Help with Water Pump

[alec_t]

D9 goes to +Vt. Would that be 12V?

It needs to be the changing Vt from the tidal timer while testing the effect of the mod(s) on the toggling, but just to see if the FET can switch the pump on connect D9 to 12V for now.

I set the power supply to 20.1V

Suggest you leave it at that permanently, even when we've got these darn pumps toggling merrily .

pump 20.6

pump 21.3

Wow! Both are higher than the 20.1V supply. You've just invented a new power source, Joe! Get your patent application in quickly .

Pump running-fuse 0.1, pump 20.6, Q1 gate 11.4, Q1 drain 0.3, Q1 source 0.3.

That's as expected.

Pump stalled-fuse 0, pump 21.3, Q1 gate 10.8, Q1 drain 0.3, Q1 source 0.1

That's the intriguing set of figures. The FET seems to be not fully on. Could be the FET is oscillating and your meter is doing its best to average weird voltages.
I'll tinker with a simulation to see if I can replicate the effect. Trouble is I don't have a very realistic Spice model of your motor.

 

[salty joe]

I have not done the D9-R13 mod yet because I am not certain what +Vt is. Is +Vt 12V?

I wanted to run everything again but with a known trimmer setting and power supply V.

All FET readings were taken with ground probe of DMM on pin 1 of Q1.

Fuse readings were across fuse.

Thanks.

Pump readings were taken at soldered connections on jack.

Alec, your post did not show up until after I posted. I"ll double chek the pump and power supply voltages.

 

[salty joe]

Wow! Both are higher than the 20.1V supply. You've just invented a new power source, Joe! Get your patent application in quickly .
.

I just double checked the 20.1V at the power supply and the DMM really does read 21.3V at the soldered connections of the jack when the motor is stalled. With the motor stalled, the DMM read 20.1V. Honest. If you guys figure this one out and make a zillion dollars, don't forget your poor old friend back in Ohio. Also double checked the gate, drain and source with DMM ground on pin 1 of Q1 and got the same numbers as before.

With the motor stalled,sometimes the alarm seemed like it was trying to go off, but could not quite get there. It was kind of chirpity chirp chirping like a stutter. If there was any movement of the PDM, it made the stutter worse. The led flickered as well. This is the first I've seen this. I have not changed a thing.

 

[alec_t]

I just double checked the 20.1V at the power supply and the DMM really does read 21.3V at the soldered connections of the jack when the motor is stalled

With the motor stalled,sometimes the alarm seemed like it was trying to go off, but could not quite get there. It was kind of chirpity chirp chirping like a stutter

Both those symptoms suggest to me that the FET is oscillating, perhaps because the 3A current limit is too low. What's your view guys?
Is R2 connected really close to the FET gate pin (which should eliminate high frequency oscillation)? We need to look elsewhere if there's low-frequency oscillation.
If you have a spare 4.7k resistor (like R8 in the Mk12), try connecting it temporarily across Q1 base-emitter. That will increase the current-limiting threshold and perhaps bring the pump out of the stalled state.

 

[KeepItSimpleStupid]

I agree with the oscillation argument. Possibly going in and out of current limit as your suggesting.

Joe also needs to know what Vt is? I tried looking for the schematic reference, but didn't find it.

 

[KeepItSimpleStupid]

Joe:

Looks like vt is the control input or Voltage(timed). See: **broken link removed**

It's the control signal that makes the pumps do it's thing. You have been setting it to +12 for the pumps to run continuously and have it connected to the (?) when it does the pulsing thing.

 

[salty joe]

R2 is in the hole right next to the hole the gate is in.
I do have plenty of4.7K resistors. I'll find which pin is base, hook up the 4.7 & let you guys know. Thanks.
Thanks for finding Vt. In my mind it was control-Vt was there on the schematic all the time...I better tighten up.

 

[salty joe]

I installed a 4.7K between the base and emitter of Q1. The pump still stalls and sometimes runs. Stalled, the pump still shows 21.3V, gate is still 10.8V, source is still 0.1V, but drain is now 0.2V.

No alarm stutter, but the ground between the PDM and control circuit was not as snug as it should have been. Could that have caused the alarm stutter?

 

[alec_t]

but drain is now 0.2V

No alarm stutter

A slight improvement, especially if we've gone from no run to sometimes runs. More thinking/simulation to do.

the ground between the PDM and control circuit was not as snug as it should have been. Could that have caused the alarm stutter?

Possibly. The current would be affected by changing resistance in the ground path.

 

[salty joe]

It was stall sometimes-run sometimes before the 4.7 K was installed. So no change there.

 

[ronv]

With a little stray capacitance from base to emitter of Q1 it will oscillate. I think the phase shift is from the pump inductance. I was able to settle it down with 0.1 ufd in series with 470 ohms across the pump leads. I tried to pick some values that you might have Joe. If you don't have a 470 ohm try something smaller.

 

[salty joe]

Yep, very nice. Here is the table formatted a bit.

Pump	Voltage	Current	Vsense (0.22 resistor)
A	16.95	0.91 A	0.2
B	17.1	0.91 A	0.2
C	17.3	1 A	0.22
D	17	0.95 A	0.21
E	17.1	0.95 A	0.21
F	17.1	0.95 A	0.21

Thanks Joe. I would consider the pumps well-behaved with run current and I doubt "Stiction" would account for anything.

So, I guess, you guys will have to concentrate on circuit design and or tolerances.

For all intents and purposes, the pumps have identical electrical properties.

Thanks KISS for putting the numbers in a clean chart.
Power supply at 20.1V, trimmer at 25K, 4.7K resistor between the base and emmitter of Q1 and toggle set at about 5 sec. on-5 sec. off;
Pumps A, C and E did not run at all, which is strange because under the same conditions yesterday, pump A would run sometimes.
Pumps B, D and F ran reliably.
I ran the test for 7 toggle cycles then plugged each pump in (in no particular order) for one cycle and got the same result. The pumps that did not run made a humm or buzz I could hear through the water. All six pumps are facing up in a 5 gallon bucketof water. To keep things as consistant as possible, I always plugged pump into jack during the off part of the toggle cycle. The alarm did not sound.

With a little stray capacitance from base to emitter of Q1 it will oscillate. I think the phase shift is from the pump inductance. I was able to settle it down with 0.1 ufd in series with 470 ohms across the pump leads. I tried to pick some values that you might have Joe. If you don't have a 470 ohm try something smaller.

So you're saying a smaller resistor and a cap between the emitter and base, is that right?

 

[KeepItSimpleStupid]

No. The stray capacitance was added to the model. Stray, meaning it is caused by the board layout.

A series RC network is also known as a snubber. The RC network (0.1 and 470) goes across the motor leads. Usually lower values of R are employed. If 470 doesn't work, make the 470 smaller. |Another 470 in parallel with the 470 would make it 235 ohms. You can try to go lower too.

We want the load (motor) to look more resistive. The added capacitance kills some of the inductance thus interacting less with the stray capacitance.

 

[alec_t]

Pumps A, C and E did not run at all, which is strange because under the same conditions yesterday, pump A would run sometimes. Pumps B, D and F ran reliably.

That suggests something is right on a threshold of operation. Could be the 3A limit needs raising a bit or the limit needs delaying. Still playing with Spice. Have now made a somewhat more complex model of the motor, including the commutation and back-emf generation. Should work better, but it's still probably some way from reality .

 

[salty joe]

Thanks very much.

 

[ronv]

snubber

If you have the parts, this is where they go.

 

[ronv]

Alec,
Try it with about 500pf from base to emitter of Q1.

 

[ronv]

Alec,
You may be right. Even though it is oscillating the current is still 3 amps. Could have sworn it started on a 6 volt battery way back when, but more might be better.

 

[ronv]

Joe,
You can always try another 0.22 ohm resistor across (in parallel) with the one that is there already. This will give 6 amps before it current limits. If that makes them all start we can search for the minimum.

 

[alec_t]

You can always try another 0.22 ohm resistor across (in parallel) with the one that is there already.

Or if it's more convenient (since the 0.22 is rather bulky) you could connect a spare 1.8k across Q1 base-emitter instead; it would have a similar effect.
But if you do try the 6A limit, make sure you switch the pump off manually after ~ 1 sec if the trip doesn't do so and the pump isn't running properly; we don't want a stalled pump drawing 6A!