On Mon, 14 Jul 2008, Jon Elson wrote:
> Date: Mon, 14 Jul 2008 11:33:37 -0500
> From: Jon Elson <[EMAIL PROTECTED]>
> Reply-To: "Enhanced Machine Controller (EMC)"
> <emc-users@lists.sourceforge.net>
> To: "Enhanced Machine Controller (EMC)" <emc-users@lists.sourceforge.net>
> Subject: Re: [Emc-users] resolver to quadrature converter
>
> Dave Engvall wrote:
>> Hi Steve, Jon:
>>
>> The Analog Devices chips have really good specs especially at low freq.
>> How they are rather pricey in todays world of inexpensive chips with
>> excellent performance. A couple of $14 to $31 chips drives the price of
>> the finished board up pretty fast.
>> There is a National tach chip ( LM2917) at much lower prices but also
>> lower specs.
>>
> This is still a unipolar (= unidirectional) F-V converter.
>> I wonder how well a digital approach with a simple moving average would
>> work. I'm assuming that a good digital filter would take too much time
>> to compute but that is someone else's problem. ;-)
> FPGAs can do digital filters in the MHz range, if you really
> need it. But, the problem is to filter a discontinuous-time
> signal (encoder counts) into an approximation of a
> continuous-time signal (velocity) with minimal delay AND minimal
> ripple at low speed. NOT trivial.
>
> Jon
A digital system that either did Count/TQuad velocity estimation, or 1/T
velocity estimation at low speeds and switched to M/TSample at higher speeds
is possibly a good way to go. Some fancier velocity estimators take higher
order differences (and maybe motor applied torque) into account to get a
better velocity estimate. If you go digital for processing, you still need to
get a wide dynamic range, low delay signal out, which probably means a DAC, so
its a multi chip solution.
Our HostMot2 firmware supports Count/TQuad velocity estimation, basically just
timestamping the quadrature edges, so that when you read the encoder counter
you can read the number of counts that have occured and are able to
accurately measure (well limited by quadrature phase distortion) the total
time elapsed time for the counts.
Some fancier stuff we've done for customers saved every timestamp in a FIFO so
the differences could be extracted, and I've seen some other commercial
hardware that saves the last 4 or 5 timestamps.
A $2.50 DSPIC could probably do all this up to a few hundred KHz (I'd probably
do the divide with a table) but you would need a SPI DAC to get the analog
output...
Peter Wallace
Mesa Electronics
(\__/)
(='.'=) This is Bunny. Copy and paste bunny into your
(")_(") signature to help him gain world domination.
-------------------------------------------------------------------------
Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW!
Studies have shown that voting for your favorite open source project,
along with a healthy diet, reduces your potential for chronic lameness
and boredom. Vote Now at http://www.sourceforge.net/community/cca08
_______________________________________________
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users
