Help with Water Pump

[salty joe]

Wow. If the pump really has 18.68 volts across it it should be running and drawing more current (higher voltage at the source) What this sounds like is that the leads to the pump are open. Maybe a bad connection around the plug?

If we can understand this one I think it will be time for the mods, but it is actually easier to look at some of these without the current limit and alarm working.

On the other measurement it would be best to just measure the voltage across the resistor. It will be pretty small -- like 0.25 volts.

I have a jack with wires soldered to it. I used wire nuts to connect the jack to the PDM and plugged the pump into the jack. I’m pretty sure the connections are good. I got that 18.68V at the jack.

KISS, so I need to measure the V at the pump, OK But what about the variable reading? Do I take an average which would really be a good guess? Worse, what about the inconsistent readings? For pump A, I got 16.14-16.37 and 5-10 minutes later got 16.78-17.01.
The reading across Rsense will be in amps so I will need to change the setting on my DMM, is that right?

 

[alec_t]

what about the inconsistent readings?

Do you have a fresh reputable-brand battery inside your DMM? If it's the one that came with the meter it could be dying.

The reading across Rsense will be in amps

No. Volts. You calculate the Amps (if you want) by dividing the Volts reading by 0.22.

 

[ronv]

I have a jack with wires soldered to it. I used wire nuts to connect the jack to the PDM and plugged the pump into the jack. I’m pretty sure the connections are good. I got that 18.68V at the jack.

Hmm. maybe a bad pump. You can try moving the wires around and see if it starts up.

 

[KeepItSimpleStupid]

Joe:

100 mV is only 0.5% of 20 V, so even a few hundred mV isn't really a significant number. For motor voltage read to the nearest 0.1 V and take the (Max+min)/2.

For 3 Amps, we would expect voltages near 0.6 V. Again, voltages probably within the nearest 0.010 V will be fine (Max-Min)/2.

As for why things are not steady. A two phase motor. Right there there is an opportunity for jumpyness. We do care about both average currents an peak currents, but it's not something YOU can do. It either requires a scope (Preferably storage) or a peak-hold meter. Don;t worry too much about the actual numbers.

I did measurements once where 1e15 ohm-cm was basically the same number as 3 e15 ohm-cm, yet the second number is 3x larger than the first.

Being, Two phase, temperature, turbulance and Additude (A fancy word for orientation of the motor) , Stiction (Start up friction) will all play a part.

Now, we are interested in starting info. Your numbers "could" be better, say if you left all of the motors in the bucket for 24hrs or even an hour before starting the test. I can't be 100% certain though.

Don;t worry about a 5% do the best you can. 10% of 20 V is 2 V, 5% is 1 V, 2.5% is 0.5V. Now the readings don't look so bad, do they?

 

[salty joe]

V pump VRsense VRsense/0.22
A 16.95 0.2 0.91A
B 17.1 0.2 0.91A
C 17.3 0.22 1A
D 17 0.21 0.95A
E 17 0.21 0.95A
F 17.1 0.21 0.95A

The Vpump numbers went up a little as time went on. I ran all pumps through the test three times until the values seemed to stabilize. Maybe the power supply needed to warm up? I used the final “stabilized” numbers.
The VRsense numbers were consistent.

edit: I am sorry for the way these numbers got jammed together. I had them nice and neatly seperated in the reply box. I hope you can make sense of it.

 

[ronv]

Very nice. All is well. Is one of those the pump that won't start?

 

[KeepItSimpleStupid]

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.

 

[alec_t]

Just to confirm the set-up for those KISS tests:
1) There was only the 24V supply, pump and 0.22 resistor in series? Or was the FET involved?
2) Vpump is the voltage across the pump? Or is it at 'Pump-' ?
3) All pumps ran properly under those conditions?

Good to know the pumps are mutually consistent, but we seem to have lost some Volts somewhere, if we started off with 24 but there's only 17 across the pump?

 

[KeepItSimpleStupid]

Alec:

I'm sure these are true. Joe described his setup in detail. Resistor, wirenut, power supply etc, The supply did have an adjustment on them. A LONG time ago, it was decided to place that adjustment at minimum. 24-10%(24v)-1V is about 24-3.4 or 20.6V. Just pullled the 10% number out of the hat. Could be 15%. Joe did say at one point he didn't have a 0.22 resistor, but later said he had at least an extra one once it was pointed out it was Rsense.

"Across the pump" is at the power cord of the pump. There is some length of 16 AWG (I think that's what I measured) of wire to the pump. Joe would have to confirm the length, but I'd suspect about 3-4 feet or 8 feet, two ways. That's another 30 mV or so.

I SUSPECT the answer to #3 is yes.

Good to know the pumps are mutually consistent, but we seem to have lost some Volts somewhere, if we started off with 24 but there's only 17 across the pump?

Power supply trim should have been set at minimum. You said to do that.

 

[salty joe]

Yes, I hooked the Rsense directly to the + side of power supply and wire nutted the + side of a jack to the other side of Rsense. I hooked the - side of the jack to the - side of power supply. Nothing but power supply, Rsense and pumps.

I took Vpump at the wires soldered to the jack.

All pumps ran just fine. I just checked V at the power supply, it is 23.8.

By RonV "Is one of those the pump that won't start?" Yes, but now they are labeled. BTW, I put a new, brand name battery in the DMM but the Vpump numbers still seemed to climb a bit with time.

Thank you everyone.

edit: KISS your post snuck in before mine. I would have sworn that the power supply was set at 20V. I have 2 power supplies and must have grabbed the wrong one.

 

[alec_t]

Thanks for confirming the set-up.
Ok. Let's assume the power supply was set at 20V for the tests. We got 17V across each pump, 0.2 across Rsense each time. That still leaves 2.8V unaccounted for

the Vpump numbers still seemed to climb a bit with time.

Not unexpected. Possibly due to the pump 'running-in', or the IC in the pump warming up.

Edit: Duh. Just realised the volts aren't really missing, because the 17V is an average reading of a voltage which is switching on/off with the coil commutations courtesy of the internal Hall IC.

 

[ronv]

Alec,

I think you have a good catch there unless the supply measurement was made with no pump running there is 6 volts missing at 1 amp.

23.8 volts at the supply.

Salty,
Was there a pump hooked up and running when you measured the supply voltage?

 

[salty joe]

“Just realised the volts aren't really missing”
That’s a relief-is the time right to try the mk12 mod?
Not pump related, but am I the only one having trouble with the reply box? I usually go to Word then copy & paste to the reply box. No problems at other forums. Just curious.

I am not sure what V the tests, other than the very latest, were run. I will set the V to 22 and run them tonight or tomorrow.

 

[salty joe]

Alec,

I think you have a good catch there unless the supply measurement was made with no pump running there is 6 volts missing at 1 amp.

23.8 volts at the supply.

Salty,
Was there a pump hooked up and running when you measured the supply voltage?

You mean the test I just ran with Rsense, right? If so, yes all measurments were taken on a running pump.

 

[ronv]

Yes, It's probably nothing but the numbers should add up. Since the output of the supply is 23.8 volts and the drop across a typical pump is 17.2 plus the drop across the 0.22 ohm at 0.2 volts. That's only 17.4. That leaves us short of the 23.8 by 6.4 volts. So we worry that there is some resistance somewhere else that we don't understand.

 

[ronv]

I think you could probably make the D9, R13 addition, but we should probably see what Alec wants to do. That change should keep the pumps from stopping after 1 toggle.

 

[KeepItSimpleStupid]

Hey guys:

We have no idea what the current waveform is AND I'll bet that Joe doesn't have a True RMS meter. I'm not going to find that post somewhere, but I think we did have some sort of discussion.

That, by itself can lead to lots of discrepancies, so we can't really trust the voltage or the current, but I think we can say that the pumps should behave nearly identically. It is, by definition an average multiplied by some unknown number.

 

[alec_t]

am I the only one having trouble with the reply box?

No problem here. I just type straight into the box.

We have no idea what the current waveform is

Well we don't know exactly, but I'd put money on it being similar to the waveforms I saw for the BLDC fans I tried when I was designing the pulse-delay speed control. In other words the current waveform is essentially a square-wave (plus switching transients) with a duty cycle of ~85-90%. I didn't check voltage waveforms. Don't know if motor back-emf can somehow get back to the DMM to affect measurements.

I'll bet that Joe doesn't have a True RMS meter

I don't either .

is the time right to try the mk12 mod?

I'd start with Ronv's D9/R13 mod (incorporated in the Mk12) first, as he suggests in post #1156. That may be all we need. If we still have a problem with FETs not switching on fully we can add the other (Q3 etc) mod later. Set the trimmer ~mid-way.

 

[salty joe]

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

I set the power supply to 20.1V, the min. setting. Trimmer is 25.0 K.

V readings with pump hooked up to PDM. Sometimes the pump stalled and sometimes the pump ran.

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

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

 

[ronv]

Looks like we are back to where we were. Lets double check the setup.
Is Vt tied to +12 or to the output of the timer?
The measurements on the fet are all made with the meter ground probe on ground?
The fuse reading is with the meter across the fuse?
And the pump reading is with the meter across the pump?