# How to make a non-linear potmeter more linear

I have an "Expression pedal" that worked fine on my old keyboard (Gem WX2+) but works almost like a digital switch on my new keyboard (Roland VR-09).

The recommended expression pedal is a Roland EV-5 which has the schema at the bottom right of this page:

I measured the pins of my old pedal and the schema is probably like this (have not opened it yet):

simulate this circuit – Schematic created using CircuitLab

An important difference is that the Roland pot is Lineair, while mine seems to be logarithmic. When close to the heel position I can easily navigate close to a wanted resistance. That becomes increasingly difficult moving closer to the toe (fully pressed) position.

I think it should be possible to tweak it to a more linear like curve by making an adapter, adding a resistor between the tip and the ring to pull the curve up somewhat. And while doing so I can just as well add the minimum volume feature as well. This is what I have in mind so far:

simulate this circuit

I'm not an expert at electronics, so I would really appreciate a peer review of this idea before I order the parts. I'm not sure if the value of the curve correction pot is ideal? At the maximum setting, the total resistance would be 9.6K which is a little lower than the 10K of the Roland pedal...

In fact I'm not sure if this idea would work at all, but since a new pedal is quite expensive and has a smaller "throw/range/angle" than my old one, I'd really like to keep my trusty old pedal with a hack if possible. Any other suggestions are welcome off course!

• The circuit diagram says 'special taper', not linear. Can you measure it to find out the actual taper? Commented Aug 16, 2019 at 1:11
• easiest way is to remove the log taper pot and install a linear taper pot Commented Aug 16, 2019 at 1:54
• @BruceAbbott Thanks, I didn't notice that. I can't measure the taper of the EV-5 since I do not own one. But you gave me an idea, I can log the midi response curve of my current pedal by doing a few sweeps with a constant speed. I guess the inverse of that curve is what I should try to go for in order to get it more lineair... Commented Aug 16, 2019 at 12:07