EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: RawCode on October 09, 2018, 05:22:17 am
-
Hi guys.
I've got a big problem. And it's very URGENT(it's the project for my exam).
I need a high pass filter at the output of a differential amplifier that cuts off the DC voltage of this stage. So I designed the filter with these parameters: R=1Meg C=100nF, with cut frequency around 6Hz. I need also a buffer stage that "isolate" what the capacitor can see, because after the buffer I have the classic inverting mixer. As opamp i'm using the TL074.
Here come the problems. I don't understand why, but the buffer stage introduces a strange heavy distortion: if i put a probe from my audio interface to the non inverting input, the sound is perfect, but if i put the probe to the opamp's output there is this a strange distortion. The voltage at the output is around 0V, so it's not clipping to the power rail.
Can you help me please? It's very urgent.
-
There are 14 pins on a TL074, you are showing three. What are the rest doing?
Tim
-
Single or dual power rails?
And where is the output connection? Are you blocking DC on that with a cap?
-
The other opamps are doing the same thing with 3 other signal source.
I'm using dual power rail (+-9V).
The output is connected to a inverting mixer, so there isn't any other capacitor
-
Hmm. It’s a long shot, but are you using a chip ceramic 100nF cap? A lot of those are inherently non-linear. Maybe replace it with a film cap? If you’re under the gun, back-to-back electrolytics might work (I.e. one positive, one negative).
-
Is it a distortion or an oscillation/noise? Do you still get it if both inputs are disabled?
If it is oscillation you can try and put some small caps between the outputs and the inverting inputs of the first and the third opamps and see if it stabilizes.
-
I'm using the classic green caps, i think they are film caps. I'll try to reduce the resistor to 100k
I think that the distortion( i think is a distortion non an oscillation, but i have to verify that) is introduced by the op-amp, because the distortio i hear is after the opamp
-
Have you tried load resistors on the output of each stage? Since every output just goes into a very high impedance, you may be running to issues when the output current approaches zero - if you describe the distortion it may help verify that.
The datasheet tests at 2k and 10k ohm loads, so maybe try a 10k resistor to ground on the output of each stage and see if it improves?
-
It's hard to explain it, but it's like there is a little cut in the waveform, but it's not clipping to the power rail
-
Ps: The output feeds an inverting mixer
-
The TL074 is not a rail to rail opamp and it will clip at 1.5-2V bellow supply voltage. Could you post a screenshot of your scope on the output? Maybe it could help
-
Yeah, should have been there along with the first post.
-
The other opamps are doing the same thing with 3 other signal source.
?
I'm using dual power rail (+-9V).
One thing that fits all the evidence (LOL), is the TL074 not being powered at all.
-
Looking at you schematic of the final output stage, as the trimpot value shrinks, that op-amp's gain can go haywire making noise throughout all the other opamps in the package. That negative inverting gain can drive the differential negative input to hell the way you have it wired as the trimpot approaches 0 ohms.
Also, having a big conducting mass like a trimpot wiper directly tied to a negative input on an opamp, without some series resistor at the op-amps input also is asking for an output with additional hiss and EMI sensitivity.
-
The first opamp has a very low value (1k) for its negative feedback resistor that is actually its output load resistor and is overloading the output of the opamp at high output levels. Why not use 10k or 100k for all the resistors on the first opamp?
The second opamp is also overloaded by the very low value (1k or less) of the gain pot.
The third opamp also has a very low value for its negative feedback resistor that overloads it.
-
Just show us the waveforms.
If the signal is OK up to the output buffer and clips at its output, then you're either not powering it properly or you're just making it saturate. As said above, it's not a rail-to-rail opamp and needs a comfortable margin between the output amplitude and the power rails.
-
To build on Brian's post, the output stage has a few problems:
1.) the voltage gain is, in the limit, the ratio of the feedback resistor divided by the trimpot resistance setting. As the trimpot resistance is lowered, the gain goes up and is ultimately limited by the gain of the opamp itself.
2.) the load impedance presented to the second stage is essentially the trimpot resistance because the imedance at the '-' input of the third stage is zero ie., it's a virtual ground. At low trimpot settings the second stage can limit current and distort the signal.
Audioguru has pointed out that the resistors are quite low in value. I'd look into increasing them to ease the load on the opamps.
Cheers,
-
Oh, yeah. Didn't see the second, whole schematic.
Couple suggestions:
1. Learn to debug your circuits. Isolate causes. You kind of did that looking at the signal before and after the opamp buffer (if I got it right), but you probably didn't disconnect the opamp buffer from the output stage. So do that first to isolate the problem. Disconnect the buffer's output from the potentiometer. If the signal is not distorted, then you know you have a problem with the output stage.
2. You shouldn't implement what I take is a volume control by a direct gain control of an opamp IMO. Especially with a mechanical pot. In this config, it's bound to quickly make all kinds of noises while you're rotating it... besides, in this config (inverter topology), you're not going to get a zero-volume. Don't know whether it's purposefully or not. And what do you think happens as the wiper's resistance goes down to zero (res. beween the pot's wiper and the end connected to the opamp's output)? Compute the inverter's gain for the two extreme positions of the pot.
3. If it's for volume control, just use like a 10k pot in a classic attenuator configuration, and the wiper connected to the input of the last opamp configured as just a buffer.
-
And it's very URGENT(it's the project for my exam).
....
Can you help me please? It's very urgent.
What was the urgent solution?