Help with Water Pump

[alec_t]

Current limiting to be really effective on motors has to incorporated with a delay. You let the limit be exceeded briefly for start-ups.

Ah, I've come up with a design for just that. How about this?

Explanation: When motor current through Rsense exceeds ~3A Q1 turns on, allowing C1 to charge over a period (~0.3 - 2.5 sec, to allow for motor start-up) set by the trimpot. When C1 voltage reaches the Schmitt threshold of gate U1a the gate output goes low, quickly discharging C2 and switching off the FET. Q1 then turns off and C1 discharges over ~ 20 sec via R5. When the FET is off so is its LED indicator. Gate U1a has its output diode-OR-ed (D4-D7) with the outputs of corresponding gates in other pump modules and when it goes low triggers the alarm module.

 

[salty joe]

KISS, you should have the dead pump in your hands by Wed. Thanks all.

 

[KeepItSimpleStupid]

Got your PM. Seems like there is a wide perception of quality including construction: **broken link removed** On the other hand people are pleased with the price/performance of the pump.

One guy found some bad solder joints on the controller.

I'm not sure if you sent the controller too, for it could have some protections built-in. I wonder if we were all unaware of the built-in controller?

Someone may have been pulling your leg too, because both models support 120/240 volts. **broken link removed**

Here is a thread with a pic of the controller. http://www.reefforum.net/showthread.php?15435-Alterar-Resum-Wave-Maker-15000 If the top right is the power supply and the top left is the motor interface, there is at least a hefty diode-like device (could be diode or TVS) and one or two inductors. The inductors could be used for RFI/EMI or could be used to limit the inrush current.

 

[KeepItSimpleStupid]

This link https://www.reefforum.net/showthread.php?21440-wave-maker-resun-15000-AJUDA is in Portugese, but you can use Google Translate and/or Chrome to read it in English. This particular controller seems to be having "Fried DRIVER" for breakfast too and complaints about poor construction. I signed up for the forum, but it needs moderator action, so I cannot view some of the pictures.

They aren't getting anywhere either and they didn't mention if there is a pump problem too.

 

[salty joe]

Someone may have been pulling your leg too, because both models support 120/240 volts. **broken link removed**

.

I had not seen these units set up for 110 before now. The transformers that came with mine call for 220-240. The seller also warned me that thes units are 220.

I have read similar things about the pumps as were mentioned in those links. I have also read about more than one controller failure. Even if the supplied controller was rock solid, it does not do what I need. I'm kicking myself right about now because I should have tossed a factory controller in with the dead pump. I'll get one heading your way.

 

[KeepItSimpleStupid]

Here is another thread (Russian) https://www.aqualogo.ru/phpbb2/index.php?showtopic=7094&st=0 that seems to talk about 3 voltages for speeds. They are also talking about creating a "nite mode". They have the p/n of the FET like thing which is a switching regulator with an enable.

They seem to suggest that the regulator LM2576HV-AJ ( https://www.electro-tech-online.com/custompdfs/2012/06/lm2576.pdf ) fails in an earlier design due to a poor quality inductor. It does have an output capability of 3A and has cycle current limiting.

At least that's my take.

 

[KeepItSimpleStupid]

Here is an interesting article on Intelligent Fan control: https://www.electro-tech-online.com/custompdfs/2012/06/01178A.pdf

Here is an interesting hall effect IC/motor controller https://www.electro-tech-online.com/custompdfs/2012/06/AH287.pdf EXCEPT it only has 400 mA output capability.

It's really difficult to find hall sensors that operate to 24 V. Here is one such sensor: https://www.electro-tech-online.com/custompdfs/2012/06/AH337.pdf

Note, we need inversion and hysteresis to run the motor.

The first article is useful because of the explanation of the suggested protection components.

It suggests that high currents could happen when the gate is first turned on if there is no resistor to drain off the gate charge. This is where the 3A current limiting of the original controller could possibly remove this requirement. Overvoltages don't appear to be handled with 5 parts. Selected capacitors and TVS diodes are suggested.

Locked rotor:
Benn trying to figure this one out, but it generally mans that the rotor is prevented from moving causing excessive current. Again the 3A current limit of the original controller will help this, but it's not a locked rotor fix. The purpose is really to prevent winding damage. We could be lucky here where the FET blew to protect the motor because of being the weakest link.

It appears that a locked rotor fix detects rotation via the hall sensor. That's the easy part for a controller to do. Recovery seems to be harder. From looking at some other data, recovery could be as simple as when locked rotor is detected, energize coil for 1 sec. Does it rotate? If not, wait 6 sec and repeat.

Guesses at this point:
Lack of current limiting and/or proper fusing contributed to the failure.
Locked rotor is a possibility.
Spikes from the turn-off of the motor coils is another.
The FET(s) just turned on full because the charge wasn't drained away from the gate.

One additional protection device is in order and that is a reverse biased diode across the power supply. This prevents backfeed of the motors into the power supply.
I've known power supplies to die because of this. It may cause the power supply voltage to increase.

1. I need to draw a circuit. Need to speculatet how inversion is done
2. I need to determine if the coils were damaged.

 

[alec_t]

The above links have been helpful to glean some clues.
This link gives a bit more info :-
https://www.google.co.uk/url?sa=t&r...sg=AFQjCNGN67FMbhlaZeRg4-0laauKFmauUg&cad=rja

Thinking about the original controller supplied with the pump, it operates over three power ranges to simulate waves. Presumably it does so by using PWM to ramp the duty cycle, hence the 'average' supply voltage to the pump, up and down (rather than just turning it on and off). It follows that the pump is not running flat out all the time, or even for prolonged periods, in its intended mode of operation. Perhaps the average power is as low as 50% of the rated power? If so, it's real power-handling ability may not be 30W (or 25W?) continuous as we've assumed from the blurb, which would mean we can't run it at 24V for hours at a time as Joe wants .

It's really difficult to find hall sensors that operate to 24 V

If the original controller uses smoothed PWM to apply an averaged voltage to the pump perhaps the pump never 'sees' the full 24V?

 

[salty joe]

Here is another thread (Russian) https://www.aqualogo.ru/phpbb2/index.php?showtopic=7094&st=0 that seems to talk about 3 voltages for speeds. They are also talking about creating a "nite mode". .

The factory controller has an adjustable repeating pattern of low, medium ang high, with IIRC 5-20 seconds at each setting. Not even close to the bulk water movement I'm looking for.
Corals certainly don't need a night slowdown mode. That's presumably so fish can get a good night's rest. There will be plenty of hiding spots for fish to slip into to sleep. Just not next to a pump, which is why the pump flick is critical.

Thank you, thank you, thank you.

 

[KeepItSimpleStupid]

Joe:

Is it possible to give the circuit a workout using:

https://www.amazon.com/Onite-GU10-lamp-socket-holder-base-2pcs/dp/B0066O1RWY Socket
and https://www.amazon.com/Bulbrite-FMW...ref=sr_1_4?s=hi&ie=UTF8&qid=1338829675&sr=1-4 35 W bulb

or similar?

One bulb should suffice until, we can find a better simulated load.

The Infrared light/heat goes out the back of the lamp so you cannot mount the base down close to a reflective surface. I know, first hand. The IR will deteriorate wire insulation except for the high temp pigtail. I'm not sure if these bulbs have an operating angle. Shining down is always good with no flat surface nearby the rear of the lamps.

It's not the same as motor. The resistance will be lower on startup though.

When I know the inductance of the motor, we can find a better simulated load, but it won't have the functionality of operating as a generator.

 

[salty joe]

OK, I'll go downstars and replace the FET and jack that got cooked, chase down a 35W halogen & fire it up.

 

[KeepItSimpleStupid]

So far from what I read about the old controller, the voltages applied are linear (switchmode regulated) and not PWM'ed. We'll find out soon enough. If it was PWM'ed then the hall devices seeing 24V all the time would make sense.

This **broken link removed** under Complementary MOSFET motor controller has resemlences of what I think the motor is doing. With access to both coils, inversion is possible depending on what ends are connected to the supply and the driver. There is two resistors that are either a pull-up, pull-down or series limiting resistor. The hall IC probably has open drain or open collector outputs, +V and GND as connections. Speculation at this point.

We still know what I need to do. If there is no coil damage, I'm wondering if I could design a replacement PCB for inside of the motor?

I'm gonna jump the gun a bit here and actually suggest a more realistic design for randomness, but the current design has to work before progressing to generation final. Although not currently within stuff I know, but I do possess the required prerequisite skill set. The PICAXE, www.picaxe.com is a very cheap processor programmable in BASIC, I did buy a development kit to play with, but so far I haven't been able to devote much time to it. Most BASICS, and probably this one supports pseudo random numbers. What this does mean is that you could get your true randomness that you desire. The technique that's used is that a random number is usually a number between 0 and 1. If you wanted a period between 6 and 10 seconds. A statement such as INT((5*RND)*10)+6 gives a number between 6 and 10.

An interface may not be needed. The constants could just be "programmed". There is really only 4 outputs to deal with.

Another totally radical way to approach this problem is to find an air motor. I don't know if they exist or not. It's just a thought.

 

[salty joe]

Random currents are useful, (Like the very first controller, which I had planned to use once a week at cleaning time to change things up a little) but what I am after is bulk water movement. A pair of pumps that can run for about 30 sec and off for maybe 5 sec for about 6 hr and then switching to an opposing pair of pumps is one thing I'd love to have.

Then another controller to run two pumps where only one pump would run at a time, with run time being adustable between 1-15 sec or so. These two pumps would be opposed to one another. The run time is adjusted to find the resonance of the tank to create a real wave. These two controllers will be the workhorses. This kind of bulk water movement is outstanding for a reef aquarium.

Interesting you would bring up air, KISS. I have goofed around with high powered air lifts and reverse Carlson surge devices and do have experience using air. I used the heavy duty type air pumps common in residential septic systems. Using air introduces a whole new set of problems such as even more humidity to deal with. But the worst problem is heat loss. For some systems that might be a plus, but for me it would end up costing a fortune to heat the tank. The submerged water pumps help me out a tiny bit in the heat dept.
OK, I really mean it this time-I'm going downstairs to see about bringing that poor scuffed up controller back to life.
BTW, you should have the dead pump, factory controller and transformer Wed., at least that's what the USPS told me.

 

[KeepItSimpleStupid]

It doesn't have to be air. It could even be water. Call it hydraulics. If it was water, the system could be closed. The air does not have to exit into the tank. It could be exhausted. A closed hydraulic system might be something to think about.

With all of the things you have planned, an IDEC SmartRelay https://us.idec.com/Product/Programmable_Relays.aspx They don't have a lot of I/O or Input/Output, but they can sport battery backup, an LCD display and an operator interface of sorts. These are programmable in "ladder logic" or a "flow chart" type of language. There is a downside of requiring a "development system" of sorts, so to get started is about $300. Various power options are available as well. You can and I have played around with the language for free in simulation. They may have an astronomical clock as well.

They might be able to offer you a huge bang for the buck in the long run. Daylight timers, automatic feeders, temperature control etc. might be better accomplished with a few of these SmartRelays. They are basically tiny PLC's or (Programmable Logic Controllers)

Since your building and starting from scratch, now is the time to think of that "ultimate system". You can use that framework and add/subtract as you see fit. A good reason of building in this fashion is that when your building, your anticipating future expansion. It may cost very little to add a "hook" for an option now than to re-design it later.

When you use standard building blocks, generally the reliability goes up.

I have a source for outdoor enclosures, but your not there yet.

Industrial control panels are usually built using DIN terminal blocks, DIN mounted power supplies, DIN relays etc. This is a "T" rail system that makes control panel construction more like an Erector set. Stuff is more expensive initially, it's easy to troubleshoot and it's totally re-useable. There is even plastic "wire duct" for raceways within the enclosure.

For instance, let's just say that you build a "dual pump interface". This could be a box that has power in, a contact closure enable, an alarm output, fuses etc. and you put it into a DIN mounted case. So, when you want to add pumps, you grab a few of these interfaces and your SmartRelay, a Power supply and your off. Just connect and program.

When I was working in a laboratory, I made things that were still around after 25 years. The stuff I made was modular and fixable and did not have a single purpose.

An example is a temperature controller.
1. 120/240 is an option - wiring
2. Two fuses were always mounted - one or two was used (120/240)
3. A separate over temperature monitor was optionally installed (room on the front panel)
4. There was a small toggle that disconnected the heater output using a relay. (same relay was used in the overtemp monitor)
There was the ability to wire a ON/OFF/(ON) switch here for power drop out.
ON - Output contactor was enabled
OFF - Output was off
(ON) - Resets the drop out relay
5. The output relay state indicator.
6. There was a small toggle that changed the setpoint to zero or enable.
7. there was a toggle that changed the setpoint to internal or external

That was he ultimate controller. The hooks required were the connectors, a 240/120 control transformer, and an output contactor, an extra TC input hole and a relay socket and some attention to wiring.
8. There was a configurable connector on the rear panel



The fancy controllers (SmartRelay) won't get you anywhere unless the simple one works.

 

[salty joe]

A jack, an FET and cap later, and the controller seems to be up & running. I hooked up 40W incandescent bulbs until I get the halogens.

 

[KeepItSimpleStupid]

Are these 12V 40W bulbs? If they are, then no need to get 12V, 35W halogens.

Keep an eye on the FET temperature.

Next time you have to place an order for something, I want you to get a diode similar to this one: https://www.digikey.com/product-detail/en/SB5100E-G/641-1421-1-ND/2075785 Basically a Schotkey, 5-10 Amps and 50+PRV. This needs to be installed at the output of your power supply, between positive and negative at the supply. The band goes to the positive side. The purpose is to prevent back EMF from reaching the power supply. If there is a sizable amount the fuse will blow, because voltage would be coming from the motor not to the motor.

 

[salty joe]

No, just regular 110 household light bulbs. Thanks for reminding me to check temp.

 

[KeepItSimpleStupid]

I'm going to throw this out to read. https://www.electro-tech-online.com/custompdfs/2012/06/4000nrim.pdf Although it's discussing a 3 phase brushless motor, I think it illustrates what I mean by ultimate. It has all of the bells and whistles where as this 2 phase motor has nearly none of them.

 

[alec_t]

That's a weighty tome! Interesting reading. I see what you mean by ultimate.

 

[salty joe]

With all of the things you have planned, an IDEC SmartRelay https://us.idec.com/Product/Programmable_Relays.aspx They don't have a lot of I/O or Input/Output, but they can sport battery backup, an LCD display and an operator interface of sorts. These are programmable in "ladder logic" or a "flow chart" type of language. There is a downside of requiring a "development system" of sorts, so to get started is about $300. Various power options are available as well. You can and I have played around with the language for free in simulation. They may have an astronomical clock as well.

.

Looks very cool but if I understand this, if the SmartRelay failed everything associated with it goes down.

https://www.amazon.com/Onite-GU10-lam.../dp/B0066O1RWY
Are these 12V 40W bulbs? If they are, then no need to get 12V, 35W halogens

The link is for 120V 35W halogen, and you mentioned 12V 35W. Just want to make sure it was not a typo. I have a 35W 120V halogen desk lamp that could easily be hooked up.