Help with Water Pump

[salty joe]

One difference between the Mk14 and Mk15 is the separation of PowerGnd and SignalGnd in Mk15. Joe, check that both are grounded. Also check that Rsense has good soldered connections (if either end has a high-resistance connection the pump won't start).

I had a wire for signal ground connected to a negative terminal on the power supply and a heavier wire for power ground connected to another negative terminal on the power supply.
I feel confident that Rsense has good solder joints. I always make sure to get the solder to flow and if I’m working with thick wires, I always double check to make sure solder flowed all around. All the solder joints are shiny and tight. Thank you Alec, I appreciate you having me check things.
I just finished going over every single connection on my PDM15. I made sure the IC was a 4093, Q1 is a 3904, and Q2 & Q3 are 3906. I could not find any diodes or caps turned the wrong way. I double checked all resistors and caps for the correct value. I looked for debris and solder bridging. Everything appears to be as it should.
I gotta hit the rack. Night all. And thanks.

 

[alec_t]

Everything appears to be as it should.

Ok. The strange thing is that the only significant differences (in theory) between Mk14 as modded and Mk15 are (1) the Power/Sig grounds and (2) the circuit connections/components around U1d. Have you checked that the two inputs of U1d are now separated, with one receiving Vt and the other grounded (at least until the optional Fault module is used)?
One possibility we haven't dwelt on so far/recently (though I think KISS mentioned it many moons ago) is that, now that we've set the current limit at >4A, the thermal fuse may be kicking in. Try a wet finger on it to see if it gets noticeably warm/hot and measure the volts across it when the pump is supposedly running.

 

[salty joe]

QUOTE=alec_t;1098115]Ok. The strange thing is that the only significant differences (in theory) between Mk14 as modded and Mk15 are (1) the Power/Sig grounds and (2) the circuit connections/components around U1d. Have you checked that the two inputs of U1d are now separated, with one receiving Vt and the other grounded (at least until the optional Fault module is used)?
One possibility we haven't dwelt on so far/recently (though I think KISS mentioned it many moons ago) is that, now that we've set the current limit at >4A, the thermal fuse may be kicking in. Try a wet finger on it to see if it gets noticeably warm/hot and measure the volts across it when the pump is supposedly running.[/QUOTE]

I completely forgot those two unused inputs. You mentioned it not long ago. I am sorry. I’m feeling about as sharp as a marble right now. IIRC, a couple attempts ago I grounded both unused inputs. When I get home tonight, I’ll ground one and run the other to Vt. Thanks Alec, another great catch.
I’ll check into the fuse too.

Am I correct to assume D4 should be re-connected before the next test?

 

[alec_t]

Am I correct to assume D4 should be re-connected before the next test?

Try with U1d inputs corrected and D4 re-connected, trimmer at max.

 

[salty joe]

. IIRC, a couple attempts ago I grounded both unused inputs. When I get home tonight, I’ll ground one and run the other to Vt.

I was mistaken there. Pin 13 of U1d has been soldered to the Vt terminal point from the getgo. I grounded fault (pin 12) and now the alarm seems to be working properly. But still the same story with the pump.

I have the toggle set at about 20 sec. run-10 sec. off. Every now and then, immediately after the LED lights up it goes off for a couple seconds, then back on and continues as normal. Could this be caused by my having two 7812 voltage regulators? (one for the wave controller and the other for the tide controller, both controllers share the same board) I did not change the controller board voltage regulator situation yet because the PDM 14 worked with this setup so I thought I’d get to it later. Is it possible that having two voltage regulators for two controllers is the root of the pump problem?
I held my finger on the fuse for three 20 sec. toggles and the pump making a quiet hum but not running. The fuse did not get warm at all.

 

[alec_t]

The presence of two vregs shouldn't be an issue unless their outputs feed a common point (directly or via a low impedance path) and so get connected together. They could fight each other. I'd disconnect one of them to make certain they don't.
The fuse probably isn't tripping, but to be sure can you check the volts across it?
Glad the alarm function is sorted; so now that part is ok try again with one end of D4 disconnected. Measure the volts across Rsense and also at the R1/R2 junction when the pump is in the hum state.

 

[salty joe]

The vreg for the wave controller is disconnected.
One end of D4 is cut loose.
The voltage across the fuse is 0.01 with the pump in hum state, (LED on) and 0.00 when the LED is not on.
The voltage across Rsense is 0.04 with pump in hum state and 0.29 with pump running.
Voltage at R1/R2 with pump in hum state is 10.72 and 11.41 with pump running.
Thanks.

 

[alec_t]

Ok. So there's only one vreg operating, the trip part of the circuit is effectively out of the equation, the fuse is intact and low resistance as it should be, the voltages are as expected when the pump runs.
With 10.72V at R1/R2 and 0.04V at Rsense the FET should be turned fully on, but it's only passing ~180mA. I conclude that the IC inside the pump is switching off in protection mode, or else the pump connections are intermittently high resistance. Are you still using the little 1A-rated power-plugs/jacks as pump connectors?

 

[salty joe]

Yes but I never got an answer as to the amperage rating. It’s this one.
https://www.ebay.com/itm/2-x-DC-Pow...600?pt=LH_DefaultDomain_0&hash=item3a68476150
I have these on hand. I don’t know if they are more heavy duty but they are metal. I’ll solder leads to it and give it a go.
https://www.ebay.com/itm/5-x-DC-Pow...807?pt=LH_DefaultDomain_0&hash=item3a79d57297

If this jack doesn’t do the trick, I’ll cut the plug from the pump cord, tin the ends of the braided wire and screw them directly into the terminal points for the pump. So later today we’ll know whether to get heavy duty jacks or go back to PDM14. The PDM14 has been solid using the same jack.
Really do appreciate the expert help.

 

[KeepItSimpleStupid]

Here is the same style and it's rated at 1A, 12 V: https://www.electro-tech-online.com/custompdfs/2012/12/typ_48402200.pdf and 5000 insertions.

When the pump is stalled, do you notice very unstable readings across the sense resistor when you look at it for say 10 seconds or more? i.e. 1 s on 5 sec off

 

[ronv]

I thought this sounded familiar. I think we had an oscillation problem with this design starting back on page 118 the cause being in this link. https://www.electro-tech-online.com/threads/help-with-water-pump.124250/#post1090043

 

[alec_t]

I think we had an oscillation problem with this design

The oscillation then was in the Mk13 version which included a delay cap C4. That cap has since been removed, but it's possible there is still oscillation from some other cause. If so, it's odd that it occurs with some pumps and not others.

 

[ronv]

Actually I think it was the MK12, (before C4) but I didn't save the simulation.

Either way having more than one version that are going to get a lot more mods is probably not a good idea. Maybe the one that works is the best.

 

[salty joe]

The voltage across Rsense is steady for the entire time the pump is in the hum mode.
Fussy pump A hooked to PDM15 with a metal jack still only runs for the initial toggle, the same as when using the first jack.
I hooked the first jack to my PDM14 and plugged in fussy pump A. It has been toggling along for over an hour. I’ll unplug it before I go to bed in a few hours. So I did not cut the plug from the pump.
The version of PDM14 that I am using has a separate signal ground and power ground. A cap was replaced with a resistor to drop the starting amps. IIRC, we used a 4.7K resistor to drop the starting amps to 4.4A. Sorry I can’t be more specific-I can’t find PDM14 saved in my documents, I hope to find it in this thread and make another copy.
In U1d of PDM15, pin 13 goes to Vt and in PDM14 pin 13 is grounded. Could pin 13 going to Vt have anything to do with making the pumps act up? Would it be worthwhile to ground pin 13 to see what happens?
It looks like the only feature I’d lose by going back to PDM14 is the fault function.
I am very lucky to work on this amazing project. Alec, you have created incredible schematics. KISS, RonV, ()blivion, Pyro, you’ve been amazing too. Thanks all.

 

[KeepItSimpleStupid]

Joe:

With the pump humming, put your finger on the various pin/pins of the IC. i.e Run you finger along the pins staying at least the pump cycle time. It won't matter if your finger touches multiple pins together. If things are oscillating, your finger might stop it. You can also try touching one pin at a time through something like holding a screwdriver by the shank.

 

[alec_t]

Could pin 13 going to Vt have anything to do with making the pumps act up? Would it be worthwhile to ground pin 13 to see what happens?

I don't expect it will make a difference, but try it anyway.

put your finger on the various pin/pins of the IC

.....but make sure you static-discharge yourself first by touching a grounded part of the circuit.

 

[alec_t]

Post deleted

 

[salty joe]

Pressing on the IC's pins didn't seem to make any difference. Neither did grounding IC pin 13.

There seems to be something about the PDM15 that does not agree with half the pumps, which makes me think the rest of the pumps are running on the edge with PDM15.

My copy of PDM14 drives all pumps reliably. I have PDM13 saved in my documents and I'm going to read through and find what changes were made to PDM13 to make PDM14. The only changes that I remember is swapping a cap for a resistor and seperating signal ground from power ground. Anyway, it'll probably take a few days for me to sort this out and get clean copy of PDM14.

 

[ronv]

Mk14

This is what I have for MK14.

Also added what I have for 13.

 

[alec_t]

Thanks for posting those ron; I'm still having problems uploading images. EM is looking into it I hope.