sanyo denki motor problem

[daedalus]

Hi,

I have a sanyo denki p5 100w motor which is being run by an amc 15a6 brushless dc servo drive and it runs fine, however when i stall the motor, i loose all torque, does anyone know why this is happening? The motor drive doesnt kick up a fault, and the current draw is what i would expect (i am limiting current to motor to continuous stall torque rating in order to avoid frying it).

After releasing the motor it doesnt start up straight away, but after rotating the shaft a bit its back up and running.

Is this something to do with how brushless drives / motors work that i am unaware of? One project i had planned for these servos requires that they are run in stall condition, so this is causing me no end of trouble. Would a sinusoidal commutation drive be any better?

[KTP]

So you are just running the P5 and the AMC amplifier open loop? Like applying +10V to the ref+ input, ground to the ref- input and having the motor spin full speed? Is the motion smooth? I am thinking either you have something set incorrectly on the AMC amp or you have the halls/phases wired incorrectly. Even open loop, the motor should keep trying to spin after a stall.

I have successfully run a P5 (1kW model) with a AMC15A8 amp and my Pixie board in a closed loop (PID control) and find the motion to be very very smooth, and of course impossible to stall the motor by any means known to man (well just about).

Has to be something in your wiring/settings...

[daedalus]

sorry typo on my part, the drive i am using is model bd15a8, same as you. I am using a signal generator to provide a square wave for pwm, with dir pin grounded (as i dont have a working pwm source yet). The motion is smooth when its running, and it is producing a reasonable amount of torque.

When i stall the motor it usually continues to output force for a few secs, but it seems to cut out, and when it does so it will not start up again until the shaft is rotated.

encoder u,v,w is wired as spec sheet for motor, and motor coils are wired as:
A - red
B - white
C - black
the drive is set to 120deg phasing <-- dont know what this does..

[KTP]

Oh! No, we do not have the same drives. My drive is a BE15A8 and yours is a BD15A8. Mine takes an analog control input in the range of -10V to +10V, whereas yours takes PWM and direction. Sorry about the misunderstanding.

The 60 degree/120 degree switch refers to the sequence of the halls for phasing. 120 is like 001 101 100 110 010 011 (000 and 111 are invalid states for 120 degree phasing).

Not too familiar with the PWM input amps though...

[Al_The_Man]

Have you tried upping the current switch setting, it sound like you may be tripping out on the current, I believe on these amps the fault light does not light if in this condition.
If you have no load connected to it you could temporarily set the current to max. for a test ( all switches on).
Al.

[daedalus]

i was using a variable psu to control max current to the motor, so that i wouldnt cook it with peak stall current. With the current limit set higher the drive doesnt seem to stall as much. could the power supply be the problem?

I have had the drive on max current whilst i have been testing, and have been relying on the psu to limit current to the motor.

[daedalus]

oh, should this be set to 60 degree or 120 degree phasing?

[Al_The_Man]

Peak stall current is usually never reached with the normal set-up, this is the max current that can be supplied to the motor before demagnetization etc. Most power supplies will collapse, trip a breaker or amp current limit will kick-in before this.
If you are testing this off load, there is not much to worry about in the way of over-current.
I would start with a stable power supply with enough capacity, otherwise you dont know what is going on.
Al.

[daedalus]

The psu im using is 0-20A 0-50v, and i have the motor running ok un-loaded. The problem is i want to use these motors as torque sources in a robot interface im building, and the motors will be running stalled for the most part. To test the motors under this condition i have been clamping the shaft to intentionally stall the motors, and that is when the torque drops off after a few seconds.

I wonder if the drive has some safety feature which disables the output stage if a commutation signal is not recieved after so long? i guess there are few people who actually want their motors running stalled. I doubt its a drive overcurrent, as the psu is currently set to top out at 4 amps, and the drive can run 7a continuous.

That or something is tripping in my power supply (although it still reads the full current draw even when the motor is producing no torque).

[KTP]

"although it still reads the full current draw even when the motor is producing no torque"

This seems to scream that you do not have the commutation signals or motor power phases wired correctly. Maybe try rotating the halls one way or the other (ie, hook what you had for hall 1 to hall2 and what was on hall2 to hall 3, and what was on hall3 hook to hall1). 120 degree should be the correct setting.

By the way, the amp might not be able to power the encoder since I don't think it can put out the large currents that these P5 encoders draw (like 200mA!!) This could be your problem! Perhaps try and power the P5 encoder from a seperate 5V supply, and just hook the halls up to the amp (and connect the neg of the 5V supply to the neg of the encoder and the pin 11 (hall ground) on the amp. This way you are not trying to pull 200mA from pin 10 on the amp which can only source 30mA. Also it looks like pin 9 on the amp needs an external 5V supply, make sure you have that hooked up correctly.


I looked at the amp data sheet, and it seems to have 3 switches that control the current limit, sw1, sw2 and sw3...what do you have these set to?

[unterhaus]

I don't understand why you would want to run the motor "stalled" -- if you mean the motor will be holding position, that is very different. If your robot holds something against a mechanical stop, then you could see this condition, but the robot isn't going to last long. Get a pulley and a rope with a weight on it, and use that to put a torque on the motor. Clamping the rotor is going to exceed the amperage setting of the drive, and the motor.

[daedalus]

Maybe im not explaining what im trying to do...

Imagine a single motor with an arm and handle attached. The idea is to hold the end of the handle, and by using the motor to generate a torque and send varying forces to your hand. Basically the control loop simulates a virtual spring between the actual rotor position and a virtual object, and uses this to generate feedback forces. When colliding with a virtual 'wall', the motor will generate near max torque, and will likely be under heavy load from the user. The rotor will not be spinning, as the handle is not moving.

Basically im a rearcher into haptics, the sense of touch, and im designing a new computer interface using scrounged parts.

an example of a similar interface is attatched. In this one the motors rotate around a cam on a cable drive, allowing forces to be generated in 3 degrees of freedom. Note that the main problem with these interfaces is the thermal issues involved with running these motors in such a way, as currently this design will shut down after 1-2 mins of max force due to motors overheating.

I have a few ideas how to improve the design of these interfaces, however i dont exactly have a lot by way of resources, so im building a few prototypes on the cheap to test things out on.