Help with Water Pump

[salty joe]

The diode test on your meter would be useful for solder bridges here too.

On the diode setting, how would a short be indicated on my DMM? In this case polarity would not matter (no difference where each probe goes), is that right?

Voltage between which two points?
.

That was taken at the LED for the pumps. The LED is hooked up like the pumps except the LED has a 2.2K resistor on the positive lead.

Later today I'm going over this circuit with a fine toothed comb. Alec, awhile back you mentioned the possible cause for the toggling LED not completely turning off, in this case toggling between 24.4V and 15.8V, could be diode trouble. Can the diodes be checked while they are hooked up? If so, I can test an unused Shotkie 1N5822 (installed by pumps) and an unused 1N4148 (installed by the 500K clock pot) and compare readings with the installed diodes.

Thanks guys!

 

[4pyros]

That was taken at the LED for the pumps. The LED is hooked up like the pumps except the LED has a 2.2K resistor on the positive lead.

You may need more of a load like the pumps them selfs to get the voltage to drop all the way. The LED may not pass enough current.

 

[salty joe]

Check also that you're getting a clock signal (0/12V toggle every 2.6 secs) at the top end of R2 (= output of gate U2a). Then if that doesn't identify the problem it's a matter of double-checking all the pinouts and soldered connections of the 4020.

Yes, the output gate of U2a toggles between 0/12V every couple seconds. I used pin 8 for that.

If the blink is too brief it may not be visible. I've just knocked up a breadboard version of this part of the circuit. With my 'C10' = 0.47u and my 'trimpot' = 330k my LED blinks on for ~ 0.2 sec.
Could be due to FET or Schottky diode leakage current, or spurious FET turn-ons due to noise (circuit presently unscreened).

I tested the four installed Schotky diodes and they registered 167-169. New Schotky diodes registered 154. The diodes for the soft start measured 683 and the same type diode for the on and off pot were 688/699. New diodes measured 668.

Measure the voltage (20VDC range) at the left side of C10 (= Q4 out from 4020). It should toggle between 0 and 12V every 21 sec or so.

Yes, Q4 or pin 7 on the 4020 toggled between 0/12V every 21 seconds or so.

After taking anti-static precautions, temporarily short the bottom end of R6 to ground. Both LEDs should light (all 4 pumps on).

Yes, both LEDS lit up when I took the bottom of R6 to ground.

I checked all IC connections for shorts with my DMM. At 2000K, conductivity was indicated between any pin on an IC and ground, so I tried 200K and it showed no conductivity between any pin on an IC and ground. At 200K my DMM told me that there are no shorts between any IC pins. I also traced every IC connection and everything seems to be connected as it should be.

I am letting the circuit run to see if the other LED will toggle. Should have that answer this evening.

 

[alec_t]

Good. We're getting there. Those checks confirm the oscillator (U2a) and the 4020 are working.

I am letting the circuit run to see if the other LED will toggle.

If you want to speed things up 10x or so while testing you can connect a ~ 47k resistor temporarily between the input and output pins of U2a.

 

[salty joe]

Just for the heck of it I measured the V coming out of the pump leads from the side that's toggling and it is 24.4 and 3.0V. That makes no sense to me since the LED on that side shows 24.4 and 15V. The leads for the pump and LED are hooked up exactly the same except the LED has a 2.2K resistor on the + lead.

Neat idea to speed thing up, but I have to run. I would have thought a resistor would slow things down if anything. Counter intuitive for sure.

 

[alec_t]

I'd like to clarify where you're measuring the voltages.
In the attached, what voltages do you get with
(1) the + probe at point '1' and the - probe at '3' (gnd) ?
(2) the + probe at point '2' and the - probe at '3' ?

Can you confirm the LED/resistor combo is connected between the +24 rail and point 1 (or 2) ?

I would have thought a resistor would slow things down

The added resistor would be in parallel with the existing R2 and pot, so overall resistance is reduced.

 

[salty joe]

The other LED is toggling! The magic you wrote in there is starting to show itself.

The voltage where the drain connects to the Schotky on the toggling side is 0V when the LED is lit and 22V when the LED is off. The LED on this side turns all the way off.

On the idle side I get a steady 21.3V, and the LED is dimly lit.

I unplugged the circuit and plugged it back in so the side with the dimly lit LED is toggling. I get 0V when that LED is on and 21.3V when the LED is dimly lit. The idle side shows a steady 22V.

I double checked the LED resistor combos and they are hooked up the same. I connected two 2.2K resistors to the 24V rail and the positive LED leads to the other side of the resistors.For the negative LED leads, I connected leads to where the drains connect to the Shotkys. (points one and two in your attachment)

On the side where the LED turns all the way off, the voltage from the pump leads and the LED leads is the same-24.2/2V.

Thank you.

 

[KeepItSimpleStupid]

On the diode setting, how would a short be indicated on my DMM? In this case polarity would not matter (no difference where each probe goes), is that right?

A short would indicate 0.00 and the meter usually beeps if under a certain voltage.

The diode mode of the meter passes a fixed current through the probes with a "compliance voltage" and measures the resulting voltage drop. A typical Silicon semiconductor junction drops around 0.6 V or so. Schotkey diodes will drop less.

You will get what might be confusing readings when using an ohmmeter to test a diode.

The diode mode typically gives you a "beep" for continuity or (Under some value of dropped voltage). "OL" is typically indicated if the leads are open.

 

[alec_t]

O.k. The LED resistor combo connections seem fine. I suspect one of the FETs isn't fully turning off.
Now (taking anti-static precautions) measure the voltages at the left side of R4 and R5 with/without toggling. What do you get?
BTW, do you have the dummy load (mains filament lamp) connected when doing these checks?

 

[salty joe]

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?

 

[alec_t]

IIRC I suggested the dummy load for your previous project. Sorry, I should have reminded you to use one in this project. Saves running the actual pump on the test bench. It's not essential, but is a more realistic test for the FET driver than just having a LED as the load!
Yes, a 40W bulb (crudely representing a pump pair) would be good. I think with that (or the pumps) in place you won't find the LED lighting faintly. One dummy load per side (but you could just swap one from side to side).

 

[salty joe]

IIRC I suggested the dummy load for your previous project. Sorry, I should have reminded you to use one in this project. Saves running the actual pump on the test bench. It's not essential, but is a more realistic test for the FET driver than just having a LED as the load!
Yes, a 40W bulb (crudely representing a pump pair) would be good. I think with that (or the pumps) in place you won't find the LED lighting faintly. One dummy load per side (but you could just swap one from side to side).

No,no... I should have remembered. My bad about the dummy load. I'm such a dummy sometimes.

So I put a 40W bulb on each side and the LED that was dimly lit now turns all the way off. Seems odd that only one LED did that.

The voltage on the 4093 side of R4 and R5 is the same. 12/0 when toggling and 0 when idle.

 

[alec_t]

So I put a 40W bulb on each side and the LED that was dimly lit now turns all the way off.

Excellent.

Seems odd that only one LED did that.

Your test with the dummy load in place confirms there's a slight leakage current via one FET. It's not turned fully off when there's 0V on its gate. That could be due to manufacturing tolerances, or a slight fault with the FET. Doesn't seem significant, 'cos the pumps won't operate on only a few microamps, and we know the LED will now go fully off with the pumps connected .

The voltage on the 4093 side of R4 and R5 is the same. 12/0 when toggling and 0 when idle.

That's as it should be. So, all systems go?

 

[salty joe]

Yes, it's fabulous except there does not seem to be any flick. At least my DMM does not show anything. I set the flick pot for a quick interval and held the DMM probes to the second set of leads on the idle set of pumps.

 

[alec_t]

there does not seem to be any flick

Is the Q4 output of the 4020 (left side of C10) toggling at ~ 42 sec?

I set the flick pot for a quick interval

The 4020 sets the interval. The pot sets the duration of the flick (e.g. 0.1 sec, which probably won't register on a DMM).

 

[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.

 

[KeepItSimpleStupid]

I'm not sure what kind of potentiometer you purchased, but make sure it either decreases or increases in resistance when you turn it.

Some pots don't have markings. When facing the shaft, there is a left lug (CCW) a right lug (CW) and the middle is the wiper.

If it's a trimmer, you may find nothing and it's marked like above. You also migh find CW, CCW and an arrow.

Make sure you use the center lug and/or arrow and one side.

Usually the wiper is tied to the CCW or CW position in practice.

 

[alec_t]

With no power on the circuit, measure the resistance between the right end of C10 and the 12V rail. Does it vary from 0Ω to ~500kΩ as you rotate the pot wiper? If not, you may have a wrong pot connection or a duff pot.
Check that one end tag of the pot goes to the 12V rail, the other end tag goes to C10 (+ = right side ), the middle tag (wiper) goes to one of the end tags.

 

[salty joe]

It was hooked up correctly and I could adjust the resistance from 0 to 529.

Thanks.

 

[alec_t]

529.

Do you mean 529k? If your pot is only 529 Ohms you won't get any noticeable flick