Help with Water Pump

[salty joe]

Nice article, Thanks Alec

+1. Congratulations on your creative idea and thanks for sharing it.

Just got done soldering the left side from here. Thank you Alec for adding a resistor to stretch out the flick function. **broken link removed**

I'm gonna start on the Alec 812 & as soon as I have four copies working, then add your speed control. But the article said that as shown, the speed control would work for .15 A motors. The pumps draw 1.3A. Will I need heavier duty parts?

 

[alec_t]

Will I need heavier duty parts?

The parts for your controller are as in post #905 (modified as per post #938). The only heavier duty bits, compared to the Article circuit, are the Rsense resistor (0.22Ω, 3W) and the FET.
(The Article is general-purpose for wider use).

 

[salty joe]

OK thanks.

 

[salty joe]

I've been getting up a little early every day to solder and am almost done with the driver circuit from post 812. (Thank you Alec for all your brilliant ideas and schematics.)
I can't figure out how to install these little bitty resettable fuses. https://www.newark.com/multicomp/mc36235/fuse-ptc-reset/dp/21T9753?Ntt=21t9753
I want to take care of the - pump and - LED leads when the fuse is inserted. Other than that, all done. I expected to see two leads coming off the fuses. I did notice a silver bar on each end, I'm assuming they get soldered somehow. I looked at the data sheet and did not find installation instructions. Baking instructions, yes.

Anybody familiar with the installation of these teeny tiny fuses? I'm also curious, how in the world do you re-set them?

 

[KeepItSimpleStupid]

Joe. I guess we forgot about that. I'd suggest getting some of these: https://capitaladvanced.com/6000ser.htm I did not pick out a model because it's bedtime. www.digikey.com does sell this brand.

In fact, last weekend, I made an SMT breadboard last week. I need to look at what solder paste I have.

In reality, you use Solder paste such as this stuff from Chip Quick. https://www.digikey.com/product-sea...k-products/solder/1310838?k=chip quick solder

The solder is overnight shipped in a cold pack and stored in the fridge.

I could solder them for you and mail them back using an iron and the solder paste.

The real method is that a Stencil is make and you squeegee solder paste through the Stencil. Then you lay the component on the stencil and put it into a computer controlled toaster oven or skillet. See: https://electronics.stackexchange.com/questions/1184/diy-smt-reflow-toaster-oven-skillet-or

The PTC fuses never really blow. You let them cool and they are good as new.

Think about it. You can try wire solder, but you have to be quick about it.

 

[()blivion]

Think about it. You can try wire solder, but you have to be quick about it.

Yup, Most surface mount parts work just as well with soldering wires onto them as they do soldering them on a board.

View attachment 66451

 

[salty joe]

Thanks KISS for that explanation and offer to solder paste. Overnight express refrigerated solder??? Wow and wow again.

I like your idea of putting legs on the fuse. I'll drink some more coffee first, to make me quick.

Great picture ()blivion. I'll try to make this happen with what I have on hand. Thanks!

 

[alec_t]

To make connections to a surface mount fuse I think I'd try (caveat: I haven't tried!) bending a length of thin tinned copper wire into a hairpin shape with a tight enough U-bend that the chip is held between the two arms of the hairpin, then soldering the arms to the 'silver bars', then snipping the bend off.

 

[()blivion]

To make connections to a surface mount fuse I think I'd try (caveat: I haven't tried!) bending a length of thin tinned copper wire into a hairpin shape with a tight enough U-bend that the chip is held between the two arms of the hairpin, then soldering the arms to the 'silver bars', then snipping the bend off.

Not a bad idea. I would snip after it was soldered in and installed.

One thing I do. I'll take the chip, line it up on the board over a few holes, then take the wire and stitch the legs down with it. Then wire them up, and solder.

To Salty
What ever method you use, you need to be sure the wires in question are already wired to where you want them to go and attached down secure and immobile before you attach your part to them. This makes it more like a PCB pad layout in real surface mount soldering. That or you will have to attach the wires to your surface mount part, then heat sink them so as not to unsolder them again. I find for me that planing it out first is easier than trying to heatsink the joint and hole you don't unsolder it.

For the picture above, I put the tinned, cut, and bent wires in my solderless breadboard and lined them all up first. Then it was a simple and easy matter of soldering on all the legs.

Keep in mind I have been soldering since I was like 7-12 years old. Which can make one quite skilled.

(PS. In case it hasn't been made 100% clear yet, that part you have was intended to be mass production assembled by robots and such, that's why it has no wires. It's a Surface Mount Component. Bottom line, cheaper and easier to automate, quite a bit harder to solder by hand.)

 

[alec_t]

Good advice ()blivion. +1

 

[salty joe]

Thanks for all that great advice, guys! Thanks for suggesting to just put legs on it. I did the hairpin shaped wire and it worked real well-thanks. I would not have tinned the wire first-thanks. And I'm pretty sure I would not have heat sinked it-thanks again. You guys saved me some hassle and time hooking up that fuse, I appreciate that.

I've got the logic circuit from post 367 hooked up to one driver from post 812. Everything seems to be working. It's just hooked to a 40W bulb for now. As soon as these **broken link removed** arive, (thanks KISS for that link) I'll hook them up in place of the light bulb before running the pumps.. When the trip test is pushed, the piezo beep, beeps intead of a steady beeeeep. I like the beep beep much better, nice touch.


Alec, the 150K (R3) you had me install certainly stretched out the flick adjustment which allows for fine tuning, thank you. But the flick adjustment is from just over 1 sec to just under 6 sec. Is there a way to have a shorter flick-just enough to give the prop a spin-and still allow for fine tuning? For C10, I used a 10u cap. The 500K flick pot is linear. I'm all apologies if you've already said, but what does the 50k trim pot in your 812 schematic do? I just noticed it says 50K Trim. I used a regular pot with a knob. It happens to be log.

I decided to definitely go modular with your special speed control. Mainly because they (that would be every circuit) are a challenge for me to build and this makes it easier to track down problems. I was thinking of simply soldering wire leads for my hooks from the drivers to speed control.

You guys are awesome!

 

[alec_t]

I've got the logic circuit from post 367 hooked up to one driver from post 812. Everything seems to be working.

That's good to hear .

Is there a way to have a shorter flick-just enough to give the prop a spin-and still allow for fine tuning?

If 1 sec is too long then you can reduce the 150k (by replacing it, or by putting another resistor in parallel with it; e.g. another 150k in parallel will reduce the effective resistance to 75k). Or you could reduce C10....I think I suggested 4.7uF, but you could go lower.

what does the 50k trim pot in your 812 schematic do?

It sets the time delay (~160mS to ~1.1s) before the trip occurs.

I just noticed it says 50K Trim. I used a regular pot with a knob. It happens to be log.

A trimmer is generally used where you want a 'set-and-forget' infrequent adjustment. A pot is used where frequent adjustments are made (e.g. a volume control for a hi-fi). Trimmers are usually much cheaper than pots, too. I explained the log/lin usage in an earlier post. It's of little consequence here.

this [modular] makes it easier to track down problems

A point which I think KISS made, and with which I agree.

I was thinking of simply soldering wire leads for my hooks from the drivers to speed control.

That will do, but both KISS and I recommend terminal blocks, as you know. If wires are constantly moved about (likely at the prototype phase) they tend to fracture.

Good luck with the speed control build. I do hope it works with your motors. As I said, I've tried it with two differently-rated (albeit small) ones, one of which I definitely know is a 2-phase type (similar principle to yours). Success depends on there being a detectable drop in current with each commutation inside the motor.

 

[KeepItSimpleStupid]

Nice to hear that construction seemed to go "a little better" this time.

Do we get to see some pics of your latest creations?

 

[salty joe]

Thanks for answering all those questions Alec. I switched out that 10uf For a 4.7, and sure enough the flick time decreased considerably. I might end up trying a 2.2uf to decrease it a little more. I think I might use a trim pot there. Thanks for telling me the difference between a trim pot and a regular pot.
Thanks KISS, I'm glad the construction went better this time too. I think my biggest enemy last time was pushing myself when I was tired.
I plan on making five more drivers; there will be two to a board. I also plan to add the logic circuit from the wave controller to the same board that the logic circuit shown is on.
I'll let you guys know how the speed control works out. It might be a little while though.
It's been great hanging out with you guys. I have learned a lot about electronics and it makes me appreciate just how incredible your knowledge and creative ability is.
A very special thank you to Alec. Thank you for your incredible amazing schematics, top-notch advice, open-minded attitude, ease of style to work with, thank you for not giving up on the project when we ran into the inevitable snags and thank you for not giving up on me when I made a lot of bonehead mistakes. You made this happen.
I am also very grateful for the help I got from KISS. Thank you for taking the time to analyze and perform an autopsy on the dead pump, and all the useful links and information and advice. Thank you for the envelope of goodies.
I'd also like to thank ()blivion for lurking about and showing up at just the right time with killer advice and pointers and humor. Also thank you for that slick soft start way back when.
Thanks also to 4pyros and Ronv for your advice and helping to kick this thing around.
Gentlemen and scholars, every one of you!
Joe

 

[KeepItSimpleStupid]

Joe:

They look really nice.

Bet those boards were copper and not HASL (Hot Air Solder Leveled) or some such nonsense.

I mix of Vinegar and salt will totally clean the copper oxidation and you can scruff things up a bit with a Scotchbrite bad. The grren pads you use for scrubbing dishes.

Something that I have done on occasion is this. Warning, it's wierd. I've cleaned the boards as a above.

Then get some "peanut oil" and a frying pan and something to sort of keep the PCB from laying on the bottom of the pan. You can heat the board to the point that solder will melt. Take it out quickly and apply solder all over. Then wipe with a rag and then reflow it back in the oil.

I've used this before boards have been drilled.

It might not be appropriate or could be more of a pain to do the solder re-flow or your board may be totally devoid of traces on the back. Dunno.

I do know that the Vinegar and salt mix will clean the board and so will Scotchbrite. That you should try.

 

[salty joe]

That makes sense to shine up the copper circles before soldering. Thanks, I'll do that from now on. Probably good luck so far, but the solder beads up on the copper circles without bridging and everything seems real solid. But yeah, I'm going to scotchbrite now.

BTW, If any of you guys ever find yourselves in NE Ohio and would like to stop in for lunch or dinner, let me know. I will make you welcome in my house.

Also, if you get stuck on a fixit project around the house I will help out with any advice I can come up with. Not saying I have all the answers or anything, but I've been swinging a hammer for a while and have repaired my share of unusual problems.

 

[alec_t]

Just don't overdo the Scotchbrite, or you'll have nothing to solder to .
Thanks for the lunch etc offer. One day..........?

 

[alec_t]

A while back you mentioned the possibility of using 6 pumps. I've expanded the Tidal timer and Alarm sections of the system accordingly. Asking a CD4093 gate to drive 3 FETs is a tad optimistic (even 2 is pushing it) so I've added U4, which is a CD4050BE hex non-inverting buffer, between the 4093 and the FETs. If you only want 4 pumps (2 pairs) you can omit it. I've shown the pin numbers of the 4050 as the pin-out for the supply is unusual.
View attachment 66521

 

[()blivion]

Asking a CD4093 gate to drive 3 FETs is a tad optimistic (even 2 is pushing it)

It wasn't before your speed control circuit. But now we gota actually switch them FET's often.

But... you knew this already.



BTW alec_t, can I get your LTspice potentiometer model and SPDT switch model from you? And sub circuit files. Thanks.

 

[alec_t]

But now we gota actually switch them FET's often.

Well if we assume for convenience the pump max rpm is 1800, then for a 2-pole 2-phase motor that's 60 pulses per sec. Agreed, switching more often than before; but still s-l-o-w by most FET-switching standards (e.g. conventional PWM) and doable without having to use high-current drivers for the FET gates (I hope!).

Attached are the models requested.