Problem triggering soundlab with midi -> cv circuit

[Lemmy]

Hi all,
I'm building a midi -> CV circuit into my soundlab. I've got the CV and gate hooked up, and I can hear the pitch changing as I play the keyboard.
I can't get the external gate to trigger the envelope though.
I'm measuring about 4.3v at the external gate every time I press a key on my midi keyboard. Shouldn't that be enough to fire the AR envelope?

I tried changing Q8, but it still doesn't work. I'll poke about with my multimeter, but any ideas on how to troubleshoot would be welcome!

Thanks.

[Pehr]

4.3V should be enough. Have you set the AR-times to minimum?

[Lemmy]

Hi Pehr,
Thank for that - I got it working. The problem was that my S&H circuit was connected to the external gate.
I put in a pot to control the output level. (See picture. I don't know if this is the best way to do put in an output level control by the way??). But when there is a connection to the junction of D7 and R199 the external gate doesn't fire.


I notice that Ray has now adjusted his S&H diagram to put a switch where my pot is. I guess I have to choose between S&H and external gate - I can't have both?

Maybe I could replace my output pot with a pot which has an off switch at 0? Or maybe I could use my S&H rotating input selector somehow so that when "None" is selected, the circuit is switched out?

Hmmm, thinks....

---

SampleAndHoldOutput.gif
Description:
Filesize:	10.8 KB
Viewed:	8263 Time(s)

---

[Pehr]

[Lemmy]

I tried that. But if i remember rightly, then the output just behaved like the rate pot, changing the S&H rate.

EDIT - maybe we are speaking at cross purposes? The output pot above is part of my S&H circuit, it has nothing to do with the midi->CV or external CV in.

[Pehr]

[Lemmy]

Hi Pehr,
I just wanted to have a way to control how much the S&H was used in the soundlab. So I designed my front panel to include an output level before I really knew how the S&H would behave. I thought I could sort of mix the S&H effect in to the existing sounds using the output pot
Now I know it doesn't quite work like that.
So I'm experimenting with uses for the output control. How would you suggest hooking it up? (So that it doesn't just work as another rate control?)
I have noticed other builders have output pots for their S&H units.

[Lemmy]

And also, getting back to the original issue, I suppose I'm going to have to switch off the S&H connection to D7 whenever I want to use midi in or external CV in?

[Lemmy]

By the way, I got my portamento working! I adapted it from Ray's sequencer, as Pehr suggested in another thread.
I mounted a spare LF444 on the kludge area of the soundlab PCB. (And just used half of it.)
Works great with my midi-cv circuit!

---

portamento.jpg
Description:
Filesize:	15.62 KB
Viewed:	8204 Time(s)

---

[Pehr]

Great that you got the portamento working! Remember if you want longer portamento times, just increase the value of the cap

0k, but if you want to control the amount of S&H you should have the attenuating pot on the CV signal in instead of the trigger.

And also, if you want to control the oscillators, you should have it attenuate towards ground and not -9V.

However, for the VCA and VCF it should attenuate to -9V, so you might want a switch to choose either.

Good luck!

[Lemmy]

Cheers Pehr, the output control is more useful like that.

Yet another question for you or anyone else who might know.
My midi->cv circuit has an auxiliary CV output. I thought I would hook that up to filter cutoff, so I can control that from my keyboard's mod wheel. It looks to me that this connection should go to the middle pin of R37.
Problem is, the voltage range of the cutoff control seems to be -9 -> +9, more or less. The range of my aux CV is 0 -> 10.
So I can only sweep the upper part of the cutoff range. Is there any way I could modify the aux CV range to fit the cutoff range?
And, is it dangerous for the soundlab to put more than 9v into the circuit. e.g. 10v?

Thanks again!

[Pehr]

You could mix the aux CV with -5V and then amplify it by 1.8. That would give you a sweep of -9V to +9V.

[Lemmy]

Aha, you make that sound easy!
Well I'm reading through some excellent Forrest Mimms books at the moment, so maybe that will give me some ideas on how to achieve such a thing.

[Lemmy]

Here's a good page I found with some info on this kind of very basic but very useful stuff!
Link

[Lemmy]

Well I got round to looking at this. (Bear in mind I'm learning this as I go along. I was quite pleased with myself to get this far.)

I measured the range of my filter cutoff again - it's about -8.13v -> 6.13v

My midi circuit's Aux CV range is 0 - 10v

The pics below show my attempt to convert the 0 - 10 range to something useable. I change the gain of the Aux CV out from x1 to x1.46 by changing the resistor values.

Then I use a couple of op amps to subtract 8.18v.
In theory I end up with a range of -8.18 -> 6.42v.

Questions for any experts out there...

1. Does it look reasonable? Would it work? I would be using half of an LF444 for the subtraction part.

2. Is there a simpler way to do it? Could I just modify the original circuit somehow? It's by Marc Bareille and you can see it here.

3. Do I need a resistor at R7? In my Forrest Mims book he has a resistor there, value = (R5xR6)/(R5+R6).

4. In this document it says

[peterw]

Hi Lemmy,

If you are learning this for the first time, I think you are doing pretty well.

The voltage converting circuit you have come up with and the 1.46 gain modification look good on paper but one thing is bothering me. The opamps are being powered by 15 volts I assume. If this is the case you will have to select an opamp which will cope with an output swing of +/-14.6 volts. Most will not swing that close to the supply rails.

One way around this is to leave the original Marc Bareille circuit as per original and change the 10K resistor connected to input V1 on your green drawing to about 6.7K. It may be best to select the exact value during the final testing and setting up stage. This way none of the opamps will be expected to swing too close to the supply rails.

As for connecting directly to the wiper of the cut off pot, that is bad practice because if the pot is set fully clockwise or anti clockwise you may be shorting out the opamp output or inadvertently connecting two opamp outputs together. That may damage an opamp. It may be best to sum the wiper voltage with your new control voltage before sending it to the rest of the filter circuit or install a break jack so that plugging in the external CV automatically disconnects the pot wiper voltage. I have not looked at the circuit of the filter so I am guessing at this last part. You could use the second opamp in the Green circuit to sum in the variable pot control voltage by connecting it the +ve input of that opamp. I bet you can work out how to do that.

You don’t really need a resistor at R7.

Don’t worry about noise etc in this circuit

Sorry about the length of time it has taken me to post this, I am a very slow typer.

[peterw]

Hmmm, I just looked at the sound lab schematics. To connect your new source of CV to the filter cutoff you could just duplicate the LFO input components S8 and R31 say S8a and R31a so that one end of R31a connects to R36 the same as R31. Input your new CV to the far end of S8a. R28, R31 and R36 form a summing circuit to inputs 1 and 16 of IC 14.

[peterw]

One last observation, the resistor values in the green diagram are a bit low. I would add a nought to the value of each one. That will bring them more in line with typical resister values used in CV processing modules. It will reduce the strain on the opamps.

[Lemmy]

Hi Peterw,

Thanks very much, your advice is really useful. I'll need to sit down with a cup of coffee and absorb it properly.

A couple of notes - the power supply is +/- 12 in the "amplifier circuit" and +/-9 in the "summing circuit". ( I built a dual power supply before realising I could power the Soundlab at +/- 12 )

If I change the 10K resistor at V1 to 6.7K will that give me the gain required? If so that would simplify things a lot.

I see what you're saying with S8a R31a etc. Here is where I get stupid though. I'm not clear on what happens in this summing circuit if there is a high voltage at my Aux CV in and already a high voltage at R36. Say 6v at one and 6v at the other. I don't end up with 12v at IC4 pins 1 and 16 do I? (That would probalby fry something?) And -6v and 6v don't cancel each other out? I'm guessing this is a job for Ohm's Law...

The Aux CV in is midi controlled, so I can't easily switch it using a jack. It will be controlled by my keyboard mod wheel, so when not in use it will sit at the bottom of the range, i.e. -8.xx
So if that is always switched in to the summing circuit above, will it affect the filter cutoff, even when it's always at -8.xx?

I'm thinking I could have a three way switch next to my midi in socket, which is:
1. midi note in
2. midi note in && midi filter control (aux cv)
3. cv in.

That way I can always use midi note in, but can choose when to use the mod-wheel/midi software to control filter cutoff, OR the soundlab pot.

thanks again,
Lemmy

[peterw]

Hi Lemmy,

If the power supply is +/- 12 in the "amplifier circuit" you will not be able get 14.6 volts output by modifying the original circuit to amplify by 1.46. Put it back as it was i.e so it gives 0 to 10 volts out.

In the “summing circuit” (Green drawing) changing the 10K resistor at V1 to 6.7K will give you the gain required. I would actually use 67K and make all the 10k resistors 100K and the two 1Ks connected to the 1458 pin 6 100k as well. The other two 1k resistors can stay at 1k.

The opamps in the “summing circuit” need a supply high enough to cope with the 8.18v max swing you will get in that circuit. +/-9 volts may be enough with better opamps (maybe CA3130 http://www.intersil.com/data/fn/fn817.pdf or possibly just try the LF442 or 4 http://cache.national.com/ds/LF/LF442.pdf – same as the rest of the Soundlab. You may need to check the data sheets though). The Marc Bareille circuit shows a +/-15 volt supply but +/- 12 should be ok.

But...

You are right to be concerned about frying the LM13700. The circuitry feeding into pins 1 and 16 of IC4 is not a true summing circuit as such. You must ensure that the max possible current to those pins does not exceed the Absolute maximum for the LM13700 (see the LM13700 Data sheet http://cache.national.com/ds/LM/LM13700.pdf Absolute maximum amplifier bias current is 2ma). You should never get close to that with the circuit you have because R28, 31 and 36 limit the current to a max of approximately 9v/50k = 180ua worse case. Adding R31a etc will not change that much.

As this is not a true summing circuit all the controls connected to it (Frequency cut off , AR level and LFO level) will interact to some extent. I see that the AR level and LFO level controls output -9v when turned right down. I think you should arrange your “level converter” circuit to do the same when the mod wheel is right down. -8.18volt looks about right as you have already said.

There does not appear to be a Pitch CV input to the VCF ? (odd)

“ I see what you're saying with S8a R31a etc. Here is where I get stupid though. I'm not clear on what happens in this summing circuit if there is a high voltage at my Aux CV in and already a high voltage at R36. Say 6v at one and 6v at the other. I don't end up with 12v at IC4 pins 1 and 16 do I? (That would probably fry something?) And -6v and 6v don't cancel each other out? I'm guessing this is a job for Ohm's Law... “

Yes it’s a job for Ohms law. It’s about the current flowing into and out of the common point, pins 1 and 16 and the current flowing into the pins more than about the voltage at the pins. (its not a true summing circuit). Controls will interact I think.

“ The Aux CV in is midi controlled, so I can't easily switch it using a jack. It will be controlled by my keyboard mod wheel, so when not in use it will sit at the bottom of the range, i.e. -8.xx
So if that is always switched in to the summing circuit above, will it affect the filter cutoff, even when it's always at -8.xx? “

Ah I see – Yes it may but you will compensate by adjusting the cut off control.

[Lemmy]

[Lemmy]

I tried breadboarding this. The results aren't quite what I expected.

The bottom line is that the maximum range I can produce from my "summing circuit" with 0 -> 10v (well really 0 -> 9.7) at V1 and -9v at V2 is -7.1 -> 7v

I'm guessing that this is down to the characteristics of the LF444, which is the op amp I'm using. Is the output voltage swing the relevant measurement? On the datasheet I found the typical LF444 ovs is +/-13. Yet I'm only getting +/-7. Hmm.

It seems also that I can't offset the voltage range further into the negative, say -8 -> 6. The lowest voltage I can get is -7, the highest 7.

That's pretty useable, but I'd be interested to know why I'm getting this behaviour. I also tried -12 at V2 with the same results.

I'm also interested in how peterw calculated the gain in the summing circuit!

Thanks anyone with suggestions.

[peterw]

[peterw]

Ah, I just noticed something.

You have calculated V2 as being R1/(R1+R2) * -9 = -8.18. That is not right because the 10K input resistor between the junction of R1, R2 and pin 2 of the 1458 is in parallel with R2. So you only have -6.92 volts instead of -8.18.

OK….

8.18/10,000 = 0.818mA

So, the input resistor for V2 needs to be 9/0.818 = 11,002 Ohms

Just remove R1 and change R2 to 1k then put the -9v into that.


P.S. You will still have to increase the supply voltage to +/- 12v get the output swing you need. If this means that V2 is derived from the -ve 12v supply then remove R1 and replace R2 with 4.670k (i.e. 14.670K total) to maintain the 0.818mA current as calculated above.

[Lemmy]

Brilliant explanations Peter! And you are right about me using +/-9 v to power the LF444.

After my last post I tried hooking the circuit up to the filter cutoff through a 200K resistor, as with the other filter inputs. After a bit of experimentation with changing the gain to a higher value, I was getting good results, and could sweep pretty much the whole range that my pot does.
I put a 47K resistor in instead of the 67K (I used 10x all the resistor vals as you suggested) to do that.

I can see now that was partly compensating for the problems you pointed out.

I'll go back and do some more experiments with R1 and R2 as you suggest, and maybe switch to the +/-12v power supply too. There doesn't seem to be much more audio signal at the bottom of the range, but there might be at the top.

(I might have another poke around at the resonance pot too, as it doesn't seem to have much effect except at the top of the range. Seems like this is a feature of the soundlab though as others report the same thing.)

Thanks for explaining it so clearly, that will help me a lot for future circuit
building.

I'll upload a finished circuit diagram here when I get the thing done, as Marc's midi->CV circuit goes well with the soundlab, so it might be useful for other people...