Yes, I’ll try with some that are a little higher. It does indeed stop working at the top.
That was me. It was also affected by the resonance. I sorted that out, though its hard to tell if it was my redesign or just building it again and not making the same mistake. @antoine.pasde2 recently asked me about that redesign and i will have to look at what i was thinking at the time. The past few days have bee rough so i havent had a chance to really look at it.
1 Like
Ah, phase reversal, our old friend…
Is it only when you crank up the resonance, or does it happen also at low resonance levels? What signal source are you using?
1 Like
I added on my filters an ATT potentiometer on the input signal, to be able to attenuate a little some signal (square for example)
So, what are good options? Use a different opamp? Limit the input with diodes? Keep using the 100k and add an amplifier at the output?
working on the performance filter and just came accross this issue
swapped out the 100k resistor for 10k resistor and it fixed the issue, but on LPF mode the input gain makes the filter shut down
thinking maybe trying out a 20k resistor instead to see if it can duck just below the panic level?
anyone found a better way of doing it?
1 Like
can confirm, just put 18k in place and seems to be fine
no audible signal drop between modes and stays working all the way to the top
5 Likes
for this problem i put a attenuator pot in my Input in all my filters, but i made them with stripboard and diy panel, if yours is the Perfomance one more diffucult with pcb and panel ready made (but allways possible i think)
1 Like
Hey is it normal for this filter to self oscillate? I used an oscope at the output of the filter, with no signal going into the input, and got something resembling a triangle wave at the output.
3 Likes
Oh, definitely. The resonance is quite severe.
3 Likes
yup yup. totally normal! 
1 Like
Awesome! and kinda unexpected 
2 Likes
I’m confused, does this only apply for the performance filter? I’ve been running into the same issue with the basic MS20 LPF. If the source is a square wave, the signal will drop out when the resonance is high.
a simple solution for this problem
couldn’t say for sure but i think that the performance filter IS the MS20 filter in a funkier configuration, so im not surprised of the same issue, seems like a weird one to keep repeating though
1 Like
People keep using the filter in various configurations, so not everyone is hitting the same issues; it’s perfectly possible to use it out of the box without any issues, especially if you’re used to attenuate your way to the right signal levels (which is something you usually need to do anyway in a full rig).
There are three level issues that people keep running into:
- The MS-20 filter has different response curves in low pass and high pass mode; in low pass, it’s two −6 dB/octave filters in a row so −12 dB/octave in total, while in highpass it’s one −6 dB/octave filter and the other stage is only used to inject resonance feedback.
- The MS-20 itself has one low-pass and one high-pass filter in sequence, so it’s not really an issue there, but here you have only half the filter, and if you flip the switch the levels will be off.
- The performance build has a much larger input resistor than the stripboard and other builds, which affects the signal level. You can swap this for lower values, but go too low and you run into the next issue sooner:
- The TL07x series isn’t a rail-to-rail opamp, and the input range (“common mode input voltage range” in the datasheet) is limited; if an input voltage goes below −8 V or so, the opamp misbehaves (this is called phase reversal). This is not a problem with inverting topologies, but the TL07x amplifiers in the filter chain are non-inverting voltage buffers, so the inputs will see the full signal voltage. Too much amplitude in or via resonance, and the filter locks up. To avoid this, you have to either manage the levels, or switch to a different opamp.
So as long as you have the right levels in and the right gain after it, the stock builds work just fine. If not, some tweaking might be needed (e.g. replacing the input 100k with 20k seems to bring up the levels to nicer but still reasonably safe levels) but you may still run into issues that are inherent in the MS-20 design and the component choices.
6 Likes
Any suggestions for opamps to switch to?
1 Like
Thanks, I’ll give this a try and see if that solves the issue
1 Like
I’ve used two TL072s for my builds, giving me the option to swap the one in the filter path for something else, but haven’t tested anything. Kassutronics uses an NE5532 but if you have quad opamp that won’t help you. I don’t think this circuit demands anything from the opamps, though, so something really basic like an LM324 might work.
(I’m a little bit tempted to experiment with using the built-in darlington stage, like in the original MS-20 design; all the filter buffers do is to convert a current to a voltage, while isolating the filter stages from each other, and if the darlingtons were good enough for Korg it might be good enough for the rest of us too )
7 Likes
Well, don’t let me stop you 
6 Likes