Help with Water Pump

[alec_t]

I see from post 678 that we now know a bit more about the OEM controller wiring/control. Does it make use of pin 5 to shut the controller down?
I'm attaching drafts, for review, of a possible PDM-Mk3 using the 1084 suggested by ronv, and PDM-Mk4 using a (hopefully-rescued) 2576.
View attachment 65081View attachment 65082

Mk3 operation
The top half of the pic is a conventional linear v-reg using an LM1084 (U1) programmed for maximum output voltage by R1/R2. If FET M1 is turned on by control voltage Vt the v-reg turns on, powering the load (pump).
Load current is sensed by R9. Anything over ~2A turns on Q1, allowing C4 to charge negatively. If the C4 voltage reaches the lower Schmitt threshold of gate U2a the latch formed by U2a/U2b is tripped (set). This switches off the FET via D4, and also turns on Q2 via R7 to give an alarm output. The time before the latch trips is settable with the trimpot(~ 0.15-0.6 sec). A push-button allows the latch to be reset. R11/C6 provide power-on reset. R8/C5 provide a slow-start switch-on of M1 (albeit not very slow, to avoid significant FET heating). The 1084 needs a heat-sink to cope with ~1W during continuous running of the pump. R9 should be a wire-wound type (at least 1W rating). If a resettable fuse is used (type not yet determined) it could be thermally coupled to either U1 or M1 as a second line of defence.
Mk4 operation
Not having a Spice model of an LM2576 I used an LT1074 model with the shut-down logic inverted. The 1074 has internal limiting at 6A, similar to the 2576.
As for the Mk3, the LM2576 output is programmed for maximum output voltage. Since there is a shut-down input no external FET is needed to switch the v-reg on and off.
The current-sense and timed trip arrangement is similar to the Mk3, the difference being that U2a controls the v-reg shut-down. The current-sense resistor R3 needs to be placed such that the high ripple-current pulses through smoothing cap C2 don't affect the trip.

 

[ronv]

I guess I'm partial to the analog implemtation as switchers on a strip board can be a problem and without a scope very hard to fix. There also seems to be a lot of "fine print" around the switcher. If you decide to go that way I'll read closer.

Just noticed, you plan to use a portion of the standard controller. Might work with shut down.

On the linear regulator:

Is the soft start FET needed? It only reduces the current for a few MS. Or are you concerned about the 1.2 volt output at shut down?

Does the latch need a power on reset?

 

[KeepItSimpleStupid]

The switcher side of things on the OEM regulator is pretty much nailed, I just have to draw it and measure the inductor.
There is 24 VAC input that runs to a 3A fuse and an 8A bridge rectifier. Then 1000 uf x 2 caps provides the raw DC for the regulator. The diode is standard, the inductor hasn't been measured and the output runs into a 1000 uf cap. An inductor (measured earlier) and a 470 uf cap provides a ripple filter. There is a bypass cap (C19) on pins one and 3 of the regulator. The switcher is heatsinked.

The voltages are provided by a fixed divider and two resistor/transistor pairs (2SC1815's) that parallel the resistor on the ground side of the fixed divider.

The shutdown pin is hardwired on.


Then there appears to be a series resistor (1K) from the AC side of the bridge that power a discrete bridge of 1N4007 diodes that create a raw DC filtered supply using a 2.2 || 220 uf caps. This eventually drives a 7805 regulator.

 

[ronv]

[Quote}The switcher side of things on the OEM regulator is pretty much nailed, I just have to draw it and measure the inductor.
There is 24 VAC input that runs to a 3A fuse and an 8A bridge rectifier. Then 1000 uf x 2 caps provides the raw DC for the regulator. The diode is standard, the inductor hasn't been measured and the output runs into a 1000 uf cap. An inductor (measured earlier) and a 470 uf cap provides a ripple filter. There is a bypass cap (C19) on pins one and 3 of the regulator. The switcher is heatsinked.

The voltages are provided by a fixed divider and two resistor/transistor pairs (2SC1815's) that parallel the resistor on the ground side of the fixed divider.

The shutdown pin is hardwired on.


Then there appears to be a series resistor (1K) from the AC side of the bridge that power a discrete bridge of 1N4007 diodes that create a raw DC filtered supply using a 2.2 || 220 uf caps. This eventually drives a 7805 regulator.

Can you draw a schematic of it? A picture is worth a thousand words since we don't have it in front of us.

 

[alec_t]

I guess I'm partial to the analog implementation as switchers on a strip board can be a problem and without a scope very hard to fix.

I think Joe is planning to build on matrix board rather than strip board, but layout care will clearly be needed.

Is the soft start FET needed? It only reduces the current for a few MS.

I'm guessing soft start is unnecessary, but others seem to think it may help.

Or are you concerned about the 1.2 volt output at shut down?

Yes. Simulation shows that grounding the ADJ pin pulls the output down to 1.2V but results in ~ 12W power dissipation in the 1084 if, for some unlikely reason, the pump failed in a way which left a coil energised during rotor lock. So that has to be halted quickly...hence the FET.

Does the latch need a power on reset?

Perhaps not, depending on start-up conditions; but the sim showed the latch set at power-on without R11/C6. A luxury maybe, but saves having to manually reset.

 

[salty joe]

A luxury maybe, but saves having to manually reset.

That is something I was a little worried about. If this could turn itself back on after a shutdown, that would be a big deal.

I was going to use the same type board as my other work, mostly because I'm familiar with it and I already have several of them.

Thanks.

 

[ronv]

Yes. Simulation shows that grounding the ADJ pin pulls the output down to 1.2V but results in ~ 12W power dissipation in the 1084 if, for some unlikely reason, the pump failed in a way which left a coil energised during rotor lock. So that has to be halted quickly...hence the FET.

I suspose.. My thinking was neither the FET or the Hall would be able to turn on at 1.2 volts so there would already be a nasty failure and since the regulator needs a heatsink anyway my favorite one @ 6C / W would keep it to 110C until the alarm was noticed.

Perhaps not, depending on start-up conditions; but the sim showed the latch set at power-on without R11/C6. A luxury maybe, but saves having to manually reset.

Hmm sorry there is one. I thought I saw it come up latched in one of the sims I ran.

When I made the motor an inductor with series resistance the linear acts strange on power up and down with current limit. ( 2.15 ohms)I haven't had much time to look at it, but it goes away if the slow start FET only removes ground from the motor and not from the regulator. I'm off now but will check back in - probably Sunday.

 

[KeepItSimpleStupid]

Looking at the diagram for the OEM switcher based system, I'll almost bet you can:
1. Add a diode easily across the motor
2. upgrade the cap easily
3. The other diode, not sure yet.

There are a few transistor positions on the board which can be cut away for pads.

There are some 4528's that can be removed. One "possibly" could be populated with your gate. The other could be populated with a "DIP Header" for the few resistors that are needed.

There are a few LEDs on the OEM board, they could also be re-purposed.

I'll bet that the 7805 could easily be upgraded to a 7812.

Most of the logic wiring could be done with wire wrap wire. There is Loctite Tak-pak for fastening the WW wire down.

I can almost see the entire board being repurposed with a Dremel, WW wire and a DIP header.

@Joe
This is an example of a DIP header: https://www.goldmine-elec-products.com/images/G18626B.jpg

The idea is that you can place 1/4 W resistors between the posts and insert into a socket. So there are two 16 pin DIPs and one 8 pin DIP location that could be repurposed. The 8 PIN dip location could possibly be used for a couple of SIP headers for logic signals to the controller.

You take a Dremel tool and cut some or all of the traces away and make them with WW wire. One difference is that each board could have it's own 12V regulator. No big deal.

I can't find a good pic of the DIP header construction technique. Actually look here: https://www.google.com/imgres?start...2&tbnw=147&ndsp=15&ved=1t:429,r:9,s:175,i:281 The bottom right hand corner and the row of resistors are placed in a DIP header about 5 rows up.

 

[ronv]

When I made the motor an inductor with series resistance the linear acts strange on power up and down with current limit. ( 2.15 ohms)I haven't had much time to look at it, but it goes away if the slow start FET only removes ground from the motor and not from the regulator. I'm off now but will check back in - probably Sunday.

Hmm. Every once in a while even simulators get a bad part. C3 in the MK3 has an ESR of 6.7 ohms. S/B few m ohms

 

[4pyros]

I can't find a good pic of the DIP header construction technique. Actually look here: https://www.google.com/imgres?start=1...:9,s:175,i:281 The bottom right hand corner and the row of resistors are placed in a DIP header about 5 rows up.

Thay are DIP Component Carriers. Look down this page;
https://www.physics.mcmaster.ca/electronics/?page=construction.html

 

[alec_t]

@Joe

That is something I was a little worried about. If this could turn itself back on after a shutdown, that would be a big deal.

Not sure if 'big deal' is good or bad We don't want the shut-down circuit setting/resetting itself repeatedly in rapid succession or things could run very hot. We may be able to use an output from the TAM (e.g. the 'flick' output) to reset the latch after a cool-down period of, say, several minutes to try to restart a stuck pump. Would that be a useful feature?

@KISS

Looking at the diagram for the OEM switcher based system

Wish we could . Any chance you could share that with us, KISS?
From you post 688 the prospect of hacking the OEM controller looks good.

I'll bet that the 7805 could easily be upgraded to a 7812.

Unnecessary? The main purpose of 12V was to have a good FET gate drive voltage. If we use the shut-down pin instead of a FET for shut-down then we can use 5V logic for trip timing and control.

@ronv

Every once in a while even simulators get a bad part. C3 in the MK3 has an ESR of 6.7 ohms. S/B few m ohms

My fault for being lazy and picking a cap model at random

 

[salty joe]

@Joe
Not sure if 'big deal' is good or bad We don't want the shut-down circuit setting/resetting itself repeatedly in rapid succession or things could run very hot. We may be able to use an output from the TAM (e.g. the 'flick' output) to reset the latch after a cool-down period of, say, several minutes to try to restart a stuck pump. Would that be a useful feature?

I meant that it would be real good if the pump could try to re-start. Sorry for not being clear. Thanks all.

 

[alec_t]

I came across this, where rotor-lock is detected within 1 sec and shut-down lasts 5 sec. So we're in the ballpark.
**broken link removed**

 

[KeepItSimpleStupid]

That chip might be "pretty close". Note that in the Datasheet it shows a +12 and a +24V application, BUT the datasheet also states it's for +5, +12 applications with a maximum of 18 V. I'm confused! If I remember, the resistors were 520 ohms, again close with difficulty in guessing the polarity. "P-channel FETS are slower which might be a good thing.

After I get some things done, I'll work on the schematic. Hopefully, Joe can post some pics of the controller.

 

[KeepItSimpleStupid]

OEM controller pics

Enjoy!

 

[KeepItSimpleStupid]

OEM Controller pics

Enjoy!

The component view is annotated.

 

[ronv]

It's invisable! Amazing!

Wait.. there it is.

 

[KeepItSimpleStupid]

I got zapped with "your message is too short". They were "quick and dirty" pics. The x-ray view could have been better. The power supply areas are obscured a bit. I didn't even use a tripod.

 

[ronv]

Nice.
Doesn't look like it has any shutdown circuits. Interesting....

Not sure I understand how the 2 transistors work for the 22.8 volts or the control circuits.

 

[KeepItSimpleStupid]

Shutdown is permanently disabled.

There is a voltage divider that sets a 10.25 V value. I think this was with a 30K and a smaller value resistor to ground.
If you parallel the smaller value resister with an even smaller value, then the voltage goes up. The transistors select a resistor to parallel the lower resistor of the 10.25 V divider with, thus increasing the voltage.

Now, I don't know how it happens or which transistor does what, but if T1 and T2 are both off you get the lower voltage. If say, T1 is on, you get the middle voltage and say if T2 is on you get the higher voltage. There is enough data in this thread to figure out which one is which and the resistor values. The calculated and measured values match. A LED will turn on, to which will indicate each active setpoint and power ON.

Missing is:
1) A revered biased diode across the pump
2) A diode to bypass the regulator.

Underdesigned?
a) I don't think the output cap is rated for high ripple.
b) It's rated for 25 V and the max voltage is 22.5.

I could care less about the control circuits: one 555 and two 4528's which are timing chips.