Path: utzoo!attcan!uunet!kitty!larry
From: larry@kitty.UUCP (Larry Lippman)
Newsgroups: sci.electronics
Subject: Re: electrically operated valves
Summary: Building proportional control valves...
Message-ID: <3165@kitty.UUCP>
Date: 16 May 89 03:50:34 GMT
References: <23526@shemp.CS.UCLA.EDU> <11170007@hpfcdj.HP.COM> <631@corpane.UUCP>
Organization: Recognition Research Corp., Clarence, NY
Lines: 81

In article <631@corpane.UUCP>, sparks@corpane.UUCP (John Sparks) writes:
> >I'm not sure if washing machine, sprinkler, or other binary
> >solinoid valves would suffice the original poster, but
> >what I would like and can't find are PROPORTIONAL valves
> >of some sort.  Something like a gate value with a stepper
> >motor attached.  Something that can take water at 80 PSI.
> >(Which leaves out $100 each heating system zone valves.)
> 
> I don't know, but we've (Corpane) have been looking for such an animal for a
> while now. We have water cooled machines and would like our on-board
> microprocessor to control how much water is let into the system. So a stepper
> motor controlled proportional water valve is exactly what we need.
> 
> We looked everywhere, I think we found one place that sold it but it was an
> outrageous price (somewhere around $1500). If anyone knows where we could
> find a cheaper version than that, Please help us out and tell us where.
> 
> I guess our next step will be to build it ourselves. But we really aren't set
> up for such an enterprise. 

	My organization frequently uses proportional control valves for
various chemical instrumentation projects.  Most of the valves we use are
off-the-shelf chemical industry control valves which require a 4-to-20 mA
actuation signal (i.e., 4 mA = closed, 20 mA = fully open).

	However, there have been a few projects where we did not wish to
use an such an off-the-shelf valve because of size considerations, and
because we wished to use a stepper motor and remain all digital for more
precise control.  Here is how we did it:

1.	First, begin with a bellows valve intended for metering service
	(actually a globe or needle valve will also work).  In one
	particular application we used a Nupro (a Swagelok company)
	bellows valve.  Nupro is the cadillac of valves; lesser quality
	valves should suffice for water control applications.

2.	The above valve required about 2-1/2 turns from fully closed to
	fully open.  To limit travel of the valve, use a mechanical
	limit stop assembly.  We used a device made by Winfred Berg, Inc.
	which cost about $ 100.00, and which has an adjustable stop range
	from about 30 to 900 degrees.  Couple the limit stop assembly
	to the valve.

3.	Provide a friction clutch between the limit stop assembly and
	the stepping motor.  Again, we used Berg for this item.

4.	Connect the stepping motor to the friction clutch.  We used a
	Superior Electric Slo-Syn (tm) series stepping motor.

	While the above items cost us several hundred dollars, with
some luck you can probably find them surplus for pennies on the dollar -
except for the limit stop assembly, which is not too common.  With
some ingenuity one can probably design some type of multi-turn limit
stop assembly for a few dollars in parts.  The clutch is optional;
you can use the inherent torque limiting of the stepper motor to
limit rotation.

	The nice part of the above design is that no feedback pots or
limit switches are necessary.  The limit stop assembly protects the
valve from over-travel, and the clutch protects the stepper motor
against overheating.

	In operation, the microprocessor first "initializes" the valve
on powerup by pulsing enough steps to close the valve plus some extra.
Since the microprocessor had no idea where the valve position was
upon powerup, we gave it something like the full open-to-close steps
(around 500) plus an additional 50 steps.  This _assured_ the system
that the valve was closed AND against the closed rotational limit
stop.  Since we were now initialized, the microprocessor merely kept
track of open and close steps (actually 1.8 deg increments) to see
where the valve was.  As long as the algebraic sum of valve increments
is maintained, there is no need for any feedback devices.

	I'm certain that many Net readers can construct the above scheme
on a low-budget basis.  The least expensive item should be the stepper
motor; just look for a junk LA-36 Decwriter or equivalent. :-)

<>  Larry Lippman @ Recognition Research Corp. - Uniquex Corp. - Viatran Corp.
<>  UUCP {allegra|boulder|decvax|rutgers|watmath}!sunybcs!kitty!larry
<>  TEL  716/688-1231 | 716/773-1700  {att|hplabs|utzoo}!/      \uniquex!larry
<>  FAX  716/741-9635 | 716/773-2488     "Have you hugged your cat today?"