Rotation sensing during coast down

[SteveGass]

1. Is it possible to reliably sense motor rotation from back EMF during coast down to determine when the motor has stopped rotating?
2. Will we see similar or at least some back emf with all typical single and three phase induction motors in the few horsepower range?
3. If the answer to questions 1 and 2 is yes, can it still be done if we use DC injection braking to stop the motor more quickly?

We are looking to design a circuit to do this in a consumer tablesaw and the motors we are using are 1.5hp single phase 120/240 and 3-5 hp single or 3 phase 240/480. I have hooked an oscilliscope up to one of the 1.5 hp motors and it does generate a significant decaying sinusoidal back emf on the supply lines during coast down. The voltage could be as high as 50 volts to start with and decays to maybe less than a 1 volts at low speed - say 60 rpm. The voltage starts large at full speed, then decays until the centrifical switch to the cap closes, then it jumps up again do decay back to zero when the motor stops.

 

[don01]

Stevegass
I dare say what your describing is available it I have seen similiar things in direction reversing motor speed controllers. But an alternative suggestion -there are a number of speed sensors designed to prove that belts on fans etc are intact and they are turning. generally they use a prox switch and a module with a adjusting pot to set trip point and a relay out. this would be simple and low cost

anyway its just a suggestion
regards
Don

 

[SteveGass]

Don,

Thank you for the suggestion. I am hoping to have something that doesn't require any extra cabling to minimize cost. This is for a consumer product and every dollar is critical. What I am trying to replace is a IR opto detector pair that looks through a hole in a shaft spun by the motor. That, however, required three extra wires to be assembled and holding the emitter and detector. Since I already have motor power wires coming to the circuit booard, I figured I might be able to pick up rotation sensing just by adding a zero-crossing detector connected to the motor wires during coast down. I'm just not sure how consistent the effect I see in a couple of motors will be in production.

 

[cbarn24050]

Hi, the short answer is no. Back emf drops off rapidly once you remove the supply.Most people use DC braking on a fixed time.

 

[tmahan]

There are several commercial and industrial quality DC injection brakes that do have zero speed sensing. They remove the supply voltage once the rotor is no longer rotating

 

[SteveGass]

tmahan,

Do you happen to know how these devices are sensing zero speed? It seems like it should be easier to detect rotation with the DC applied than if the motor was just coasting, however, the DC in an injection brake is pretty ugly to begin with it seems like, so extracting the back emf modulation seems like it might not be too easy.

 

[tmahan]

Steve
The brake that I am most familar with is actually looking for a current pulse on the third phase. How braking, and more to your needs, the zero speed sensing would work on a single phase motor I don't know. The brake that we use is only usable on three-phase systems.

 

[jbartos]

Suggestion: If you need more reliable braking to zero speed, then it will be better to look into an electromagnetic brake addition, or use motor that is designed for plugging. The actual plugging, the reverse rotation, can be designed such that it does not materialize, e.g. by a mechanical stop.

 

[jraef]

SteveGass,
In answer to your question 1, No, because the field decays too quickly as cbarn mentioned.

The answer on #2 is no with relaion to question 1, but read on.

3. Yes, you can do it with DC injection braking because now you are providing field excitation again, so back EMF is generated and can be measured. Returning to Q2 then, yes, when using DC injection braking it is there and somewhat predictable, albeit not very accurate.

Now for the disappointing answer to the issue you brought up to tmahan. What you want to do is done all the time by people in the business of manufacturing DC injection brakes. I work for one of those companies and I know how we do it. We have expended significant resources in perfecting a method of detecting zero speed without the need for external sensors, and we found a very reliable and repeatable one. We SELL that feature in our products. No offense intended, but why would I tell you how to do it for free?

I understand that what you want to do is for a consumer product, but what you are asking for is free engineering. One thing I will tell you however, is that most of the power mitier saw products that I see at Home Depot now have DC injection braking built in. I'd be willing to bet that they accomplished it in one of the following 2 ways:

a) they spent the money to investigate and develop it themselves legitimately, or
b) They cheated by buying a competitve product and reverse engineering it!

The good news is that now you know that it CAN be done. Quando Omni Flunkus Moritati

 

[SteveGass]

jraef,

First, it is good news to know that it can be done - however, the real question is can we switch and brake the 1.5 hp motor for $5 or so.

Second, it might be that you had the requisite knowledge and could share it without feeling like you were giving away some type of trade secret. I am novice enough on motor control design to not know what might be common knowledge to others, but just not to me! Unfortunately, it looks like coast down rotation sensing falls into the proprietary category for you. I'd be happy to pay for engineering but it would be nice to find someone who wasn't completely reinventing the wheel. Does your company do engineering consulting work?

On the rotation sensing, since my original post I have electrically captured the emf decay over an entire coast down. What you see is a rapidly decaying (amplitude and frequency) sine wave that you could easily track zero crossings in down to 1 rotation per second or so. The P-P voltage is still 500 mV up or so even at 1 rps, which is about were the last zero crossing occurs. The result was consistent in both single and three phase motors.

If anyone would like to see a screen shot of the emf decay, I would be happy to email them a word file.

steve@sawstop.com

 

[jraef]

SteveGass
I looked at the videos of your product at work. Very impressive. You obviously have already conquered the braking part of this, and trust me, it is no small accomplishment that you are able to stop that saw blade in 5ms.

Now my question for you is why do you need to detect zero speed? If you just relied upon a timer to shut down the brake in 2 seconds or so, it seems to be a given that your blade will be stopped at least by then. 2 seconds of extra time with DC on the motor is not going to do any appreciable harm. The technology that we use needs to wait about 25ms before we even begin to sense motor speed. You will be done and drinking a beer by then, no?

I may be missing something here, like perhaps you are using electro-mechanical braking now and are considering using electronic DC injection braking instead. If so, I may disappoint you. You will be hard pressed to achieve anything near that level of performance with DC injection. The quick rule of thumb is that you can expect DC injection to stop a motor in roughly twice the time it takes to start it. So unless your motor starts in 2.5 ms you may be barking up the wrong tree.

We stress very strongly to our customers that DC injection braking should not to be considered for life / safety purposes unless just to remove the danger of having a spinning blade exposed for 20 minutes of so, as in a lumber mill band saw or wood chipper. Your focus, as demonstrated by the hot dog only getting a slight nick, is definitely life / safety. Quando Omni Flunkus Moritati

 

[SteveGass]

jraef,

It turns out that on a table saw the blade will continue to coast for 10 seconds or longer after you shut off the saw. There are a fair number of accidents during this period of coast down. We therefore need to have our contact detection system active during this period. However, someone wanting to change the blade may touch the blade shortly after it has stopped and we don't want to trip the brake (an aluminum pawl that gets shoved into the teeth of the blade) if it has already stopped. It costs about $60 to replace the (single use) brake cartridge, so we don't want to trip it unless absolutely necessary. As a result, we want some way to determine within a second or so when the blade has stopped. The DC injection braking just has the potential to reduce the exposure time by making the blade stop a little sooner. Now maybe we can get a predictable enough coast down time with DC braking that we can deactivate contact detection after a predeterimed interval.

 

[jraef]

Got it. Actually, that makes you a very good candidate for DC injection. The $5 part seems a bit optomistic though. Our products are not built for 1 phase applications, but since you will not likely be a competitor I will contact you outside of the forum and share some ideas for you to look at on your own. Quando Omni Flunkus Moritati

 

[cbarn24050]

Hi Steve, detecting a table saw blade rotation is easy you just need a magnet with a coil on it, mount it near the blade teeth.

 

[SteveGass]

jraef,

I will look forward to hearing from you.

cbarn,

Thanks for the suggestion. However there are two issues with putting something near the teeth of the blade. First, the blade can be deflected by an 1/8 of an inch or more, so you can't be very close. That makes it hard to accurately detect the small teeth on 200T plywood-type blade. Second, it would require an extra wiring harness and sensor. The prior iteration saw has a IR emitter/sensor pair mounted to look through a hole in the arbor. That meant I had to have wire, sensors and someway to mount it all. That all adds to the cost and it seemed simpler and more elegant to read it directly back from the motor if that was feasible. I have been able to come up with a prototype circuit that generates short pulses through an optoisolator at every zero crossing in the back emf and it can detect rotation down to roughly 1 rotation per second or less where the peak to peak amplitude of the back emf is down to a couple hundred millivolts.