Help with Water Pump

[salty joe]

I just double checked the flick pot and it is 500K. I took the 529 reading with my DMM on the 2000K setting.

 

[salty joe]

The Q4 (pin 7) on 4020 does toggle every 42 seconds or so, but the 500K flick pot does not seem to be working. With the pot turned almost all the way to the left, I counted 20 seconds of a 12V toggle. With the pot turned almost all the way to the right, I counted 19 seconds of a 12V toggle. Probably the exact same toggle time.

I started to think maybe I was operating the wrong pot last night so I ran the flick test again. Adjusting the 500K flick pot did not affect the 12V toggle from pin 7 on the 4020.

But this is wierd. This morning when I checked the 500K flick pot, I removed the dummy loads. I forgot to replace the dummy loads when I re-ran the flick test and now the LED that was not turning all the way off is turning all the way off. I'll re-run the flick test with the dummy loads and if it has a different result, I'll post in a few minutes.

 

[alec_t]

Adjusting the 500K flick pot did not affect the 12V toggle from pin 7 on the 4020.

It's not supposed to. The flick pot merely adjusts the width of a brief negative-going pulse (at the right side of C10) from zero to ~ 0.25 sec. As the DMM probably can't detect that short a pulse, so will read a steady 12V, your only indication (if you haven't got the pumps or dummy load connected) will be the LED of the idle pair of pumps. Looking for a 0.25 sec flash every 42 sec is not easy. Blink and you've missed it. Speeding up the clock as I suggested earlier would probably help.

now the LED that was not turning all the way off is turning all the way off.

Strange indeed. Perhaps driving the dummy load passed enough current to burn away a microscopic metallic whisker in or around the FET and cured the current leakage.

 

[salty joe]

I soldered a jack on the idle side pump leads and plugged a pump into the jack. I watched the pump propeller for a couple minutes with the flick pot adjusted almost all the way to the right and anotheer couple minutes with the pot adjusted almost all the way to the left. The propeller did not move. Any ideas?

 

[salty joe]

double post sorry

 

[Boncuk]

Really amazing ...

how long it can take to get a simple water pump running.

Hi salty Joe,

from the first post until today more than four months have passed and the pump(s) still seem not to run as desired.

Good you don't have make your living off making pump control circuits.

Here is my pump control circuit (original Thai patent):

Wife climbs the ladder to have unlimited view into the water tank. She yells out to the daughter to switch on the (piston type) pump being fed by a water well. She checks the water level from time to time. She yells again if the tank is full or water spills over the tank to switch off the pump.
Daughter turns off the pump.

House distribution is done my way (European): Open any faucet: water is supplied and if the pressure in the reservoir the pump gets low the pump is cycled on. The pump is cycled off after the faucet(s) is/are closed and reservoir pressure has risen beyond the set point.

This "system" has worked well and tank refilling will be done automatically if I can convince my wife that technical solutions are reliable.

She's already one step closer to get convinced. Her motorbike has made 55,000km without any failure because she has it maintained the way I prescribed.

One package of cookies less for the daughter buys an oil change.

Regards

Boncuk

 

[alec_t]

@Boncuk
Any chance you could get your family to give Joe a hand?

Any ideas?

Maybe the flick pulse is too brief (the ~0.1 sec value was only a guestimate). Try temporarily connecting a higher value cap, say up to 10μF if you have it, in parallel with C10. That should extend the pulse to several seconds and make its presence more obvious.

 

[salty joe]

@Boncuk
Any chance you could get your family to give Joe a hand?

LOL I could use the help!

[/QUOTE]
Maybe the flick pulse is too brief (the ~0.1 sec value was only a guestimate). Try temporarily connecting a higher value cap, say up to 10μF if you have it, in parallel with C10. That should extend the pulse to several seconds and make its presence more obvious.[/QUOTE]

I do have some 10 uf caps. I'll throw one in tomorrow & let you know. Thanks for that plan.

 

[salty joe]

You are a genius Alec! Installing the 10uf cap made the flick function work. The entire schematic is amazing and brilliant. Thank you so much for guiding me along. I struggled on this one for sure. Your advice to diagram the components turned the corner for me. At that point I could see how to break this job down to several smaller pieces. This controller will be a workhorse for the tank. I can't thank you enough.

 

[KeepItSimpleStupid]

Nice job!

I do see one "potential problem". If there is a need for a heatsink, there isn't enough room on the board to install one.

For reference: These pins http://search.digikey.com/us/en/products/T42-1/M/V1070-ND/38995 are generally used for 0.042 holes for mounting of individual leaded components. There is an insertion tool for them as well. Yep, they are expensive.

I threw in a picture of a breadboard I did some 40 years ago in high school for an Analog Computer that I built for a Science Fair project. It displayed a "bouncing ball" on an oscilloscope in X-Y mode. You can vary parameters like initial height, the medium, the horizontal velocity and the springyness of the ball. You can actually see the ball deform. I have no idea if it still works. It's quite dusty. It's mostly made from 741 OP amps in the now obsolete package. Those terminals have been around for a long time.

 

[KeepItSimpleStupid]

Sorry, uploaded an unsupported pic format. Fixed.

 

[alec_t]

Glad we got there in the end, Joe. Better start a new thread for your next project!

@KISS

I do see one "potential problem". If there is a need for a heatsink, there isn't enough room on the board to install one.

Joe has confirmed that the FETs don't get hot. They only switch once in a blue moon after all.

 

[salty joe]

Nice job!
.

Thanks, but all I did was follow directions. Well, try to anyway.

Glad we got there in the end, Joe. Better start a new thread for your next project!

.

Alec, I really do appreciate you doing these killer schematics for me and I very much appreciate you hanging in there through all the mistakes and delays. You are extremely talented and extremely generous. What a guy!

 

[alec_t]

Thanks for the thanks. One thing I should have said is that the wires going to the source and drain pins of the FETs (in both your projects) need to be rated to carry the heavy current that the pumps demand, and they should be connected close to those pins (so that voltage drops due to wire resistance don't affect other parts of the circuit). Maybe you've already seen to that.

 

[4pyros]

I didn't know I needed a dummy load. Sorry about that. Especially sorry if a dummy load was mentioned in this thread. Anyway, I have 25W and 40W incandescent light bulbs, would one of these be a good dummy? Would just one dummy bulb per side get hooked up?

I mentioned that a few days ago.

Post #322

You should have a dummy load that uses about as much power as the motor, or use the motor.

Andy

 

[salty joe]

Thanks for the thanks. One thing I should have said is that the wires going to the source and drain pins of the FETs (in both your projects) need to be rated to carry the heavy current that the pumps demand, and they should be connected close to those pins (so that voltage drops due to wire resistance don't affect other parts of the circuit). Maybe you've already seen to that.

Would 22 gauge be heavy enough? If not switching to heavier wire should be no biggie. Thanks for the heads up.

I mentioned that a few days ago.

Post #322

You should have a dummy load that uses about as much power as the motor, or use the motor.

Andy

Sorry I missed that Andy. Thanks for your input on this project, I certainly appreciate it.

 

[4pyros]

Would 22 gauge be heavy enough? If not switching to heavier wire should be no biggie. Thanks for the heads up.



Sorry I missed that Andy. Thanks for your input on this project, I certainly appreciate it.

You should look up a wire current chart on the web. The right size wire is a function of the current, wire lenght and temp. There are lots out there.
I dont remember what your design current is but the the wire needs to be big enough to handle the max current that can be supplyed through the current protection device.

IE;
If the power is fused at 10 amps the wire must be rated for that or more so it will not burn. If the wire run is 100 feet it must be bigger. If the wire is bundeled with others or in a hot enviornment it must be bigger. It all starts at the fuse or braker. Andy

 

[alec_t]

22AWG is rated 8A by Farnell. 4pyros makes a good point about fusing. I think I'd fuse the whole project at 10A and run individual 22AWG wires to each pump pair, i.e. one +24V wire to one pump pair, another +24V wire to the other pump pair, one ground (-) wire to one FET source, another ground wire to the other FET source.

 

[salty joe]

Thanks guys. For the drain on the FETs, I bent the drain leg and soldered it directly to the schottky diode. The source leg is soldered directly to the 22AWG ground wire. Pump leads and all the wire is 22AWG.

It's good to know about a current chart, thanks.

 

[4pyros]

22AWG is rated 8A by Farnell.

Then you should fuse it at 8 amps.