If possible (without complicating it too much) I would like the
potentionometer to be one of those ones that can spin a full 360 deg,
so after a full spin, the resistance would go back to 0 (or max
resistance if turend the other way), and the converter would need to be
smart enough to detect this and compensate so it sends the correct
quadrature.
Any ideas?
> Any ideas?
Is there some reason you do not want to just BUY a rotary encoder with
quadrature outputs? I think I have seen some about the size you want.
Jeff.
--
Jeff Shirley
spam-...@mindspring.com
"Bill Gates is filthy rich, but that doesn't mean I want to be married to him."
An easier way to do it would be to just use a rotary encoder and forget
the pot. There are lot's of rotary controls out there that behave
(mechanically) like pots but produce a quadrature (or absolute) output
based on position, rather than a variable resistance. They're plenty
cheap -- I think digikey has zillions of these (search for "rotary
encoder"). There are both quadrature encounders and encoders that output
an absolute position, and there are both optical and mechanical versions.
Is there any reason you have to use a pot?
Otherwise, go to www.digikey.com and search for "rotary encoder".
--
(Replies: cleanse my address of the Mark of the Beast!)
Teleoperate a roving mobile robot from the web:
http://www.swampgas.com/robotics/rover.html
Coauthor with Dennis Clark of "Building Robot Drive Trains".
Buy several copies today!
Pretty easy. You can do that with a PIC or ATtiny with 8 pins. They come
with 8 or 10 bit analog inputs and any kind of output. You will have
some problems at 360 degrees when the pot goes back to 0, because this
is usually not just a tiny jump, but several degrees of dea area without
readings. Also, you have to find someone to write the program for you
and flash the controller. That will run you *far* over $80 for the encoder.
The QEDS-7148 (HP) rotary encoder has 500 counts/rev and is about the
size you want. They are USD $15.00 apiece from C and H sales. Go to:
And check their online catalog. I have a couple of these, and they are
quite nice.
I think you'll find that it's cheaper than a POT and a PIC, assuming your time
isn't worthless. Remember that you need a programmer for the PIC as well, which
isn't free (although it can be done cheaply, or you might already have one).
Also, this unit is optical so it shouldn't wear out any time soon.
One other thing you should understand (and from the tenor and subject of your
messages I thought you might not) is that quadrature output is not absolute --
it merely gives you distance and direction of travel. This unit is quadrature
output. Hit google for more information on quadrature output (if you need more
info).
Hope that helps -- tAfkaks
I checked out the encoder, the only problem is it is 40 mA, and I only
have 5 mA to work with (though it matches on 5V).
The PIC IDE (w/ assembler) from microchip is free. You'll pay for a C or basic
compiler, though. If you go avr, you can use the gcc toolchain. In both cases
you'll need a programmer -- there are a lot of options, so use Google for
details. I'm not aware of a part that runs on .5ma (did you mean 5ma?) -- but
again, go to the Atmel or Microchip sites to check the datasheets. Most parts
run on 5v.
Is there any reason you're so limited on current draw?
Mad Scientist Jr wrote:
> I think I meant to say ADC, not DAC. One more question - how
> difficult/expensive would it be to build a 8-bit analog-to-digital
> converter that you plug a potentionometer into, and it simply
> opens/closes 8 switches to send an absolute value to the KE72 or
> whatever input device you are using?
>
> Mad Scientist Jr wrote:
>
>>I'm limited by the power supply of the keyboard interface that I'm
>>using to read my inputs (Hagstrom KE72 module, see
>>hagstromelectronics.com). Also they have told me that high counts per
>>revolution might be too fast for the keyboard throughput (limited by
>>the PC). If that's a problem, I would lean toward simply using a simple
>>8-bit DAC for each paddle but it kills me to have to use 8 inputs for a
>>stupid paddle.
>
>