Hello all
I have a distortion problem with an opamp Sallen Key high pass filter that drives me crazy!
The filter is a 2nd order high-pass (fc=400Hz, Q=0.71). It has a transformer at its input (transformer volume control, TVC). The set up is shown in the attached schematic.
The problem is that the signal coming out of the Sallen Key filter is heavily distorted (point C in the drawing). The distortion also appears at point B. I confirmed that the signal is clean at point A (see attached plots with 400 Hz sine signals at points A, B and C). The distortion is most apparent at low volume settings of the TVC, but less for higher volume settings.
I don’t understand why/how the signal gets distorted so badly, and how to fix this. Although I don’t have a lot of experience with opamps, I have a gut feeling that the problem is related to the complex impedances of the capacitors and the TVC in series with the opamp input (note that an otherwise similar low-pass filter with the capacitors in parallel to the opamp input does not show any distortion, see below). But maybe I am on the wrong track.
Does anyone have a clue what’s going on here? If anyone has suggestions or ideas to to debug and fix this, I’d be happy to hear them… HELP!
Here’s a bit more information about the setup:
Here’s what I’ve tried so far:
I have a distortion problem with an opamp Sallen Key high pass filter that drives me crazy!
The filter is a 2nd order high-pass (fc=400Hz, Q=0.71). It has a transformer at its input (transformer volume control, TVC). The set up is shown in the attached schematic.
The problem is that the signal coming out of the Sallen Key filter is heavily distorted (point C in the drawing). The distortion also appears at point B. I confirmed that the signal is clean at point A (see attached plots with 400 Hz sine signals at points A, B and C). The distortion is most apparent at low volume settings of the TVC, but less for higher volume settings.
I don’t understand why/how the signal gets distorted so badly, and how to fix this. Although I don’t have a lot of experience with opamps, I have a gut feeling that the problem is related to the complex impedances of the capacitors and the TVC in series with the opamp input (note that an otherwise similar low-pass filter with the capacitors in parallel to the opamp input does not show any distortion, see below). But maybe I am on the wrong track.
Does anyone have a clue what’s going on here? If anyone has suggestions or ideas to to debug and fix this, I’d be happy to hear them… HELP!
Here’s a bit more information about the setup:
- The audio source is a DDDAC digital/analog converter with a passive I/V conversion (load resistors RL convert the balanced current output to a balanced voltage signal).
- The balanced signal voltage (across the load resistors) goes to a Sowter 9335 TVC TVC for volume control. Note that the TVC actually has 27 taps on its secondary, but I’ve shown only 6 for simplicity.
- A 27 position switch connects the output of the TVC to the input of the Sallen Key high pass filter using RCA connectors/cable.
Here’s what I’ve tried so far:
- I found an NE5532 opamp in my junk box and tried this instead of the OP275. The distortion was still there, but it was a different shape.
- I inserted a series resistor (1kOhm) in between the TVC output and the filter input (between C1 and point A). This reduced the distortion a bit, but it also affected the transfer function of the filter.
- I replaced the standard RCA cable connecting the volume switch to the filter input with another cable I found in my junk box. This is a cheap and VERY long cable (5m or so), and it has “video cable” printed on it. With this cable, the distortion is reduced considerably.
- I replaced the DDDAC/TVC part with a signal generator (50 Ohm output). The distortion went away.
- I used a LOW-pass filter with the DDDAC/TVC. This is essentially the same as the high pass shown in the attachment, but with resistors in place of C1+C2 and capacitors in place of R1+R2. This does not show any distortion.
- I inserted a buffer stage in between the TVC output and the the filter input. This reduced the distortion, but not to zero.
Attachments
Try a buffer stage between trafo and filter. A follower will be OK.
Transformer have uncontrolable parasitic elements, resistances, inductances, capacitances and all are non linear in general speaking, which can combine with filter characteristic to produce undesired and unpredictable things.
Transformer have uncontrolable parasitic elements, resistances, inductances, capacitances and all are non linear in general speaking, which can combine with filter characteristic to produce undesired and unpredictable things.
I tried inserting a buffer (voltage follower) already, didn't help. I think that's pretty crazy, so I just tried it again. Still didn't help. In fact, the output of the buffer / voltage follower already shows some distortion. And then after filter, the distortion looks about the same as without the buffer. How on earth can this be?
Not sure I understand this properly.
- What do you mean by "loaded too heavy"? Too much current flow or power transfer through the transformer? Naaah...
- Do you mean the input impedance of the buffer / voltage follower? That's pretty high already (I'd guess in the Mega Ohm range).
I don´t think the high pass filter is distorting by itself but either the DAC does not like the complex impedance it shows or can´t drive it through the transformer volume "pot" which also shows a complex impedance , or both combined, as in the DAC might drive fine either the exotic "pot" or the highpass alone but not both together.
Excuse my minimalism, usually expressed in "keep it simple" , why don´t you just use a standard resistive volume pot?
Personally I´d avoid a tapped transformer volume control like a pest; I see tons more possible problems it can introduce than what it can solve.
Maybe it was a solution to some problems in 40´s vintage Cinema recording equipment, but later much simpler and better solutions were developed.
Just in case, test a buffer inserted between DAC and stepped transformer.
EDIT:
God only knows where will the output sit at, although quite possibly it will slam against one rail or the other.
If anything: "replace R2 with a higher value resistor, say, 100k"
This does not solve the problem, because the OP wants a highpass filter there, not a buffer.
Excuse my minimalism, usually expressed in "keep it simple" , why don´t you just use a standard resistive volume pot?
Personally I´d avoid a tapped transformer volume control like a pest; I see tons more possible problems it can introduce than what it can solve.
Maybe it was a solution to some problems in 40´s vintage Cinema recording equipment, but later much simpler and better solutions were developed.
Just in case, test a buffer inserted between DAC and stepped transformer.
EDIT:
Well,I can guess that it will distort horribly because IC1 will be unbiased, having lost the ground reference.
God only knows where will the output sit at, although quite possibly it will slam against one rail or the other.
If anything: "replace R2 with a higher value resistor, say, 100k"
This does not solve the problem, because the OP wants a highpass filter there, not a buffer.
Last edited:
Just tried with a 1k5 resistor, no change. Then tried with a 10 Ohm resistor (pretty low for my tastes), and distortion increased.
But then the signal coming of the DAC or the transformer would already show the distortion, wouldn't it?
(and it does not)
This point is well taken. I am trying hard to keep it as simple as possible. The transformer is needed to convert the balanced DAC output to unbiased. Adding a volume pot after the transformer does work, and I've used it like this before I went for active speakers that need the active filter. The volume pot would require a buffer stage before the filter.
I was hoping to avoid the pot and the buffer by replacing the (fixed-level) transformer by the TVC.
Ok, that's pretty much the alternative I described above. And I am pretty sure it would work (well). But I really don't want to give up on the TVC so easily, at least not as long as I don't understand why it doesn't work.
Yep. But I might try this just to understand and learn what's going on. But I have to make dinner for the kids and wife first...
Ah, change R2 for 100 kOhm. It will be a working high input impedance buffer to see what the output signal is.
If there is still distortion one can check other things. The OP wants a highpass filter there, not a buffer (except for this test) and also no advice to use a potentiometer. I think adding extra opamps is heresy with guys that use a TVC
Last edited:
In principle it should work, therefore a few questions:
What power supply is used for the opamp and are the supply pins properly decoupled with 100nF.
I see only a signal line going from left to right but no signal ground.
There should be a firm connection between the trafo earth and the earth on R2.
Be careful in mixing earth and signal ground, they are not the same.
Hans
Hans
What power supply is used for the opamp and are the supply pins properly decoupled with 100nF.
I see only a signal line going from left to right but no signal ground.
There should be a firm connection between the trafo earth and the earth on R2.
Be careful in mixing earth and signal ground, they are not the same.
Hans
Hans
Signal GND is drawn...
Let's just hope it is not lacking decoupling and PSU is +/- 9V or higher
Where's the signal GND, I don't see it.
I only see two earth symbols.
Only the output signal has a reference to R2 earth.
What type of caps did you use for the 100nF.
Hans
Think of the ground symbols being signal GND. They are connected....Draw an imaginary line between the symbols.
I think I understand the confusion. Here we draw PE/safety earth with that symbol and GND as just a thick stripe. Even when the right symbols were used they still would be connected.
I think I understand the confusion. Here we draw PE/safety earth with that symbol and GND as just a thick stripe. Even when the right symbols were used they still would be connected.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.