Help with Water Pump

[alec_t]

I don't think Joe's budget runs to multiple wire colours (colors, for you lot over the pond), KISS

 

[KeepItSimpleStupid]

I did find somewhere, https://www.allspectrum.com , (search for wire) a cheap source of 24 AWG wires in multiple colors (colours). It's even tinned copper and a Velleman product. Radio Shack sells the copper stuff which is IMPOSSIBLE to solder too at times.

 

[salty joe]

Yeah, I'm OK with just black wire. Except the 19 gauge I have is blue and red. Not that it matters to me.

The only little trouble I've had getting solder to flow on wire has been with the thick legs on the schottky diodes and to a lesser extent, the 19 guage wire. I run the soldering gun pretty hot so I can get in and get out, it just takes longer for the thick stuff to heat up.
BTW KISS, maybe you got a bad batch of Radio Shack wire. I've had no issues at all with the 22 gauge solid strand from RS.

Hey! I have a few minutes to sneak downstairs and start in on the third driver.

 

[alec_t]

Tinned copper wire is preferable to bare copper, but even the tinned stuff can benefit from cleaning before use. If ancient it can tarnish.

 

[KeepItSimpleStupid]

Silver plated wire is worse. Rare stuff. https://www.apexjr.com/wire.html and PTFE (teflon) insulation is a real trip too. Been there, done that.

 

[()blivion]

PTFE wire.... BAH! sure it's heat proof. Just try cutting it though. You need a special tool that has like two opposing razors and is only good for a few uses.

For PCB work, I use enamel coated magnet wire and sand/scratch/dremal/burn off the insulation. It is better for when you have multiple wires laying across each other that you have to solder, the enamel won't melt. I have been bitten using every day normal low grade plastic insulated wire before. If you have two laying over each other and solder one, they can melt together almost invisibly. Nowadays, I don't lay wires over each other as much. You learn to use both sides of the PCB no matter what wire your using.

Here is something I'm working on right now that demonstrates what I think is "the right way" to use point to point board....

View attachment 66655

View attachment 66656

Not nearly as many connections, but there will be some more soon enough.

 

[KeepItSimpleStupid]

That's a pretty neat idea. I've never found the solvent for Forvar (the "enamel insulation). Scraping and Dremel works.

The Dremel is just a poorly built tool. Too sloppy and wobbly. One model, I had, I told Dremel it's not worth fixing because it's so poorly designed. Then I had to buy an ENTIRE case of stuff because I could not just buy the tool.

Too bad I didn't know about Proxxon. https://www.proxxon.com/us/html/38472.php

Dremel has lots of imprecise attachments.

 

[()blivion]

Yeah, I guess I should have used the term "rotary hand grinder". Mine is not an actual Dremel(tm).

Mine is a cobbled together piece of crap honestly. But one thing that is good with mine is it's a well designed pencil and cable style head, and the end comes apart giving me access to the bearings and such. So what I did is took it apart and put some thrust washers against the inner housing to shim up the bearings. This put some axial force on the races and COMPLETELY removed any and all slop. Basically, the inner part of the bearing is lightly forced one way, and the outer part is lightly forced the other way. Now It has zero movement in any direction except rotation. I'm sure the bearings won't last with the riding surface area reduced by half, but the whole assembly costs ~15$ to get new and I have packed the bearings with real bearing grease. What ever happens will happen, can't use it if it wobbles in any case.

But that's all off topic.

 

[4pyros]

The only little trouble I've had getting solder to flow on wire has been with the thick legs on the schottky diodes and to a lesser extent, the 19 guage wire. I run the soldering gun pretty hot so I can get in and get out, it just takes longer for the thick stuff to heat up.

Thats normal Joe, the bigger the mass to heat up the longer it takes. A varible iron would help.

 

[salty joe]

I have four of these drivers built, but one of the drivers does not work.
**broken link removed**
I have two drivers on one board and since they drive a pair of pumps, the control inputs of the two drivers are connected. The problem driver does not toggle and when I pushed the trip test, the alarm sounded but would not turn off. The other driver is good.

There is nothing touching the heat sink and I could not find any solder bridging or debris between solder joints. Is it possible that a bad fuse could cause this? It showed no resistance, but that's with it hooked up. It's one of those fuses that need baked if moist. The moisture indicater said it was OK, but still...

So before scrapping the it, I wanted to see if you guys had any suggestions to maybe save the driver.

The other pair of drivers on one board works just right.

 

[KeepItSimpleStupid]

Joe:

The fuse would be my first guess.

Verify the positions of D4 and D5.

Check the voltage between S and D of M1. It should be ~24V or whatever your supply is set to, when the motor is not supposed to be running.

If the above passes, jumper S & D and see if the motor LED comes on.

 

[ronv]

The problem driver does not toggle and when I pushed the trip test, the alarm sounded but would not turn off.

If I understand correctly the problem is the alarm won't turn off, but the pump did turn off. If so the latch is not resetting. If you can measure the voltages at input and output of U1 A, B, C, and D.

 

[KeepItSimpleStupid]

we have to watch terms like showed no resistance.

Does no, mean close to zero or infinate? I have to ask. Open, Shorted or close to zero ohms are better terms to use.

And, Joe you can always compare:
e.g. compare the in circuit resistance of the fuse with the bad and good driver.

 

[salty joe]

Thanks guys-I'll run those diagnostics tomorrow.

I did compare the fuse on the bad board to the fuse on the good board. I got the same ohm reading from both. Close to zero. Just not sure if testing an installed fuse for resistance is a valid test.

 

[4pyros]

Thanks guys-I'll run those diagnostics tomorrow.

I did compare the fuse on the bad board to the fuse on the good board. I got the same ohm reading from both. Close to zero. Just not sure if testing an installed fuse for resistance is a valid test.

Yes you can check the fuse installed. The circuity connected to the fuse will not effect a short circuit reading.

 

[alec_t]

The problem driver does not toggle

The tests the others have suggested should indicate whether the FET stays on or stays off. Compare with the good driver.

when I pushed the trip test, the alarm sounded but would not turn off

R5 and R6 form a potential divider intended to bias the voltage of C3 at just below the lower Schmitt threshold voltage of U1a (so that the very first cycle of the trip delay is virtually the same length as subsequent cycles). Because of component tolerances this may not be happening and the alarm would then latch on. Try reducing the value of R6 (e.g. by putting a 10k-15k resistor in parallel).

 

[salty joe]

I had D2 backwards so I replaced it. The alarm still sounds, and now the LED toggles - as a strobe light. It seems to be synchronized with the alarm.

I know I said I'd run the diagnostics tonight, but I'm tired. Tomorrow for sure. Thanks all.

 

[()blivion]

I had D2 backwards

I was literally going to suggest this, but I have other threads I am babysitting and you have tons of help. Sorry I didn't.

 

[salty joe]

Everything looked good with the LEDs and 40W light bulbs, so I replaced a light bulb with a pump. The pump only spins when the power supply is first plugged in. The LEDs continue to toggle happily along. This happened for both the flick and tide.

For this bench test I used the wave driver because the inputs are not connected (there are two drivers to a board-the tide driver's inputs are connected) allowing me to see the flick and tide at the same time.

Any idea what the problem might be?

BTW, I have not spent much time on the problem driver, but I did build two more so I have 5 good drivers. But I guess that's on the back burner now.

 

[KeepItSimpleStupid]

How about those suggested tests?

Did you socket U1? If you did, pull it and see what happens.