Help with Water Pump

[salty joe]

Joe:

Will you order f the following parts as well:

(8) 78K 1% 1/4 W resistor
(4) 4K 1%, 1/4 W resistor
(4) 10K, 5% 10-15 Turn trimmer, something like this: https://www.digikey.com/product-detail/en/PV36W502C01B00/490-2888-ND/666515

This will allow you to modify the OEM controller to output voltages from about 8V to <25V and if a 78K is put in parallel with the 78K, the range will be approximately 4.65 to 13V

This should definitely give you the ability to test. I can't write up modification instructions today.

Are these parts close enough to work as planned?

https://www.ebay.com/itm/50-x-Resis...460?pt=LH_DefaultDomain_0&hash=item3a677a316c

https://www.ebay.com/itm/5-x-10K-OH...775?pt=LH_DefaultDomain_0&hash=item3cbb4905d7

https://www.ebay.com/itm/50-x-Resis...538?pt=LH_DefaultDomain_0&hash=item3cb8b628f2

 

[KeepItSimpleStupid]

Nope for the resistors. The tolerences in this case matter. For instance if the pot was 10% high and the divisor was 5% low then it could be off 110/95 = 1.15 or 115% so a setpoint of 24 would end up being 27.

I tried to get the widest range of adjustment possible while keeping the top end near 24V. I was not able to accommodate the entire expected range desired unless I paralleled a value. The resistors are cheap, on the order of 0.10. For the POTs, at most you would need is 6 anyway. If alec has his way then you won't need anyway.

I linked to a 5K POT, the value is 10K. Here is that one: https://www.digikey.com/product-detail/en/PV36W103C01B00/490-2875-ND/666502 A tolerance of 10% would be OK for this since 5% is too expensive. The pot you picked out would be OK because it's +-10%.

Joe:
I attached a spreadsheet where you can play with the tolerances. You can see from the formula what numbers would be bigger or smaller. It's currently set for the parallel combination of the fixed resistor. You can play and see how the numbers work out. The OEMC does use 1% resistors/

 

[salty joe]

I envisaged all the speed control stuff as a separate module for handling 4 pumps. Since the design isn't finalised I'm not sure how many components will be used (but I've re-jigged the design to minimise components, hopefully), so I suggest you allow three 3 terminals of a multi-way terminal block for hooking into the speed module. Presumably you will have a terminal block for making all connections to/from the post #812 PDM?

I was thinking of soldering the parts on the same board. If that's not such a good move, is this the kind of terminal block you had in mind?
https://www.ebay.com/itm/3-x-DG300-...645?pt=LH_DefaultDomain_0&hash=item3a77d4373d
Thanks.

 

[KeepItSimpleStupid]

At a minimum, yes.

But take a look here: http://www.tycoelectronics.com/comm...CS_1308389_EUROSTYLE_TERMINAL_BLOCKS_0607.pdf on PDF page #7 and look at pins for unshrouded header.

These are screw terminals that unplug easily. I've been there and done that.

When you use stranded wire on some terminals, the terminals need "wire protectors". It is a flat piece of steel that doesn't mash stranded wire. Professionally, when i do terminal wireing with stranded wires, I also use wire ferrules. These are compressible little tubes that effectively turn a stranded wire into a solid wire under the screw.

The screw connectors are very appropriate for power and the motor contacts, but remember a backwards wire means poof which is where the depluggable version is also good. Using extra pins say for power in and motor out also helps to identify what connector goes where. i.e. Use a 3 pin version (add ground, so your not guilty) for power and a 2 pin version for the motor.

I sent you a sample of a tiny female plug and a header. That could be used for the logic signals from the sequencer to the motor controller. These were not polarized. Any kind that you get should be.

 

[alec_t]

I was thinking of soldering the parts on the same board.

Your choice. On the plus side it avoids some wires/sockets. On the minus side it's not so 'modular'; hence less future-proof.
Here's a rough idea of the board space needed per pump for the Pulse Width Section (PWS) for speed control:-
View attachment 66178

And here's the latest incarnation of the full Speed Module Mark 1 (SM_1). Note that it is made up of four PWS (only one shown, together with part of a PDM to show how they interconnect) plus one Speed Set Section (SSS):-
View attachment 66179

Circuit explanation
At each commutation the current through the pump motor briefly drops, so Isense goes low. This triggers a monostable circuit made up of op-amps U1a, U1b (acting as comparators), C2, R3.
C2 voltage initially jumps high, 'SpeedCon' goes low and FET M1 turns off completely; so the motor current is now zero. C2 voltage then drops back. When it drops below the 'spd' voltage the output of U1b goes high and M1 turns on.
Thus the start of each current pulse through the motor coils is delayed by the monostable period. The result is reduced average coil current, reduced average voltage across the pump, and reduced motor speed.
The 'spd' voltage is set in the SSS.
When 'Fast' is connected to +12V the 'SetMaxSpeed' trimmer, in conjunction with R8, allows the average voltage across the pump to be set up at 18V (i.e. the average used in the OEM controller and therefore deemed safe for long-term reliability).
When 'Fast' is connected to ground the 'SetMinSpeed' pot allows a suitable low speed to be set up.
The 'Fast' input can be controlled by a mechanical switch, by a logic circuit or by a PWM waveform.

Since the LM324 is a quad op-amp package it can be shared by two PWS if you like.

Observation
In designing this module I searched the web for speed control techniques for BLDC motors but didn't come across anything using the method proposed here. The published methods (other than the OEM method of reducing the DC voltage supply to the motor) all seem to involve greater complexity, e.g. using a micro with indirect (back-EMF or optical) detection of rotor position, or using a dedicated many-pin IC with direct access to the coils. So I think this method may be of general interest and I'll submit it in modified form as an Article.

Edit:
BTW I've tried the circuit with two differently-rated BLDC motors (albeit not as powerful as a pump motor) and both could be smoothly controlled down to low revs. As this method reduces the average motor current, not instantaneous current, there is still good torque and consistent starting even at low speed.

 

[()blivion]

Observation
In designing this module I searched the web for speed control techniques for BLDC motors but didn't come across anything using the method proposed here. The published methods (other than the OEM method of reducing the DC voltage supply to the motor) all seem to involve greater complexity, e.g. using a micro with indirect (back-EMF or optical) detection of rotor position, or using a dedicated many-pin IC with direct access to the coils. So I think this method may be of general interest and I'll submit it in modified form as an Article.

Indeed, You have effectively got inside the motor without even opening it. Very cleaver Alec_t.

 

[alec_t]

Thanks, ()blivion. I can't believe something so simple hasn't been done before: but I couldn't find evidence of it .

 

[()blivion]

Yeah, I have never seen it before.

 

[4pyros]

Nice job Alec_t. If we get an Article out of all this then maybe it was all worth 908 or so posts.

 

[KeepItSimpleStupid]

alec:

I wonder if a variation on this technique could work for a AC PSC motor that's used in ceiling fans? i.e. Run the phase shifted winding at full speed then zero cross detect at where an triac would naturally turn off. Now delay the start of the next firing of that coil.

Do the same for the phase shifted one.

In a PSC motor,one of the windings has a slightly different winding resistance depending on whether it's the one phase shifted by the cap or not. Somewhere I have a paper to that effect. I think the phase shifted one is lower because the impedance of the other winding is part capacitor.

 

[salty joe]

Congratulations Alec for a novel idea. And thank you for sharing your creativity. Thank you a LOT!
I don't understand what you did, but I do know that new ideas are sometimes worth money. Maybe you could patent this BLDC motor idea.

 

[ronv]

Slick, Kinda like a triac dimmer.

 

[salty joe]

Nope for the resistors. The tolerences in this case matter. For instance if the pot was 10% high and the divisor was 5% low then it could be off 110/95 = 1.15 or 115% so a setpoint of 24 would end up being 27.
/

Would a 75K resistor installed inline with a 3K resistor be the same as using a single 78K resistor?

 

[4pyros]

Would a 75K resistor installed inline with a 3K resistor be the same as using a single 78K resistor?

Yes thay add in series.

 

[salty joe]

Thanks 4pyros.

Somewhere on this thread is a link to a wire chart. The "search this thread" doesn't work-the search bar gets behind the first post on the page. So I searched the web and found plenty of wire charts but they had different amperage for chasis wiring and power transmission. Would an electronics project like this be considered chasis wiring or power transmission?

I'm trying to figure out if 19 AWG single strand wire is heavy enough for the + & - leads to the pumps, as well as the main ground loop. The pumps are rated at 1.3 A. They will be starting and stopping a lot.

 

[KeepItSimpleStupid]

SJ, resistors do add and so do the tolerances

Meaning:
75K 5% in series with a 3K 5% means it's a 78K 10% resistor now

The issue with the formula
1.23 * 1 + (78K 1%)/(4K 1%+ 5K 10%)

You can see that 78K+1% and 4K-1%, and 5K-1% gives you the highest voltage and

78K-1%, 4K-1%, 5K-10% gives you the lowest voltage.

Not sure about dividing.

78K||78K will be used to get feed voltages.


The wiring is considered both. It basically means wires close together and protected, so your gizmo is called chassis and the wire from the gizmo to the pump and from the power supply to your gizmo is called power wiring.

I would size for 6 A. Not sure what everyone else thinks. You don't have control over the wires used with the pump. I'm not sure what the pump AWG is. My GUESS is #20 or #18. Probably more like #20. I measured it at 20 AWG.

Wire length matters too. The farther away, the larger the wire should be.

Here is the chart: https://www.powerstream.com/Wire_Size.htm
The reason why the "power transmission" is higher is that the wire is not enclosed. So

 

[alec_t]

@KISS
Hmm, I don't see this method working with an AC ceiling-fan motor because the coil pulses are necessarily at (fixed) mains frequency. I was under the impression that ceiling fan speed is varied by changing the number of poles/coils ?

 

[KeepItSimpleStupid]

No it's not. A PSC motor consists of 2 coils and a Cap that phase shifts one of them. The CAP value is changed to change the speed.

In other ceiling fans taps are changed. Here are wiring diagrams for 3 types: https://www.hurontel.on.ca/~taitg/pages/cfan.html Changing the cap seems to be the most common. Here's some equations used in the design: https://www.electro-tech-online.com/custompdfs/2012/08/psc.pdf

 

[Dean Huster]

Really? How can a thread with 918 posts be attractive to anyone but the few who are braiding it? What had made this such an attractive thread? Would someone please begin a NEW topic with a similar title to provide a synopsis of this thread so that I don't have to wear out my mouse and eyeballs over the next two weeks reading each post? I"m not being hateful or negative ... just curious (and lazy) .... well, and I am a curmudgeon.

 

[()blivion]

Really? How can a thread with 918 posts be attractive to anyone but the few who are braiding it? What had made this such an attractive thread? Would someone please begin a NEW topic with a similar title to provide a synopsis of this thread so that I don't have to wear out my mouse and eyeballs over the next two weeks reading each post? I"m not being hateful or negative ... just curious (and lazy) .... well, and I am a curmudgeon.

Not that I'm accusing you of suggesting otherwise, but this was not something we chose to have happen. It was a natural phenomenon that simply occurred. But even if it was intentional, consider this...

(1): Mod's/admins don't seem to mind at all (that we are aware of)
(2): The OP is still more or less getting the help he needs.
(3): We are not bothering anyone else on the forums.
(4): There are search functions, and things can be reposted if need be.
(5): People are free to participate, or leave if they want without question.
(6): It will take as long as it takes, regardless if it's in a hundred threads or one.
(7): Having things in one thread makes it all contained to one spot, which I think may be preferred.
(8): Believe it or not things are getting close to some kind of conclusion.
(9): It's more or less our problem, not really yours. You don't need to be worrying about it that much.
(10): You are 100% correct, this thread is far to long. But whats done is done.

In the end, adding a summery, synopsis, making a new thread, or anything like that sounds like a great idea. However, it's actually adding to "the problem" because new posts are new posts no matter what or where they are (even this post). What actually needs to be done is people need to post less "one line" comments that mean nothing and start posting more concise and thorough comments. Or just not post at all. This way the important comments will be closer together and there will be less garbage to sift through.

As to why this thread is so attractive.... I have no idea myself. It just is for the time being.
-()blivion


Edit: I also believe there is a way to make more posts show up on one page. Then you can just Ctrl+F then type # and the post number. But that's assuming you know exactly what your searching for.