[Aion "Luna" Tremulus Lune] Help debugging a possibly broken NSL-32

[AdamFenwickSymes]

Hi guys,

I've built an Tremulus Lune from Aion's PCB, (schematic here: https://www.dropbox.com/s/78vm5lrt7rxt3h9/aion-luna-tremulus-lune-documentation.pdf?dl=1) and I'm having some problems with it. I'm using an NSL32 instead of a VTL5C2, which the docs say is fine.

I built the circuit and alligator-clipped an in and out cable to it, and it worked great! But it suddenly the audio cut out, and it hasn't worked since. The rate LED still flashes perfectly, so I guess somehow I broke a component in the audio path, but I'm not sure how. The audio path is not very complicated, so it shouldn't be too hard to find ...

When I clip the output cable to the input side of the photoresistor the signal comes through fine, great! When I clip the output cable to the other leg, I get complete silence. So I guess the NSL32 is broken? Maybe I put too much current through the LED side, but I'm not sure how, since I built it exactly according to the specs.

When I measure the resistance of the photoresistor with the circuit unpowered it's apparently higher than 20 mega ohms (my highest setting). When I measure the resistance with the circuit turned on it seems to fluctuate 400-600 ohms, but I feel like measuring resistance while the circuit is on is not allowed.

I guess the alternative is that the signal is just extremely quiet after the photo resistor and the subsequent amplification stage is not working. I thought about sending the input in on the out leg of the NSL32 to check the rest of the signal path, but I'm a bit reluctant to randomly send the signal through random bits of the circuit in case I break something.

Sorry to post a generic "help me debug" post. Some concrete questions:

- how can I tell if the NSL32 is broken?
- how on earth did I break it? Since they are kinda expensive and I don't want to break another one.

Many thanks!

[bean]

You can use the NSL32 however you should make these adjustments: make the gain trimmer 5k and C4 1n5. The on resistance of the NSL 32 is much much lower than the 5C3 (a few hundred ohms vs 24k or so). So, the adjustments are to compensate for the needed gain recovery on the circuit.

For your NSL32 are you sure you attached the leads correctly? The LED cathode is the lead with the grey dot which connects to the junction of R13/R14 on his schematic.

[AdamFenwickSymes]

Thanks for your reply.

Just double checked the orientation of the NSL32 - the grey dot side goes to the junction of R13 and R14 like you said. But it worked initially so I don't think it could be a "put it in the wrong way" issue anyway.

I have a 25k gain pot, which I guess is not ideal, but it shouldn't do any harm, right?

My C4 is a 330pF, since the schematic doesn't say anything about changing it for a NSL32. I do have C2 jumpered though, since the schematic said it will turn into a high pass filter otherwise.

Any idea how to test the NSL32 without removing it from the circuit?

[Andrekp]

Just by coincidence, I am currently building a Shoot The Moon/Trmulus Lune hybrid.  It sat on my breadboard for a while while I played with getting it right.  I used a home-rolled "vactrol," so maybe what I learned is useful to you.

It sounds like, if I understood you, the vactrol is working fine because you get varying resistance.  You probably just need to balance things out a bit.  Or you can roll-your-own vactrol.

The key to this thing is getting the balance between the LED, LDR, the two LED resistors (R13 & R14) and the gain pot.  It really does not seem otherwise all that picky about the LDR, so long as you get a light resistance of maybe a couple K or more, and a dark that is significantly higher.  You can balance the rest to work with that.

I made my vactrol with a 3mm red LED and a GL5537 (I'm pretty sure) LDR.  I tried a few and they measured all over the map, for whatever reason, so I picked one and played with it as below.

What I did initially is replace the vactrol LED with just a spare LED on its own.  Then I adjusted the values of R13 & R14 so that the light was not so bright.  I made it so the light was just about solidly on with the depth pot at minimum depth, and so that it started actually pulsing with just a tiny turning of that pot.  With the stock 1K, there was just too much brightness and the depth pot didn't really have effect for the first bit of turning, and the LED didn't really get very dark on the lows.  I settled on about 12K for both resistors.  Remember, looking ahead in the process, you will want to end up with maybe 3-10K as your low resistance on the vactrol.  You may need to readjust the resistor values to get that as well.  YMMV.

EDIT:  I also upped that low ohms resistor in the depth circuit.  I think I settled on 560R.  That seemed to adjust the full on brightness range on the depth pot.  With the stock value, the LED stayed fully on for the first fifth of a turn or so on the depth pot.  I think I made that adjustment before adjusting the othe two resistors to get the overall curve I wanted.

Step two was putting the LED side of the vactrol in circuit and measuring across the legs of the LDR to get an idea of what the resistance was.  You need to set depth the minimum and speed to maximum so that you get the LED in the vactrol to be mostly entirely on.  As on as it will be.  In my case, it measured about 4K with the light on.  I think it's better to do this in circuit, as opposed to just wiring it up to 9 volts, to see what you are actually getting.  dark resistance will be very high.  Once you have the numbers, put the LDR side in place.

The theory of this thing would then suggest a 10K gain pot, and that probably would work fine.  But what I found was that when the depth was very deep, the output volume dropped quite low.  Just a result of the curve of my home made vactrol probably.  So I upped the gain pot to a 25K.  This gives plenty of gain for anywhere I need to have it.  Use whatever you need to have usable output volume when the depth is deep and the speed is slow. 

So your vactrol may be fine, but if it's not, you can experiment with rolling your own.  I don't think it's particularly picky once you get everything balanced out correctly.  Otherwise, you might just need to start playing with some values as above to get things how they should be.

Good luck!

FYI I also added a spacing pot, which I think the TL comes with. 

I found that with the wave shape pot, that moves from sine to square wave, it might just as well be a switch to going from one extreme to the other.  No need for a pot.  But when I was experimenting, I found the shapes were better with a 180K in place of the pot, rather than a 500K, which is the pot value.  FWIW, the sine was exactly the same, but the square side became actually square, so I went with that. (180K instead of 500)  So if you stick with the pot, you may find that a 200K or so works well.