LMNC MS20 Filter Low Output

I used:

1000pf capacitor for the 1nf capacitor
.1uf capacitor for the 100nf capacitor
4700pf capacitor for the 4.7nf capacitor
.47uf capacitor for the 470nf capacitor

Those are correct, right?

2 Likes

sorry it’s 2 hours of the night in France, i go to sleep :sleeping:
look at this :
http://pencho.my.contact.bg/start/comp/uF_nF_pF.htm

2 Likes

Yes, capacitor values are correct.

1 Like

The important ones are 1 nF and 470 nF, and ideally you’d use polyester or similar there, but I guess quality multilayer ceramics might be ok (i.e. avoid no-name ceramics of unknown class/dielectric in the signal path).

If you have an oscilloscope, looking at the signal over the 1 nF caps while tweaking the cut-off knob might provide some clues (e.g. probe at pins 5 and 12 of the 13700).

As for gain stage, the filter consists of two RC lowpass stages (13700 + 1 nF cap), separated by voltage buffers (TL07x). In high-pass mode, the input signal is fed into the second stage via the capacitor, turning it into a CR high-pass filter instead. Whether the 1× buffers qualify as gain stages is a matter of definitions, I guess :smiley:

3 Likes

Thanks fredrik! I used ceramic capacitors instead of polyester on the 1nf and 470nf…The 1nf is a Vishay and the 470nf is a TDK. Do you think that would make that big of a difference?

I also just double checked the resistors and all of the values are correct.

I’m starting to think it has to be a soldering error somewhere. It’s the only other thing I can think of. I’ll have to go over the board with a magnifying glass because everything looks good from just my eyes so far!

Also, just to confirm since there were conflicting answers…what exactly should the LEDs be doing? Should they BOTH be lighting up? Should only one or the other be lighting up based on I/O? In Sam’s video, it looks like only the OUT LED is lighting up when the filter is working?

it’s not a led out, it’s 2 led who clipping when resonance is high
like you said check your solder under the stripboard .

Vishay and TDK are quality brands so at least no no-name stuff there. How worried you should be depends on what class and dielectric type you’re using, but I’m 99% sure you’re fine.

If you don’t have a scope, I’d check the circuitry on the way from vc in and the cut-off pot to the 13700s next. If you have issues there the 13700s (acting like variable resistors) might be stuck at high resistance.

3 Likes

Ok, then I guess they’re both doing what they’re supposed to because they both light up nice and bright when the resonance is high.

1 Like

Thanks fredrik! I’ll check that…and thanks for the reassurance on the caps!

1 Like

If the filter bit is working, you can hear it dampen the frequencies if you change the cut off frequency or hear the resonance change if you turn its control, then logically speaking:

1: the input signal is too small or attenuated in an unwanted way resulting in a low output

or

2: the output signal is too small or attenuated in an unwanted way resulting in a low output

If you have an oscilloscope, obviously that would be a good way to trace the input signal to the output and see what is happening to its levels.

If you don’t have one, what you could try is bypass the input stage or bypass the output stage and see whether that helps. What I mean is inject the input signal directly in the filter part (using a capacitor in series with the input signal to get rid of any DC voltage) and see whether this makes the output signal higher and from the diagram try to figure out whether this is to be expected or not. And/or use a capacitor and an external amplifier and connect it to various parts of the filter to find whether there are locations with higher signal voltages than the official output carries.

1 Like

So after a local oops :smiley: that required a bit of debugging, it struck me that a quick way to check the CV converter is to measure the bias current (aka “Iabc”):

  • set your multimeter to microamperes (µA) if it supports that (otherwise mA might work, if it can measure with at least 2 decimals).
  • power off the board, unplug CV in, and set the cut-off potentiometer to somewhere in the middle
  • carefully remove the 13700 from its socket
  • power up the board, and measure the current between pin 1 (and/or 16) and ground, while sweeping the cut-off pot between the end points

You should see currents between ~0 and 60-80 µA. If you see no movement, or you’re getting a lot more (more than 1-2 mA or so) something might be broken. Very low currents here might also affect the overall output volume afaict (I should do some calculations to figure out what an acceptable minimum bias is).

2 Likes

Good point something could be sinking current.

After finally finishing my “inspired by LMNC” MS-20 build and fixing a couple of PCB glitches, I was a bit puzzled that my notes said 10k for the input resistor, but my schematics had a 100k there. Trying to figure out where I got these values from, I checked some other schematics, and it seems like the stripboard layout uses a 10k resistor on the input (as does Schmitz), while the performance layout uses 100k.

These resistors form a voltage divider with a 220 ohm resistor, so for a full 10 Vpp in you end up with either 220 mVpp or 22 mVpp (ignoring the feedback loop). The former is a bit high (not that you’ll damage anything, but you’ll get plenty of distortion with the full 10 Vpp in (EDIT: and you also risk lockup, see below) unless you attenuate the signal, but then distortion isn’t necessarily bad in a gnarly filter). The latter is really low, though.

So if you’re getting a low output, it could just be that you have too much attenuation on the way in – check if you have 100k resistors there, and consider replacing them with a lower value if you do. You could just solder another 100k in parallel with the existing one to check if that improves things.

EDIT: Tested with a 100k and the output level dropped through the floor. I don’t trust my attempts to simulate/calculate, so will have to get out the scope later, and measure some measurements.

6 Likes

I changed my inputs to 10K and all my troubles are over. It’s nice and loud and the pots give me the attenuation if I need it. Sounds much nicer too, I think, with a bit of saturation.

3 Likes

Hey Christian! Are you talking about changing the 100k potentiometers to 10k?

1 Like

No, the 100k resistors in series with the input pots on the PCB version.

2 Likes

it’s maybe an error on Performance version because it’s 10k on the Schmitz website
and also on the simple one Sam stripboard

https://www.schmitzbits.de/ms20.html

1 Like

(yeah, that’s what I said four posts up :sunglasses: )

4 Likes

Do you notice any issues with the 10k replacements with too loud inputs? I remember a Patreon post where it was said that the Filter stops working when the input is too loud/high…

1 Like