Hi guys, I’ve just finished building my midimuso cv-12 for my eurorack system, and just plugged it in and ran into the problem of the 10.6V test point only going up to about 10.43 volts. The PSU I have (4ms row power 45) doesn’t have a trimpot to increase its output voltage, but because -12V isn’t used as much (and the midimuso draws lots of power), I was wondering if it’s possible to take -12V rail and use an op-amp to “convert” it to +12V for the midimuso. If it was less than 12V I see no reason why it shouldn’t work, but given TL072’s take in +/- 12V I don’t know if it would be capable of flipping the voltage by that much. I also think I read somewhere that you can replace a resistor on the midimuso with a slightly lower value and that will get it up to 10.6V, but if that could be confirmed that’d be great.
Even with a full +12 V on the power rail you can’t get +12 V out of a TL07x. And with less than +12 V on the rail, you’d max out at less than that. Nor would it supply enough current to act as a power rail.
Ok, I thought that might be the case. Do you reckon lowering a resistor value somewhere would work?
From looking at the schematic I think that lowering the value of R1 (the 220R resistor), I should be able to get closer to 10.6V from my PSU. This should work, right?
There’s some discussion here
about using 270R or 220R, nothing about going lower but it sounds like it might be worth a try in your case.
To invert a power rail you use a charge pump, not an op amp.
Typically they’re used to invert a positive rail, I have no idea if they work the other way around though, you might have to connect the pins differently for that. Be sure to get one that doesn’t switch at an audio frequency.
Looks like that thread is about Sam’s version of the midimuso. I’m using the CV-12 kit from midimuso themselves. Here’s the manual with the schematic, I’m looking at changing R1 to a 150R instead of 220R (that’s the next lowest value I have).
With R2 (the 5k trimpot) basically completely “off”, I’m getting 10.41V at the 10.6V test point, so lowering the value of R1 should get me a bit higher right?
The original midimuso is broken by design; see discussion here:
TL;DR the LM317 isn’t guaranteed to work with 12 V as used in that circuit, and odds are it’s not even regulating; you’re just seeing the supply voltage minus a collector-emitter voltage drop across the internal darlington transistor.
You may get around it by trying different LM317s, but the only robust approach is to increase the supply voltage (which means changing the supply if the one you have is too good).
In a quick test, I needed 12.25 V in to maintain regulation at 10.60 V out:
but the exact drop depends on the current; the datasheet suggests a 3 V difference for the regulator to work properly across the full range.
Ok, I think I understand that. Two possibly stupid questions:
Is it possible to swap the LM317 with something that’s a bit more capable of getting 10.6V from 12V, if so what would you recommend.
Otherwise, would lowering R1’s value not increase the voltage going to the LM317, even just by a slight amount?
Many thanks for your help!
Assuming it has enough input voltage to work with, the LM317 wants 1.25 V on the ADJ pin between the OUT pin and ADJ when the OUT is what it’s supposed to be. R1 and R2 forms a voltage divider that lets you dial in that voltage. To hit 1.25 V, you need the trimmer’s wiper at around (220+5000)×1.25/10.6 = 615 ohm away from ground (~12% rotation), so lowering/removing the resistor at the other end won’t help. No, I used the wrong reference in that calculation, it’s more like 4600 ohm (see below), but changing R1 still won’t help. You cannot get around the voltage drop.
There are low dropout (LDO) regulators that might work here, but the TL431 circuit Sam ended up using is inexpensive and a lot better; roughly this, but perhaps with 220 ohm instead of 270 to provide a bit more headroom.
(more here)
Ok, so would this mean replacing the LM317 with a TL431 or would I have to add to the circuitry? I’d assume I’ll have to replace the LM317 with a TL431, then adjust the resistor values a bit.
Would I be able to do all of this on the midimuso CV-12 PCB or will I have to mess with extra bits of stripboard?
Through hole footprints are different, the LM317 is TO-220 and the TL431 is DIP-8. Connections are different too, LM317 connects directly to the 12 V rail and not to ground, TL431 connects to the rail via an additional resistor and to ground.
Right, so basically if I want to get closer tracking I’m gonna need to make the circuit in that image on stripboard, then connect it to the midimuso PCB. With how I’ve made my midi-cv module it barely fits inside my case, so I’m not sure if an extra bit of stripboard would be viable.
Just tested the voltages I’m getting from the LM317. When I set the trimmer so the adjust pin gets 1.25V as mentioned earlier, out pin is sending 2.58V which obviously isn’t enough. When I set the trimmer as far as it will go, the out pin sends out around 10.43V, with the adjust pin getting 9.4V. I presume you meant the adjust pin needs 10.25V rather than 1.25V. Either that or I’m just getting confused and have dodgy components…
Yeah, seems I misread my earlier post 
ADJ should be 1.25 V below OUT, not 1.25 V above ground. I should have measured in circuit in my supply instead 
This means that you want the midimuso potentiometer to be near (220+5000)×(10.6-1.25)/10.6 = 4600 ohm (or 4410 ohm if you short out R1) which is well within range, but only works if the input voltage is high enough.
When I set the trimmer as far as it will go, the out pin sends out around 10.43V
12 − 10.43 = 1.57 V which is close to the 1.65 V drop I measured in circuit (with a ~12 mA load). Your load might be lower, the input voltage very slightly above 12 V, or our LM317s differ slightly.
You can do it freeform, probably in less space than the TO-220.
- Remove the LM317.
- Add a 220 ohm resistor from where LM317’s pin 3 was on the PCB (IN) to pin 2 (OUT).
- Connect the TL431’s cathode to pin 2 (OUT).
- Connect the TL431’s anode to a nearby GROUND (e.g. the one on the trimmer).
- Connect the TL431’s reference pin to LM317’s pin 1 (ADJ) or directly to the trimmer’s wiper.
- Change R1 to 5.1k. Leave R2 as is, and use it to dial in 10.60 V.
Alternatively, contact the midimuso folks and see if they have some suggestions for an drop-in LDO that works with 12 V. IIRC there was a post here earlier where someone had contacted them, but cannot find it right now.
EDIT: Something like this:
My psu gets just below 12V, I’ll check tomorrow but IIRC it’s something like 11.8V give or take a few 0.1’s.
What do you mean about the load? I might be inferring the wrong information but I read that as if having less load on the psu means the midimuso will operate further from the desired 10.6V. I do only have a couple of modules currently as I’m just starting, and doing as much DIY as I can. Would the V-Out pin increase in voltage when the PSY has more load? That’s how I inferred your message, but that seems counterintuitive to me. Still fairly new to electronics!
Oh, there are many voltages here. I’m talking about the voltage drop across the LM317 (IN−OUT) when it’s not regulating (i.e. not producing the expected output, since it doesn’t have enough input voltage to work with). This drop depends on the current through it – the more current, the bigger the voltage drop. This is why they recommend a 3 V margin in the datasheet to make sure it can regulate also at the full 1.5 A, but we’re seeing 1.4−1.6 V drops with low loads (a few milliamps).
(when it’s actually regulating, the voltage drop will of course be larger)
The power supply may indeed drop its output voltage (=the input voltage to the LM317) with higher overall load, but you’re already below the voltage needed to produce 10.6 V so you can ignore that for now… (and most supplies should be fairly stable).
Ok. Seems like I’ll have to do some modifications if I want to get the tracking better which is a bit annoying, so I might just leave it for now and come back to it in future when it starts to struggle under the load of a few more modules. Thanks for the help everyone, time to make a filter!
You can run the CV-12 on 15v DC and then there is plenty of power. I use an external dedicated power supply just for the CV-12 to make sure everything is stable.