Help with Water Pump

[()blivion]

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!).

True, I never actually tried to do the math on switching frequency, was thinking in my head PWM speeds for some reason. So then... were you worrying more about instantaneous gate charge current? A quick look says that 4000 series logic IC's can source/sink about 3mA constantly, but instantaneous current is much much higher, about 50mA pulsed. I have no idea for what duty cycle though, could be on for ages.

Bah... whichever, the buffer IC is not to much work to add and is certainly going to be stronger. If it were me, I would tie some buffers together and do a few logic gates per FET. Even if we are not talking about >1Mhz operation, flipping our FET's through the linear as fast as possible is the most energy efficient way to go... Actually.... with your speed controller... aren't we basically doing zero crossing switching? Or at least something close? Or is it more like SCR's?

Attached are the models requested.

Thanks, now I can play with the circuit too Not that my input is needed.

But first.... SLEEP.

 

[4pyros]

Thanks also to 4pyros and Ronv for your advice and helping to kick this thing around.

Glad I could help its been an interesting project.
The boards look nice from the top but I would like to see how well you did on the back.

 

[4pyros]

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.

Thanks for the invinte, maybe next year for Kellners demo. I am not to far here in PA.

 

[salty joe]

Thanks Alec for that modified circuit to run six pumps. That is extremely cool of you to give me that option if I need it.

Andy, I'll flip those boards over tonight to get a picture.

Where is Kellners demo? I am 10 minutes west of Fairlawn.

 

[alec_t]

aren't we basically doing zero crossing switching?

We are for the turn off. But then we need to turn on again at the end of the monostable period. So ideally quick switching at that point. However, Joe's now got some nice heat sinks . Gate charge current is limited anyway by the series 33k, which is used with diodes to OR the various FET gate control signals; so maybe the buffers aren't justified. Running a sim I'm only seeing a few mW average dissipation in the FET.

 

[4pyros]

Where is Kellners demo? I am 10 minutes west of Fairlawn.

Joe; Its in Wesley pa. north of Pittsburgh. There Demo nights are the best around. Its 350 miles from my house. But worth the trip.
https://www.kellfire.com/

 

[()blivion]

However, Joe's now got some nice heat sinks .

I saw that... :is jealous:

 

[salty joe]

Here's the back of those boards, 4pyros. The tangle of wires is 3 ICs next to each other on the logic circuit. Since taking the picture of the top of the boards, I've added most of another drive circuit.

 

[salty joe]

Oops, let me try again.

 

[4pyros]

Looks pritty good Joe. Love the bus wire running around the outside of the boards.

 

[()blivion]

It could be a tad cleaner. But it's not worth it to take it apart now just for that. Especially if it works when you plug it in. But if it doesn't work, you're going to have to sift through all that to find that one wire that zigz instead of zagz. And forget having some one else help you spot the mistake via image.

 

[salty joe]

And forget having some one else help you spot the mistake via image.

Not so fast. I can always get a different angle.

By cleaner, I assume you mean to use longer pieces of wire between connections and run them at right angles?

 

[alec_t]

Nice work on the boards, Joe. As you've found, using dot-matrix board does tend to result in a spaghetti nest of overlapping wires!
My preference is strip board (if not using a pcb). Fewer wire overlaps, but it does entail cutting some strips and using jumper wires. The end result looks less spaghetti-like but usually takes up a bit more board real-estate.

 

[4pyros]

Joe; Point to point wirring is an art form you will learn. It mostly has to do with the layout of the parts next to each other to have the least amount of wires crossing each other then thay all can be flat like a printed circuit board. For your first board I dont think you did to bad.

 

[()blivion]

By cleaner, I assume you mean to use longer pieces of wire between connections and run them at right angles?

More or less.

It should be less "spaghetti nest" as alec_t put it. But as he also said, it's harder to do such with that board type. The trick to point to point (dot matrix) board is to lay some wires on the top and some on the bottom, making all of them on one side go one direction, an the other side go the other direction. It's similar to the way you build with strip board only you don't have to cut any strips.

As I said before, I dislike strip board because you have to jump/short some connections and cut/open others. This may or may not be more accident prone, but in either case, you'll fry more things if you forget to cut a connection than you do if you forget to jump one. So I find that point to point board is safer if you end up making a mistake, which as we all agree is almost guaranteed to happen at least once.

It's no big deal either way. What ever gets the job done.

 

[KeepItSimpleStupid]

Yea, the "fun part" or what it's known as in some circles:"20/20 Hindsight"
happens when you have to determine "where did this wire go"?

 

[KeepItSimpleStupid]

Congratulations:

I don't see any duct tape.

**broken link removed**

 

[salty joe]

Thanks for all the comments. I think I better stick to the dot matrix board, even if a stripboard is neater. With work and kids, I can only find 30-40 minutes a day to solder and I'm not fast at it. But I'm nicklel and diming it and as soon as I get six drivers and the wave logic built, we'll see how well these pumps move water. I will let you know. Thanks.

 

[salty joe]

Thanks KISS. LOL!

 

[KeepItSimpleStupid]

I have to do something once and I while. This whole thing worked out really well with "salty Joe" our project manager. Unlike most managers, he let the engineers do the engineering.

But, like most project manages, he's concerned about budget. It's a good example.

I hope I won't have to say it again:
1. You should have used the T42 push in terminals - Extra $40, Tool+100 terminals
2. You should have used connectors - Again, possible crimper, terminals, housings, drill bits
3. You should have used de-pluggable screw terminals - again cost.
4. The potentiometers look as big as the circuit - Next time trimmers (like maybe 10 turn). They are tiny.
5. All black wire? That happened to me once when I was troubleshooting. It turned out that EVERY wire was numbered. Stamped when the insulation was made. Then there was the home made "ALL YELLOW WIRE". That had a different ending.