Detecting AC Motor Speed 2

[Updraft]

I am interested in detecting motor speed on an AC motor. We do not have the EE expertise so I turn to you in the Eng-Tips forum. We used a meter in the lab to show that there are detectible pulses generated by the motor when we spin the shaft with no power applied. Our question is this; can we take advantage of this to determine the speed of the motor while it is powered and under load?

Ultimately we are interested in being able to determine when the motor or the system needs some sort of attention by sensing the motor speed or some other indicative parameter. Others have proposed using a Hall Effect system, but we are trying to make this as simple as possible.

Thanks in advance!

- - -Updraft

 

[peebee]

If you're just trying trying to detect load, then you could just measure the magnitude of the current. When overloaded, the current should rise sharply. That would be much easier than detecting motor speed, any old ammeter could do that for you, and would seem to meet your need of "determining when the motor needs attention" (could you be more specific there?).

If you do truely need to monitor speed, then you would do better to use some speed sensor at the motor rather than trying to determine it from the motor leads.

 

[Updraft]

Our goal is to have some built-in diagnostic indication of when a remote system needs some attention. It is typically a belt driven system, but there are also direct drive versions. The typical problems occur when the belt breaks or stretches and is too loose. In this case the load on the motor would be below nominal. Another problem is when the system is burdened due to accumulation of debris or even jamming. In this case the load would go up appreciably. We want a way to indicate either of these out-of-normal conditions, preferably one that indicates the type of deviation (high or low) so the maintenance crew has a better idea what they will be facing before they head out.

If sensing current is the best way then is there a cheap/easy way to integrate this into an indicator at the control panel? (We are thinking of LEDs rather than meters.)

- - -Updraft

 

[DRWeig]

Updraft,

Current-activated switches with adjustable setpoints are not very expensive anymore.

Here's a wide variety:

http://www.kele.com/webservices/index/index.aspx?section=4#13

You could also do current transducers and use a transducer to signal 1) Motor spinning, no belt, 2) Motor spinning and loaded, or 3) Motor not spinning (power out).


Let us all know how you decide to go...

Good on ya,

Goober Dave

 

[Skogsgurra]

Current is not a very good load indicator in induction motors. It is the power factor that changes with load and there are plug-in relays to supervise that. Often combined with real (active) power supervision.

Google "broken belt", "load drop" etcetera.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[CJCPE]

Here is another alternative:

http://www.loadcontrols.com/whatisaloadcontrol/whatisaloadcontrol.html

 

[peebee]

Skogs: Re "Current is not a very good load indicator in induction motors" -- granted, BUT... I think current sensing should be OK to differentiate between no-load and overload, wouldn't you agree?

Not sure about sensing the difference between an unloaded belt and a broken belt though, as the motor probably would see very very little difference there -- If sensing a broken belt is key, you'd probably do better to put some kind of a sensor on the belt itself (monitoring the belt rather than the motor).

 

[Updraft]

These are all great suggestions. A big part of what we are trying to do is make use of available conditions. We want to avoid or at least minimize the "extra" equipment to add to the systems. We also do not need the highest resolution detection. As described there are only a few conditions we are concerned about. If we can find something to sense/detect with suitable separation between conditions then we are good to go. Sensing the power per CJCPE's suggestion looks pretty appealing so far, but we must investigate the separation and costs.

Some of our applications are very sensitive to motor speed, thus the comment in my first post. Is there a way, without adding a Hall effect sensor, to sense AC motor speed, preferably by sensing something going on in existing the wires? If we HAVE to add something to detect speed is there anything more cost effective than the Hall effect system?

- - -Updraft

 

[Skogsgurra]

For peebee: Yes, current is good when it is about detecting overload. But the OP did also ask for other conditions to monitor. The power (or power factor) is sensitive to most other conditions.

Do have a look at an Ossanna diagram. It will tell you a lot more about it. There is a discussion on these topics, including a few useful links here:

http://www.eng-tips.com/viewthread.cfm?qid=126286&page=1

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[ScottyUK]

As a broken belt detector we've successfully used plain current monitoring relays. If it is mains motor then if possible use a CT to isolate the transducer from the power circuit. Skogs is right - a true power sensing relay is the best option if the budget is available to buy one.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[electricpete]

If you do an FFT of the current, there will be several clues as to the speed. One is the pole pass frequency sidebands around line frequency. Pole pass frequency = 2 * s * LF. Another is sidebands around RBPF, I think. Another is running speed sidebands around 1*LF. All of these sidebands will likely be very low magnitude, and you will need large dynamic range and log scale to find them. I may have oversimplified the picture, but there are certainly people who use current spectrum processing to determine the speed - discussed before on this forum I think.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[itsmoked]

Current is lousy. Use temperature!!

It's cheaper and more reliable. It avoids high voltage.

If you want really nice. Measure the ambient and the motor temp. Calculate the rise.

This will tell you how hard you are pushing the motor's specs. It can tell how much the load is changing with time. It can show bearing issues, etc.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[DickDV]

ABB and Matrix both make motor load monitors to do just this kind of thing. Do a search. They work very nicely.

 

[roydm]

Current should be fine for a fixed speed motor but if it's on a VFD I prefer power as the current change due to speed is not as linear as the power due to change in Watts caused by the V to freq ratio.
Many VFDs allow you to pick power for the 4-20 mA output.
Roy

 

[Updraft]

By the way, I am talking about motors from 1/4 to 15 horsepower.

- - -Updraft

 

[electricpete]

Some things that might be relevant in deciding between current and temperature:
- current provides earlier warning.
- using current saves you from having to run an extra set of leads to the motor - may be an advantage if the power leads are accessible.
- a set of temperature contacts (thermistor) is pretty darned cheap compared to a current sensor. If current, what kind of current sensor would it be? CT is non-intrusive but more expensive. Shunt provides simple voltage output, but a little more intrusive and perhaps introduce some new failure modes for both the motor and the sensing circuit.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[electricpete]

For sensing current, I guess you could also use something similar to an overload heater. Except use the contact for your indication, rather than for tripping the motor... and setpoint would be down below normal loaded current, rather than above.

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[electricpete]

All in all I'm picturing Keith is right. Especially the smaller the equipment.

Is this to be plugged into 120vac outlet? Where is the remote indication vs the equipment?

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[electricpete]

Although if you go the FFT approach on current, there can be a wide variety of degrading conditions that are potentially detected and analysed. This is somewhat a new and emerging field (monitoring motors and their loads based on analysis of their current).

=====================================
Eng-tips forums: The best place on the web for engineering discussions.

 

[Updraft]

The application is industrial more than residential. Think of a wall or roof fan, a pump for effluent, a mixer or a conveyor. The motor is remote from the control by some distance. Whether the sensor is near the motor or near the control (as it might be possible to do for current, but not for temp) doesn't matter to me so long as we can adequately detect conditions that require attention/maintenance. The indication can be very simple such as a green LED for normal status and multiple red LEDs or a flash code on a single LED to indicate the type of problem (loose belt vs. overloaded motor). Even a piezo speaker with different tones might be appropriate. The indication should be near the control such as next to or even integrated into the switch plate.

Communication between the sensing and the indication should not be a burden either. We were thinking of wireless (Zigbee, bluetooth, others) or through the power line (power line modem) rather than running separate communication wire. Most of our application would be through retrofit so not having to run a separate signal wire is a huge benefit.

As a matter of fact, the retrofit is a huge driver. We want to sense operating conditions that need attention. We want to do this on existing equipment so it should be something simple to attach at/near the motor and at/near the switch. It should utilize as much existing hardware (motor, wiring, switches) as possible. If it is cost effective to replace an existing switch with the new indicating switch that would be okay, but we want to avoid having to change or crack open a motor or running new wires. Opening the wiring box on a motor or attaching something to the motor or splicing it into the wiring is acceptable.

This is ultimately for optimizing service/operation of motorized systems.

I hope that helps give a clearer picture of what we are trying to do.

- - -Updraft