Re: 3vdc motor PWM problem

On 3 Oct 2006 00:40:13 -0700, "randomname"
<randomname12345@xxxxxxxxx> wrote:

This is interesting to me, so I'm going to build one. :-)

Welcome aboard :)

I'll use a drop detector (either reflective or transmissive IR, I'm
not sure which, yet) and white LEDs to strobe the drops, which I'll
probably leave clear right at first.

If you get this part working, could you let me know how it's done? I'd
love to try drop detection in the future... but with it you really need
a PIC to get the slow motions forwards/backwards. That's why I'm
gunning for constant drip :)

---
I'm planning on using an IR opto switch, with the drop breaking the
beam as it falls. Basically an IRLED as the emitter and a
phototransistor as the detector. The IRLED will keep the
phototransistor in saturation until the drop interrupts the beam, at
which time the detector will come out of saturation, its output will
go high, and that high-going edge will be used to start the strobe
delay.

If you substitute an LDR for the delay timing resitor then you can
change its resistance by shining an LED (or an incandescent lamp) on
it. That way, you can use, say, a slow ramp generator to illuminate
the lamp and change the delay, so the column of drops will rise and
fall slowly, automatically.
---

My goal will be to generate 1/4" diameter drops that fall through
12" with a new drop being generated every 16ms or so.

1/4" drops sound a little big... you should probably use a McDonalds
straw for that (which I'm going to try tomorrow). I'm using 3/16" OD
aquarium tube, and my drops are 3/8", which are a very nice size.

---
3/8" drops?
---

As per the 16ms... that will prove to be very difficult. The fastest
drip I can get before it turns into a stream (in which case the drops
are extremely erratic and tiny, but much more frequent) is such that
there are 3 drops in the air at any given moment. Considering the fall
is 3.5 inches.. that means there's a drop roughly every 45ms. It
should be tough to cut that in third.. and if you manage somehow do let
me know!

---
Change of plans... I want to have the 12" fall, and I want only one
drop to be visible when it's halfway down, so I'll need to generate
drops every 6", which is 176ms apart. That's about 5.7Hz, so
there'll be a lot of flicker, but that's where I want to start. :-)
---

Another issue you'll run into is that 12" is pretty high. Mine was
originally this height. I figured: hey, I want to watch the drop for a
long time and see a nice splash at the end, too.

The problem here is that even at the 45ms rate, where the drops hit is
a very chaotic area. You'll get bubbles crowding up, and splash
everywhere, and since these arent periodic it will ruin the effect.
I've found the best height to be a little under 4". The impact region
really is the most impressive part of this piece, IMO. It gets ruined
when the drops come from too high, too fast. (Since the drop rebound
hits the next drop falling)

---
I'm more interested in manipulating the drop(s) in the column than I
am in the impact, right now, so I plan to diffuse the drops when
they return to the tank so they won't create pressure pulses which
could affect the pump. Maybe that's part of the problen you were
having originally with your pump's stability?
---

the appearance of a single drop suspended in the center of the "drop
zone" with a drop rising out of the catch basin, following the
center drop when the center drop appears to rise, or a drop
appearing at the top of the drop zone, following the center drop
when it appears to fall, all depending on the detect-to-strobe
delay.

If you're rate of dripping is 16ms, and the height is 12", then you
will have 16 drops in the air at any given time.

---
Right. My error, thanks.
---


I have a submersible fountain pump which can deliver about 1.5
liters per minute with a 12" head, so that's enough volume VS time
to generate the quantity of drops needed. It runs off the 120V
mains so I'm not going to mess around with PWM'ing it, I'll just
control the flow with a needle valve I had laying around waiting to
be put to use.

I wish I had a valve. My valve is a c-clamp tightening the rubber
tube.

---
Nothing wrong with that. Clever, as a matter of fact!
---

Anyway, best of luck. Can't wait to see how you do!

---
Thanks!-)


--
John Fields
Professional Circuit Designer
.

Relevant Pages

  • Re: LVDT Transfer Function
    ... What matters is the output of the LVDT electronics as a ... From delay and period, calculate phase. ... be recoded so the transfer function can be determined. ... filter at the detector output. ...
    (sci.engr.control)
  • Re: Variable frequency delay
    ... uses a magnetic tach sensor which has a 0 - 5V output. ... goes directly into the strobe. ... ms delay would cause a different angular shift in the image I see. ... slow and high speeds. ...
    (sci.electronics.basics)
  • [patch] timer-irq-driven soft-watchdog, cleanups
    ... 11-seconds delay and stayed silent during a 9-seconds delay. ... If you still get warnings even with this patch applied, ... this patch makes the softlockup detector purely timer-interrupt driven, ... this code detects soft lockups: incidents in where on a CPU ...
    (Linux-Kernel)
  • Re: OT: Steve Wozniak
    ... Transitions in the "strobe" signal should take care of that. ... a fairly long delay for all the bits to settle and the delay can be a ... Remember when serial interfaces used a separate clock signals for ...
    (comp.os.vms)
  • Re: PLL convergence problem
    ... gardner's "phaselock techniques", meyrs "Digital communication ... the most important part in the closed loop ... is the "phase detector" or "error metric". ... which have the same delay. ...
    (comp.dsp)
Loading