Low-power tuned variable resistor

[Steve83]

I don't know if this is a parts ID topic, or a circuit design topic, but I'm guessing parts.

I need a rotary variable resistor that changes from ~16~160Ω in 180° (~75Ω at 90°) with only a short pendulum (<2" radius) on the shaft. It can rotate farther, but it needs to hit those numbers at those angles. At ~5V, it has to pass ~80mA (~1/2W) to operate the 100Ω 12V gauge I want to use. I know how to build the circuit, and how to calculate the specs, but I don't know how to find the actual pot with the characteristics I need; or if it doesn't exist: how to add resistors to give it those characteristics. I only need 2 right now, so cost isn't a real issue, but I don't need a $20 pot since I might build more.

I've looked thru the Mouser catalog, but I'm going cross-eyed because I don't know the terminology, and they don't list the specs I want to see (angle, torque, power...). This is one of the pages I've looked at:
http://www.mouser.com/Passive-Component ... d8Z1yzvm30

One of the data sheets I looked at said the wiper was only rated for 20 cycles - mine needs to stand up to near-constant motion for years, and I'd like to immerse it in silicone oil for mechanical dampening & to protect from dust/humidity.

I'm gonna try to embed a linked thumbnail of my circuit:



The Pre/Parallel/Post resistors may not be necessary, depending on the pot, but I wanted to indicate them just in case.

[pebe]

The idea you have outlined will not work, but there is another way of doing it.
Are you interested?

[Steve83]

Yes, I'm very interested in why a pot's range can't be adjusted.

[pebe]

Your lowest value, 16 ohms, can be added to the 100 ohms of the meter. The next two values of 75 ohms and 160 ohms, are increases of 59 ohms and 85 ohms respectively, so follow a log law – although with less of a curve than an audio taper. But the way you describe can only be used for making a reverse log law pot. See here:

http://www.geofex.com/article_folders/p ... tscret.htm

If you could get a log law pot you may be able to modify it to be more linear using your method, but you would need a wirewound pot because it is doubtful you can get a carbon-based pot with the accuracy you require. I have search the ‘net everywhere and cannot find any wirewound pot that is non-linear – I suppose the practice of winding on a tapered card has long since ended.

The alternative is to voltage-drive the end of the 100 ohm winding, and I can see a way of doing that. I have done some calculations and could carry on with them and come up with a solution, if you are interested.

[Steve83]

Then let me rephrase...

Such a resistor has already been built (in the millions), and similar ones are currently in-production. But they all work through a 90° sweep, and I need a 180° sweep. Here's a pic of one (in an album with several more photos & captions):



It drives its original gauge (the one I plan to use) exactly the way I want, other than the range of motion. As you can see, it's simply an etched & tinned PC board with conductive (probably Carbon) paint applied such that the resistance increases across the contact points (like a wirewound). So it has all the best characteristics of both designs you described. Note that each contact point extends into the resistive paint, and that there are test pads every ~6 points for quality control. So last night while waiting for my post to be approved (what's up with that, anyway?), I bought these:

http://www.ebay.com/itm/230757137716
http://www.ebay.com/itm/370754688466

If the wire's resistance isn't high enough to build a wirewound strip that I can bend into a 180 arc and keep it small, then I'll buy a cheapo etching kit (like RadioShack sells) and use the paint to mimic the original with a longer sweep. If neither works, I'll solder wires from my etched 180 board back to that original 90 board and just use the original resistive paint through 66 tiny jumper wires.

BUT...
I'd much rather just buy them, if they exist or can be created by trimming existing rheostats/pots with external resistors.

[pebe]

Such a resistor has already been built (in the millions), and similar ones are currently in-production

But they are not made the way you suggested in your original post! I was not aware at the time that you were prepared to make a rheostat from scratch – that’s quite a different ball game. I saw a production sample for one that was made for Ford cars to accommodate an unusual shape of fuel tank. The track was made of conductive ceramic and was a weird shape. Each was individually laser trimmed.

If you use resistance wire you will need to work out the taper for the former you intend to wind it on. But if you use that wire from ebay it has a resistance of 33.4ohms/ft so you will need to wind about 5ft of it.

Good luck.

[Steve83]

They're made exactly the way I suggested, other than the sweep angle. I took all my measurements from one like that last pic, and I'm prepared to take measurements from all 66 wiper contacts if necessary. Would that help ID a mass-produced long-sweep rheostat that can be trimmed to the output I need?

[pebe]

Maybe the item you need is specific to those sorts of gauges and would probably only stocked by vehicle spare-part stores, and doing individual measurements may not help. I think it very unlikely you would get them from an electronic component supplier.

But are you quite sure that the item is made in the way you described? A linear law pot would give a resistance increase proportional to the increase of rotation. The figures you gave of 75ohms at 90% rotation and 160ohms at 180% rotation show that an increase of angle by a factor of 2.0 gives an increase in resistance by a factor by 2.133. That is a log or semi log law

But if you look at the thumbnail you showed of a rheostat with a fixed resistor in parallel the result is different. Assume a rheostat resistance of 2kohms overall with a parallel resistor of 1kohms, then at 50% rotation you have 1K and 1K in parallel (500ohms), and at 100% rotation you have 1K in parallel with 2K (666ohms). That gives a ratio factor of only 1.33 and so it has an inverse log law.

Whatever values you choose for the resistors you cannot change its overall law from inverse log to semi log. But if you were to use a transistor as the variable resistor element then it might be possible to drive it with an inverse log pot. Would that sort of arrangement be acceptable?

[sachinelectrostar]

I am also interested.....

[pebe]

In that specific gauge?