Soft-start for AC motor with an Attiny85?

[Chriss]

Hi!
A friend of mine asked me to make a controller for him to soft-start an AC motor which is around 900w.
He give me some parameter what I should achieve with an uC.
I should power a Triac with a 0-100Hz pulse which would practically rise in time from 0sec to 5sec. When 5sec is reached the signal should be 100Hz.

This is not a problem. It looks like a sweeper.

But I'm not familiar ( have no practice) in AC motors, my friend has, my problem with this setup is:
I can make this sweeper but the motor will probably not run even after reached the 100Hz after 5 sec cos I'm not sure how long should I keep full open the triac to put enough energy into the coil to start to spin the motor.

Am I righ ?

Somehow I should know how long should be the on state of the triac I suppose, it is not enough to generate just a signal of 10-20-30..Hz.
Somehow I feel there in his logic is something wrong.

Can somebody help me make me mor clear the logic what kind of signal should I create to make this think work?

Thanks for any advice and help.

[rooppoorali]

Pardon me because I myself never worked on such a high-level project. Here are some contents that you may find helpful. At least you may get some ideas.

http://www.jetir.org/papers/JETIRBF06002.pdf

https://www.researchgate.net/publication/4297167_Implementation_and_Analysis_of_Microcontroller_Based_Soft_Starters_for_Three_Phase_Induction_Motors

[Ian.M]

Singular "... a Triac ..." implies it could be a single phase motor.  There are several types of single phase induction motors, differentiated by whether the auxiliary/starting winding^* is permanently powered, and by whether or not it has a series capacitor to provide the phase shift, with some switching in a starting capacitor in parallel to provide additional auxiliary winding current and thus torque during start and run-up.  The remaining type, shaded pole induction motors is very unlikely to be used at this power level due to their low efficiency.

If it has a switched starting winding or start capacitor, using any sort of electronic or time delay switching, that isn't likely to 'play well' with a TRIAC soft start circuit, so you'd need an extra TRIAC to manage the starting winding switching.   If it uses centrifugal switching, that should work OK, and your circuit will not need to take over the switching function.

Note that for induction motors with a switched starting winding, (rather than a permanently connected auxiliary winding), it typically has a fairly low duty cycle limit, and can burn out if the motor is restarted too frequently or if the start time is excessively extended, e.g. by shaft load that's marginal for the starting torque of the motor.  For those with a switched capacitor, (whether in parallel with the auxiliary winding's run capacitor or on its own simply in series with the starting winding), it is typically a non-polarized electrolytic and also has a low duty cycle limit as its can size is almost invariable insufficient to dissipate the heat from the I²R losses due to the start winding current flowing through its ESR.  If you let it absorb too much energy, its electrolyte boils and the capacitor vents, sometimes explosively - game over!

If the motor isn't protected by an overload relay, then your soft start circuit needs to perform that duty as well, which it may need sensors for, e.g. for motor speed, winding currents, and stator winding and/or start capacitor temperature.

* The second winding is generally known as the auxiliary winding if its powered while the motor is running and the starting winding if its only powered during start and run-up.  Some reversible motors use identical   main and auxiliary windings connected at one end by a run capacitor, and the other commoned together, the direction being determined by which side of the run capacitor you apply  power to.

[pcprogrammer]

Controlling a triac with a variable frequency from 0 to 100Hz over time sounds odd to me since the triac switching off depends on the 50 or 60Hz mains frequency.

So you would need some synchronization with the mains frequency to make any sense of the control signal?

Just my thoughts, no expert on the subject.

[Chriss]

I was also come acros the thinking about the control signal frequency and the main frequency problem.
However, I still trying to understand what kind of device could this be, and I comes to an idea, it could be probably likely more a 900W dimmer then a real soft - start device.

Could it be more an 900W AC dimmer?

[pcprogrammer]

I know of two ways of dimming mains power with a triac.

One is phase chopping, where the triac is turned on somewhere in the sine and turned off on the zero crossing, the other one is a kind of pwm where a number of half periods are skipped. So the triac is turned on right after a zero crossing and turned off on the next zero crossing. Then kept off for half a period or more to reduce the amount of power, but this also lowers the frequency.

No idea if either is good for making a motor start slowly. I guess it depends on the motor what is needed to make it work.

https://www.homemade-circuits.com/how-triac-phase-control-works/

[Benta]

Big question is:
What kind of AC motor is it?
Induction? Universal?
Until that is answered, all replies are shots in the dark.
But the "0...100 Hz Triac triggering" idea is dead in the water from the start.

[Chriss]

It is an induction motor.
A motor from a home water pump for garden/plant watering.
The type is: W-GP 900,Womax Pro Power.

[Benta]

Seems to be a pretty standard capacitor-run single-phase induction motor.
Soft-start is not easy, and like I said, the 0...100 Hz Triac idea is out.
Why is soft start even wanted? it doesn't save power, just the opposite. DOL start works best.

[Chriss]

I asked my friend why is he needed a soft start for this motor. Still waiting for answer.

It was weird to me from the beginning...

[Seekonk]

The major need for soft starters is limited inverter power off grid and small inverter generators to give the motor time to get up to speed.
Good ones have a relay that shorts the triac after up to speed.

I think Great Scott on youtube has a micro based soft start if you want to save some time. I never looked at it.
NOTE: I just checked and nothing looked like what I had seen. Maybe he pulled it after problems.

Back in the 60's I made light shows with a unijunction triggering a SCR at different frequencies and let the two beat.

[pqass]

No MCU necessary.  You just need in-line power resistor(s) + relay contacts that short the resistor(s) after ~5 seconds.   Choose a small value resistor(s); just enough to stop the breaker from tripping (if that's your problem) but not enough to stall when turned on.

See: https://sound-au.com/articles/soft-start.htm
Specifically, Figures 5 or 6 (transformerless) below.

[Chriss]

Here is how I understood I should make this working:
1. Need to monitor the 0 crossing point on the main
 Probably I can use an opto coupler for trigger an inlut pin on the uC
2. I have to wait for some t1 time and then trigger the triac.
The Triac will shut down on the next 0 crossing point.
These two point from above should happen on every half period, the positive and negative too.
The t1 is responsible how much power will have the motor.
But, there should be a t2 also, which should be probably the needed time to power on the G of the triac, so he conduct properly.
This t2 should be practically be written in the datasheet of the triac.

I don't know if I'm righ with my thinking but something like that should I do.
Or am I wrong completely?

[Seekonk]

Good for transformers, but not motors as shown. Replace resistor with something like NTC 20D20, maybe a couple in parallel.

[Seekonk]

I found the video I was thinking of which was micro based.

[Zero999]

Use a timer relay to bypass a resistor.
https://www.amazon.co.uk/220VAC-Timer-Switch-Module-0-10Sec/dp/B07N2T77C1?th=1

[Chriss]

Thanks for this video.
This is exactly what I was somehow think it should work, but here is everything explained.

The main reason for a soft start to this water pump is not to stress the filter part which is setup on the piping which is under the ground deep in the earth.
When the pump starts to sucks the wather it create a huge vakuum and is stressing the filter part and cos of this effect the filter will go bad much earlier then usually.

I'm not an expert for such of water pumping stuff and this was explained to me why is needed this soft starter.

[Benta]

Sounds like bad design, but I'm not an expert on pump/filter systems either.
It looks like a three-phase pump with proper soft start or even VFD is the answer. Cheaper than digging deep to replace a filter.

[NiHaoMike]

The main reason for a soft start to this water pump is not to stress the filter part which is setup on the piping which is under the ground deep in the earth.
When the pump starts to sucks the wather it create a huge vakuum and is stressing the filter part and cos of this effect the filter will go bad much earlier then usually.

Sounds like the filter is too small for the application. Replace it with a bigger one next time.

Or have some sort of valve to restrict/recirculate flow on startup?

[Chriss]

Can somebody help me to understand the difference between these two opto coupler:
1. Zero crossing
2. Without zero crossing

I saw schematics where they use opto's with zero crossing feature to detect the zero crossing point of the half period of the sine wave.

I dont really understand the difference between these two opto's.
Both of them are LED controlled and have a triac inside...

Thanks.

[Zero999]

Zero crossing means it only switches on when the mains voltage is close to zero. Without zero crossing means it will switch on, whenever the IR emitter is on.

[themadhippy]

alternatively how about  loop of pipe that bypasses the pump with a electrically controlled valve

[Chriss]

Zero crossing means it only switches on when the mains voltage is close to zero. Without zero crossing means it will switch on, whenever the IR emitter is on.

Ok, just to clarify if I understood you correct:
Zero crossing: the triac won't swich ON even if the G is turned on with a signal but the main sine wave is not near the zero crossing point?
Practically the ON signal to the G should be at the momentum when the sine wave is near zero? Otherwise the triac will miss the ON stat.

[Zero999]

Zero crossing means it only switches on when the mains voltage is close to zero. Without zero crossing means it will switch on, whenever the IR emitter is on.

Ok, just to clarify if I understood you correct:
Zero crossing: the triac won't swich ON even if the G is turned on with a signal but the main sine wave is not near the zero crossing point?
Practically the ON signal to the G should be at the momentum when the sine wave is near zero? Otherwise the triac will miss the ON stat.

Yes, with zero crossing the idea is the TRIAC is only triggered, when the main voltage is near zero. This helps to reduce electrical noise, because very little current is flowing, when it turns on and switching large currents suddenly is what generates electrical noise. Just one point: it's not the TRIAC which does the zero crossing detection, but a circuit inside the opto-coupler. It's designed so you don't have to worry about doing the zero crossing yourself, just turn on the IR emitter when you want the TRIAC to turn on and it will do so within a mains cycle.

[Benta]

Even simpler put:
Zero-crossing yes: only usable for on/off switching of an AC voltage.
Zero-crossing no: can be used for phase-controlling an AC voltage (eg, dimmers), but needs additional circuitry.

[Zero999]

Zero crossing just means it only switches when the AC waveform is near zero. It can be used for PWM of a heater to control the temperature. Non-zero crossing can be used for phase control, which can be used to dim an incandescent lamp.

Here are a couple of waveforms showing the difference.

[Benta]

Umm... I think you should point out the different timebases on your plots. Otherwise they are highly confusing.
A tip: don't combine two plots in this way.

[Chriss]

I understood that zero crossing thing.
Thank you.

Im thinking about software part and uC.
To built this project. It is interesting for me. :-)

I will use a small 220/12v transformer for testing.
Is it advisable to detect the zero crossing moment with an uC interrupt i/o pin?
Or is it better to use a timer and check the state of the i/o pin let's say every 10uS or so?
I think there could also be noisy signals and if the signal is feed into an interrupt pin that could do very nasti things .

[Benta]

I understand your interest, but:
this motor is not suited for soft-starting.

[Chriss]

This project getting a new dimension.
I don't care anymore about the pump motor of my friend, I'm interested in learning and get new experience with this. I found this exciting to make this project alive. 😃

[Zero999]

Umm... I think you should point out the different timebases on your plots. Otherwise they are highly confusing.
A tip: don't combine two plots in this way.

How would you suggest I go about showing the difference? If I use the same scale, it's difficult to see the phase control waveform.

I understood that zero crossing thing.
Thank you.

Im thinking about software part and uC.
To built this project. It is interesting for me. :-)

I will use a small 220/12v transformer for testing.
Is it advisable to detect the zero crossing moment with an uC interrupt i/o pin?
Or is it better to use a timer and check the state of the i/o pin let's say every 10uS or so?
I think there could also be noisy signals and if the signal is feed into an interrupt pin that could do very nasti things .

If you use an opto-isolator with built-in zero crossing, then it's not needed. Ideally the microcontroller circuit should be powered of an isolated 5V supply, because it will make testing and debugging the circuit much easier and safer.

It's possible to use a capacitive power supply to power the microcontroller, then no opto-isolator is required and the MCU can do the zero crossing. This does work out cheaper, but it means the whole circuit needs to be treated with the same precautions as mains.

[Chriss]

I don't understand what you mean "if I use zero crossing opto, then I don't need.".
What is I don't have to need?

Bytw. I will everything separate, isolate from the mains so I always stay on the safe side.

With the 220/12V transformer I mean to use itduring the testin and building up the project.
This transformer will be powered through a 220/220V isolation transformer.
Just to be safe.

[Zero999]

I don't understand what you mean "if I use zero crossing opto, then I don't need.".
What is I don't have to need?

If you use a zero crossing opto-isolator, it will do the zero crossing for you, so there's no need to worry about coding it in the MCU.

Bytw. I will everything separate, isolate from the mains so I always stay on the safe side.

With the 220/12V transformer I mean to use itduring the testin and building up the project.
This transformer will be powered through a 220/220V isolation transformer.
Just to be safe.

Just for educational purposes, it's possible to power the MCU off a capacitive power supply.

Here's an example. Connect the zero label in the schematic to an IO pin. Program the MCU to trigger the MCU for 2ms just after zero has gone high or low, which is when the mains goes through the zero crossing point.

[Benta]

Umm... I think you should point out the different timebases on your plots. Otherwise they are highly confusing.
A tip: don't combine two plots in this way.

How would you suggest I go about showing the difference? If I use the same scale, it's difficult to see the phase control waveform.

How about just posting the plots as two images instead of one photoshopped image? Plus a comment for each? That would be a start.
I know that visualizing is not easy, I've been working with that most of my life. But always think about the recipient.
Even though I know about phase-angle and on/off AC control, it took me a bit of time to decode the time scales. Not everyone has a 12800 x 8000 screen.

[Benta]

Just for educational purposes, it's possible to power the MCU off a capacitive power supply.

Here's an example. Connect the zero label in the schematic to an IO pin. Program the MCU to trigger the MCU for 2ms just after zero has gone high or low, which is when the mains goes through the zero crossing point.

@Zero999, have you gone mad?
That circuit is lethal. You might a well say "stick your left and your right index fingers into a mains receptacle."

I'm out of this thread.

[Zero999]

Umm... I think you should point out the different timebases on your plots. Otherwise they are highly confusing.
A tip: don't combine two plots in this way.

How would you suggest I go about showing the difference? If I use the same scale, it's difficult to see the phase control waveform.

How about just posting the plots as two images instead of one photoshopped image? Plus a comment for each? That would be a start.
I know that visualizing is not easy, I've been working with that most of my life. But always think about the recipient.
Even though I know about phase-angle and on/off AC control, it took me a bit of time to decode the time scales. Not everyone has a 12800 x 8000 screen.

Okay, next time I'll post them as two images. I only have a standard HD display by the way.

Just for educational purposes, it's possible to power the MCU off a capacitive power supply.

Here's an example. Connect the zero label in the schematic to an IO pin. Program the MCU to trigger the MCU for 2ms just after zero has gone high or low, which is when the mains goes through the zero crossing point.

@Zero999, have you gone mad?
That circuit is lethal. You might a well say "stick your left and your right index fingers into a mains receptacle."

I'm out of this thread.

No, the circuit is perfectly safe, when properly constructed in accordance with the regulations. Treat the entire circuit, along with everything connected to it, with the same precautions as the mains and all will be be well.

[pcprogrammer]

As long as the other electronics like the mcu are connected to the 5V and ground of this circuit

[Zero999]

As long as the other electronics like the mcu are connected to the 5V and ground of this circuit

Yes, the MCU should be connected between +5V and ground. Obviously the ground symbol on my schematic is just the common for the MCU circuit, not actual safety ground/earth.

I don't know why people panic when they see this sort of circuit. It's fair enough to point out the hazards, but it's perfectly safe if build properly. The original poster is already talking about using an isolation transformer, which is a good thing.

If the polarity of the mains is known, then +5V should be connected to the mains neutral, rather than phase. Obviously this should still be treated the same as live, because broken neutrals aren't unheard of, but it will make it slightly safer.

[Chriss]

Thank you guys for so much valuable infos, I really appreciate it and you helped me really.

I would go back to the post #14 to the this video from YT :
https://youtu.be/HzSlUDdiTSc

I have two questions:
1. The owner of the video put two resistors in parallel R1 and R2, for what is this any good there?
I mean, one resistor is enough to limit the current for the LED.

2. In the video, he is talking about the "firing angle", ok, that is the triac G signal.
If I understand correctly how it is working:
Practically the half wave in 50Hz is 10ms long.
The G signal starts for the 10th ms and is counting backward and is triggering the G in this way:
1. trigger in 10th ms
2. trigger in 9th ms
3. trigger ins 8 ms ... etc. to 1st. ms
and the cycle is starting again.

If this is as I understand, how long could be the High time of triggering signal,
which is represented on the oscilloscope on "Chanel C" ?

Thank you.