[SOLVED] Three phase reversal detection and correction

Hi,

I want to design a system to detect phase reversal (in a 3phase systeM) and correct it. 3 phase line is fed to a 3phase 10HP motor (Input AC Line voltage is rms 110v/50Hz) and hence all 3 phase sequence must be maintained even if someone connects L1,L2,L3 incorrectly. I looked online for ideas but I could find ideas for detection but not correction. I guess, we need a set of contact relays which needs to be controlled using micrcontroller based on detection of phase in each line fed-in. So here is the unit block diagram in my mind:


I have some ideas for phase monitor using micro controller (like using 3 opto couplers to detect zero crossing time and taking the time difference between all 3 to find if the phase is in sequence. I am yet to finalize logic though). But I need your help to see if correction unit can be implemented practically. Kindly suggest ideas/links if anyone already implemented it.

You'll interchange two phases. Two contactors required, one for forward and one for reversed connection. Quite similar to a regular reversing motor contactor circuit.

Hi,

for phase reversal/correction: just two phases need to be switched. So a simple relay should work.

But why does it need to change automatically. Doesn´t the motor have fixed wiring?

Klaus

The dashed line symbolizes a mechanical interlock, not necessary required but provides fail-safe protection.

You can do it with two opto isolators and a single D type flip flop.

Take any phase, call it "A"
When "A" goes through its zero crossing in one particular direction, it can clock a D type flip flop.
At that instant, phases "B" and "C" one will always be positive and the other always negative.
That relationship reveres if the two phases are reversed.
You only need monitor for example phase "B" to see if its on its positive or negative half cycle to know if phase rotation is correct or reversed.
If two phases follow on in correct sequence, the third must be correct.

The state of the flip flop then determines if a relay is operated or released.
The relay contacts just transpose phases "B" and "C" to achieve correct phase rotation.
Simple and cheap to implement.

As this is most likely safety related, a simple mechanical device such as a sprague clutch might be a good idea.
The clutch could either lock the motor shaft if it tries to reverse, tripping the motor overload circuit breaker, or decouple the load from the motor if the motor ever runs in reverse. Either would be a pretty fail safe backup.

@ Klaus, having a proper mechanically interlocked 3 phase contactor system is a lot safer in the field than using relays to switch only 2 phases.

I imagine OP is thinking about shipping product to various places and having mains input wiring done by others - thus to guarantee motor direction right from the get go - phase direction must be established and set.

Hi

Let's imagine it's a simple motor (initially without electronics and relay):
Then the effort is rather big:
* power supply to operate relays and electronics
* 1 or two power relays
* control electronics
... means a lot of cost and the risk of malfunction of any of these devices.
(One has to consider how often (many times per minute?) it is switched ON/OFF and how long (24/7?) the device is ON. Also mind the voltage peak when switching OFF an inductive motor ... that all energy goes to the power supply / optocouplers... if no additional effort and cost is taken for overvoltage protection.
It's all a benefit for a (dumb) customer while all the risk and cost is on the manufacturer.

******

But if it's a device that already includes some control logic and power relays:
Then I'd add a rather cheap, low power (capacitive supply) unit that just disables the activation of existing power relays. The motor does not work when wrongly wired. A LED may say "reverse direction".


*******

Also consuder specil cases:
* loss of one phase
* connected to a VFD
* and situations one can't consider now

******
I don't say it doesn't make sense at all. If it's an expensive device or safety related device .. or portable... it may make sense.

Klaus

I was thinking of some kind of potentially dangerous machinery, that might be portable, or mobile, or even rented.
The type of thing that might be plugged in just about anywhere by people with little or no training.

Three-phase monitoring relays are standard products in industrial control, see e.g.

It's a matter of design objectives if you prevent machine start in case of phase reversal (usual) or perform automatic correction (possible). Phase loss would block operation in any case.

Two points about contactor usage. It's not safe to use a single 2xNO/2xNC contactor to switch phases, it likely causes a phase short due to delayed extinction of contact arc. A mechanical interlock provides additional safety against control signal failure and welded contacts.

Easy peasy said:

I imagine OP is thinking about shipping product to various places and having mains input wiring done by others - thus to guarantee motor direction right from the get go - phase direction must be established and set.

Click to expand...

Yes this was my intention.

KlausST said:

(One has to consider how often (many times per minute?) it is switched ON/OFF and how long (24/7?) the device is ON. Also mind the voltage peak when switching OFF an inductive motor ... that all energy goes to the power supply / optocouplers... if no additional effort and cost is taken for overvoltage protection.
It's all a benefit for a (dumb) customer while all the risk and cost is on the manufacturer.

Click to expand...

Hmm. All I am thinking as "protection" is just to use varistor across all 3 hot wires and ground.

FvM said:

Two points about contactor usage. It's not safe to use a single 2xNO/2xNC contactor to switch phases, it likely causes a phase short due to delayed extinction of contact arc.

Click to expand...

I never used interlock before. I googled and understood how it works. So if I have to use interlock then, I guess I need 2 units of 4pole (one pole for holding contact) contactor?


Thank you all for insightful comments. Considering diagnostics and indication point of view and all your inputs I rigged up a quick - schema to get clarity. I want to go with microcontroller approach but will use logic suggested by @Warpspeed. Also as @KlausST mentioned I can take care of loss of phase diagnostic too if required. VFD case is very good point. I am not sure if the same logic works or not. Since this forum doesnot encourage micrcontroller discussion, I won't discuss variable frequency detection in SW here(But I am attaching the logic I plan to do in controller)

I have 3 phase available with me here in office. I'm gonna try this next week with light resistive load that I have. Thanks again for the support.

Hi

vishweshgm said:

Hmm. All I am thinking as "protection" is just to use varistor across all 3 hot wires and ground.

Click to expand...

Sounds simple.
I still recommend to design carefully.
* Connected t the input or at the motor
* expectable energy
* expectable current
* expactable voltage
* expectable duration

Just to avoid that the whole added circuitry causes more trouble than without it.

Klaus

Warpspeed said:

Take any phase, call it "A"
When "A" goes through its zero crossing in one particular direction, it can clock a D type flip flop.
At that instant, phases "B" and "C" one will always be positive and the other always negative.
That relationship reveres if the two phases are reversed.
You only need monitor for example phase "B" to see if its on its positive or negative half cycle to know if phase rotation is correct or reversed.
If two phases follow on in correct sequence, the third must be correct.

Click to expand...

Hi,

So here's the updates on topic:
1. I made a simple PCB with above mentioned logic (the optocoupler output rising edge is detected. I had put 2 relays on board for controlling a 3 phase motor that draws 1A. Each phase LinetoNeutral voltage is 220V,50Hz.
Block diagram is as shown below:


Note that This is just a prototype. Warpspeed's Logic worked like a charm here. Very accurate and No mis-detection whatsoever. I first sense L1 zero cross point rising edge and check the status of other two lines to determine phase reversed or not.
ACR1 and ACR2 are contact switches, whose coil is energized by AC voltage. So I used one of the phase linetoNeutral L2 through my on-board PCB relays to energize the coils. L1 is directly connected to motor, as the motor won't start without other 2 lines.

Also I have a logic in controller to see if any one of the lines are switched off (and thus shut off power to motor) and here's where things get interesting. When one of the PhaseLine(either L1,L2,L3 orN) is switched off without motor connected at the load, I can detect it by monitoring zero crossing on each line and it works fine and I trip ACR2 immediately.

But when the same phase line is switched off during motor run, there is still AC Voltage on the line that was switched off, coming from Motor side (probably due to other 2 lines still running the motor now at high current, imbalance torque). This is the design flaw, I was not aware earlier. Due to this voltage, my zero cross detection circuit still see rising edges and won't shutoff (ACR2) power lines to the motor resulting in catastrophic failure.

So I need your support for this please....

Monitoring zero crossings doesn't provide reliable detection of halfways balanced three phase voltage, practical monitoring relaysare also applying a magnitude criterion.

Your ACR1 circuits runs also into the problems mentioned in post #9, it's not safe to use a single contactor for phase reversal. The circuit can work if the positions of ACR1 and ACR2 are flipped.

Agree with FVM, the original concept was for a phase rotation check only.

Where there is no neutral, as is usual with motors, how about feeding all thee motor phases combined through a single fairly sensitive current transformer ?
If all three phases are evenly loaded, there should only be some small residual harmonic energy present.
This is something I have never actually tried myself, but I suspect a completely missing phase will show a large current imbalance which should be readily detectable.

FvM said:

Your ACR1 circuits runs also into the problems mentioned in post #9, it's not safe to use a single contactor for phase reversal. The circuit can work if the positions of ACR1 and ACR2 are flipped.

Click to expand...

Sorry, I donot undertsand how flipping ACR1 and ACR2 can work but the other way won't. My thought process was that I ensure ACR1 must be switched before closing ACR2, how position of ACR1 and ACR2 matter?
For both relays I am using CA2KN22M7 Relay - Datasheet

Warpspeed said:

Where there is no neutral, as is usual with motors, how about feeding all thee motor phases combined through a single fairly sensitive current transformer ?

Click to expand...

As @FvM mentioned, when one of the phase is cut-off during motor run, I see it's LinetoNeutral voltage drops significantly. In my test, It dropped from 230Vrms to 150Vrms. I rigged following voltage divider and sensed it on my low-cost microcontroller. (I used 47uF capacitor to get pure DC).




So now I measure both adc and zcd to handle phase reversing and correction. It is working fine now.
I am attaching the complete schematic here.

What about this one
It works flawlessly.

ZASto said:

What about this one
It works flawlessly.

Click to expand...

Hi, thank a lot for the schematic, this works. But design that I posted works with only 2 switches (that too only NO/NC). Costwise I would still go for my design, don't you think? And safety-wise I wasnot sure. I thought it is safe as long as I use High voltage switches like this one. CA2KN22M7 Relay - Datasheet. However @FvM disagreed to this.


B.t.w can you please explain a little more? I understand K1 or K2 is switched depending on phase reversal state. But what are PC1 and PC2? Are these supposed to be off-shelf modules that detect phase reversal and switch R1?

Simulations of 3-phase. One has generators oriented correctly, the other has a generator reversed.
Six led's light up depending on polarity of AC cycles.

Notice generators have a white dot indicating orientation.

Generators oriented correctly. Voltage & Amperes display expected waveforms.



Generator #1 reversed. Notice its white dot is adjacent to that of generator #3. Generator volt levels are unchanged. However Ampere waveforms are abnormal.

Just an update : I have delivered this product to the site and its working fine. Thanks @Warpspeed for initial idea. I am closing this thread.