either external or internal.
Most of the time, too hot because of oscillation exists with no signal. Seeing as you're not seeing that... I can't rule out oscillation, but it doesn't look to be the case with the little information you've given.
If it's getting hotter than expected at high frequencies, it's probably just an effect of driving a capacitive load. Either external or internal.
Most of the time, too hot because of oscillation exists with no signal. Seeing as you're not seeing that... I can't rule out oscillation, but it doesn't look to be the case with the little information you've given.
Sorry guys!
Im at work and its slack today, but even so....
Thats my circuit (from my Filter Q thread...)
I have made up on stripboard, and switched to LME49740 ICs from the NE5532 I was using. This was mainly because I got them as samples and the Quads made the cct easier to implement in stripboard.
Mods to cct that arent shown in schematic:
1: I have added output resistors of 150 Ohm (as AndrewT or another helpful soul suggested)
2: I have just added 22k from sig to gnd also (just in case...)
3: I cut the common signal input to isolate the other feed to the second filter cct.
Im running +/- 12VDC rails, and I was driving with 10 Vpk-pk input, but I have since reduced it to 4V pk-pk.
I think running too close to the supply rails maybe?
Running at 4V pk-pk the devices are cooler, but not COLD.
Running into a scope im surprised they are warm at all.
Im at work and its slack today, but even so....
Thats my circuit (from my Filter Q thread...)
I have made up on stripboard, and switched to LME49740 ICs from the NE5532 I was using. This was mainly because I got them as samples and the Quads made the cct easier to implement in stripboard.
Mods to cct that arent shown in schematic:
1: I have added output resistors of 150 Ohm (as AndrewT or another helpful soul suggested)
2: I have just added 22k from sig to gnd also (just in case...)
3: I cut the common signal input to isolate the other feed to the second filter cct.
Im running +/- 12VDC rails, and I was driving with 10 Vpk-pk input, but I have since reduced it to 4V pk-pk.
I think running too close to the supply rails maybe?
Running at 4V pk-pk the devices are cooler, but not COLD.
Running into a scope im surprised they are warm at all.
Last edited:
Yes thats correct Mooly, I have 22k from sig to gnd.
Testing with a Square wave as advised was a bit weird.
No overshoot or anything like that, but instead at lower freqs I get a strange sawtooth type wave, moving to a couple of kHz and it becomes sinusoidal.......
Slewing? thats what it looks like. I cant see how that can be the case with LME49740.
Must double triple check layout and wiring, something isnt right here.
Testing with a Square wave as advised was a bit weird.
No overshoot or anything like that, but instead at lower freqs I get a strange sawtooth type wave, moving to a couple of kHz and it becomes sinusoidal.......
Slewing? thats what it looks like. I cant see how that can be the case with LME49740.
Must double triple check layout and wiring, something isnt right here.
Last edited:
Haha that is true.
I am more used to either just sine frequency responses in my amateur projects or thyristor traces using scopes here at work.
I cant say ive ever used square before. Except digital work
So yes more practical is always good. I'm 35 and only now am i learning the theory i should have studied after a levels...
Regardless, a HNC/D isn't enough. Hell, i don't think the degree im on is enough.
Audio is a funny game.
I am more used to either just sine frequency responses in my amateur projects or thyristor traces using scopes here at work.
I cant say ive ever used square before. Except digital work
So yes more practical is always good. I'm 35 and only now am i learning the theory i should have studied after a levels...
Regardless, a HNC/D isn't enough. Hell, i don't think the degree im on is enough.
Audio is a funny game.
I tested the lowpass seems ok. I checked the highpass today, and there is some overshoot.
So i rechecked for solder bridges, found a couple of suspect areas. Cleaned them up and removed flux residue, and the problem is reduced.
Since then ive made some small changes to the cct, removing the a level pot which i think was faulty as it caused severe clipping past midpoint. Clueless why.
Since then i haven't checked for ringing again. Tomorrow if i get time.
I'm currently considering an RC at say 50khz, although in not 100% if that will solve it.
Any advice or links that could help me here?
So i rechecked for solder bridges, found a couple of suspect areas. Cleaned them up and removed flux residue, and the problem is reduced.
Since then ive made some small changes to the cct, removing the a level pot which i think was faulty as it caused severe clipping past midpoint. Clueless why.
Since then i haven't checked for ringing again. Tomorrow if i get time.
I'm currently considering an RC at say 50khz, although in not 100% if that will solve it.
Any advice or links that could help me here?
Last edited:
When I upgraded the op amps in my disco mixer from 4558 to 33078, I knew I had a problem because the fan on the PA amp was running when there was no music. I had a working scope for about a month, and using a quasi sine source, the oscillation showed up as a grey band around the actual stable wave.
Now that I don't have a working scope, I use my analog 200 kohm/volt VOM to look for oscillation. Play some music that doesn't change volume (Pacobel Canon is good) into the circuit, look at the output with a .047 uf cap in series with the meter on 2 VAC scale (or 20 VAC on power amps). Take a reading. Reduce the size of the cap to .0047, then to 470 pf and 220 pf. If you've still got a significant reading with a 470 pf cap in series with the meter, you've got an oscillation.
Scope needs new electrolytic caps on the sweep (and everything else probably) but the boards are glued in the frame. !#@$% Waste of $40.
To kill the oscillation I had to 1. put ceramic .01 caps near the op amps location, about 1" away, 2. put 33 pf ceramic caps around the feedback resistor at 50x gain for the mag phono circuit.
Now that I don't have a working scope, I use my analog 200 kohm/volt VOM to look for oscillation. Play some music that doesn't change volume (Pacobel Canon is good) into the circuit, look at the output with a .047 uf cap in series with the meter on 2 VAC scale (or 20 VAC on power amps). Take a reading. Reduce the size of the cap to .0047, then to 470 pf and 220 pf. If you've still got a significant reading with a 470 pf cap in series with the meter, you've got an oscillation.
Scope needs new electrolytic caps on the sweep (and everything else probably) but the boards are glued in the frame. !#@$% Waste of $40.
To kill the oscillation I had to 1. put ceramic .01 caps near the op amps location, about 1" away, 2. put 33 pf ceramic caps around the feedback resistor at 50x gain for the mag phono circuit.
Last edited:
Thanks.
I think ive been a dunce again. I haven't decoupled close to the supply pins.....
The ground is not so close, so i may just settle with a small cap across the split rail supply under the IC socket. To be totally honest, i wasn't sure i could do that, instead of going from rail to gnd. Reading around here and i find that it should help.
Usually I'm making digital circuits in micro board and decoupling each IC with small 'lytics.
I don't know how i forgot the decoupling.....i knew i had to...
I think ive been a dunce again. I haven't decoupled close to the supply pins.....
The ground is not so close, so i may just settle with a small cap across the split rail supply under the IC socket. To be totally honest, i wasn't sure i could do that, instead of going from rail to gnd. Reading around here and i find that it should help.
Usually I'm making digital circuits in micro board and decoupling each IC with small 'lytics.
I don't know how i forgot the decoupling.....i knew i had to...
I installed the ceramic caps rail to rail about an inch from the op amp, drilling through the PC board and bending the legs of the cap over the trace and soldering. I got away with it. I also got away with one cap for two op amps. Yeah, my ground wasn't anywhere convenient, either, also 4 holes drilled instead of 2 would be a nuisance. My drill chuck won't fit #61-#80 drills, so I have to do the drilling by rotating a pin vise with fingers.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.