Problem with Ray Willson midi to cv

Cfish · Joined: Feb 24, 2016 Posts: 477 Location: Indiana

I used a Ray Willson schematic to build my MIDI to CV converter. Switched the digital side to arduino and a shift register. Built the analog side as described in the schematic.

I am getting substantial drift in the span adjustment for volt per octave with just small changes in temperature.

I used the older of the 2 schematics he had listed. With the tl084 doing most of the analog work

Wondering if this drift in span was inherent to this circuit? or if I am doing something wrong?

As I have had drift issues with the TL07x series of op amps, I almost built this with 4 independent opamps, so I could change out to OP177s. But decided to try it as is.

Any suggestions would be very much appriciated.

Grumble · Posted: Tue Oct 04, 2016 2:13 am Post subject:

Do you have a diagram?

Cfish · Joined: Feb 24, 2016 Posts: 477 Location: Indiana

Sure do. Sorry I forgot too poste it in the first place

IMG_0522.PNG

Description:

Filesize:

551.3 KB

Viewed:

430 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

Grumble · Posted: Tue Oct 04, 2016 11:39 pm Post subject:

I don't think the opamps will be the source of the drift problem, because they hardly amplify in this diagram.
An R2R network is a source of all kind of problems, like switch artifacts, (digital) noise and drift and were used when DAC's were expensive and tedious to use.
I don't know which shift register you use, but I would suggest using the MC74HC595A, this ic accepts SPI serial input and outputs 8 bit parallel.
Or better, use a proper DAC like the MCP4901 from Microchip, this will cost you about 1€50 is single supply powered (5Volt) and lives in a DIP package.

Cfish · Joined: Feb 24, 2016 Posts: 477 Location: Indiana

Thanks for the advice grumble

The shift register we used was SN74HC595N.

This is the only module I have built that required me to pretty much rely on my brouther for half of it.

This was his first atempt at midi receiving.

His specialty has ben midi controllers. ( used a junk Polly six to control a virtual Polly six software) all buttons and knobs work correctly. Same thing with a Hammond B3 and draw bar set.

Has been an education for both of us.

Cfish · Joined: Feb 24, 2016 Posts: 477 Location: Indiana

Question to grumble, or anyone else that may know?

I decided to move the volt per octave trim to the face plate, seeing that way I could at least make things useful.

Odd thing is, it seems to have fixed the problem all together. Switched from a boarns trimmer to a regular potiometer.

What is the chance the trimmer was causing the drift?

I have 2 volt per octave keyboards that have a similar problem and they both use a similar control system. They also both use boarns trimmers from the same batch I ordered.

Nope, I spoke to soon. Must have been hotter than I thought last night when I tuned things. Turned AC on and the drift began once it dropped a couple of degrees

Grumble · Posted: Thu Oct 06, 2016 2:14 pm Post subject:

I'm pretty sure the serial to parallel converter is the problem: It's not designed for analog cirquits.
It is either the temperature dependence of the output levels or the dependency of the output voltages to the power supply voltage (is this stable?) because any change of the supply voltage will immediately give a change of output of your r2r network. (which still is a bad choice imho)
I think you could better try the DAC I suggested, just a minor change in software is necessary

Cfish · Joined: Feb 24, 2016 Posts: 477 Location: Indiana

Thanks grumble.

Have already drawn up my next one with the DAC you suggested, am really just trying to get this one the best possible before I move on. I have learned a lot from it already.

Am currently setting up to test the voltage from the resistor ladder while forcing temperature change. Thanks for the recommendation. Will monitor the 5v supply at the same time.