Danger: Exploding Capacitors

[Eurisko]

Now that I have your attention...

I was rereading Gecko's Step Motor Basics PDF.

Here is some very useful information for anyone considering building their own power supplies to power their stepper drivers :

From Gecko:
"...There is a special consideration if the power supply voltage will be at or near the maximum
voltage rating of the drive. If the motor will be rapidly decelerating a large inertial load from a high
speed, care has to be taken to absorb the returned energy. The energy stored in the momentum
of the load must be removed during deceleration and be safely dissipated. Because of its
efficiency, the drive has no means of dissipating this energy so it returns it to the power supply. In
effect, instead of drawing current from the power supply, the drive becomes a source of current
itself. This current then may charge the power supply capacitor to destructive voltage levels.
If more than one drive is operating from the power supply this is not a problem since the other
drive will absorb this current for its needs, unless of course it is decelerating as well. For this case
or for a single drive it may be necessary to place a voltage clamp across the power supply in the
form of a zener diode. The voltage of this diode must be greater than the maximum expected
power supply voltage..."


Strange, but I've read a lot of posts on The Zone, don't recall seeing this mentioned before.

[cjmerlincnc]

Hi, Yep you can take this info as gospel. The reason you probably havn't heard of it is that higher end servo drives ie: the expensive ones the big boys use, have a thing called a dump resistor built in so when the drive is decelerating the excess energy is dumped across the resistor to ground.

I'm currently building my first servo project using Gecko's and have read nearly all there is on the forum to gain knowledge and this was the only thing that stumped me. My voltage will be 74V which is near the limit of the Gecko drive and am quite concerned.

I'm assuming that I could probably get away without problems because my servo driving the largest load is under 500W so not that powerful and it's only moving a 800mm mill table and if I keep the accel/decel times in Mach3 longer it wont raise the power supply voltage.

The other thing you need to watch out for is using a motor whose voltage rating is larger than the drives rating. Looking on Ebay for motors as I have done, most are rated at about 90-110v once you go above the 24v mark which is a bit above the Gecko drive's rating of 80v.

The thing most people don't realise is that the motor once it's driven by a mechanical load and not the drive ie: during deceleration, becomes a very good generator and if the motor's rated voltage is larger than the drives then you'll end up with a black crispy thing where the drive used to be.

There are several things that can be done with using a drive without a dump resistor built in which is,

Longer accel/decel times fixed in the software ie:Mach3
Connecting a voltage clamp across the power supply or for the electronically minded what about a voltage sensing circuit that either switches in a resistor placed across the power supply or that disables or faults the drive if the voltage gets too high.

Personally I'm going for the 'adjust the software approach' and see what happens.
If I end up with a burned crispy thing where the drive was I'll certainly be posting my findings on the forum to let you guy's know.


Good Luck..

[Eurisko]

Thanks for the additional info.

I'm not sure that I would trust software to do the job. All it takes is one wrong digit in a program to bury a cutting tool where it doesn't belong. That kind of deceleration would surely be bad news.

Zener diodes are very cheap insurance, compared to the cost of 3 Gecko drives.

Does anyone know of a source for a 60 volt zener? Also, what should the wattage rating be?

I realize that I could wire a few lower voltage zeners in series, but I'd rather keep it simple.

[Eurisko]

Found a source for 68 volt, 1 watt zeners.
30 cents apiece from Ocean State Electronics.
I've ordered from them before, great source for hard to find items.

Minimum online order is 10 dollars (before shipping).

http://www.oselectronics.com/ose_p79.htm

And my collection of router stuff continues to grow...

[kreutz]

I don't recommend to use a zener diode as a regenerative voltage dump load (even with higher rated voltage than the expected power supply voltage). Gecko has a diagram on their website for a simple voltage clamp circuit using a power resistor to dissipate excess energy, a lot more useful and safe solution.

You can also use a Transzorb (Transient voltage suppression diode) rated at 1500 W, better than a standard zener diode but having the same effect of destroying itself in order to protect your drive.

[ger21]

Originally Posted by cjmerlincnc The thing most people don't realise is that the motor once it's driven by a mechanical load and not the drive ie: during deceleration, becomes a very good generator and if the motor's rated voltage is larger than the drives then you'll end up with a black crispy thing where the drive used to be.

I'm assuming that you're saying that the motor will generate a higher voltage than the drive can handle? Correct me if I'm wrong here. I know that a servo's rpm is proportional to the supplied voltage. Supply a 90V motor with 70V and you'll get 7/9 of the motors rated rpm. So wouldn't the generated voltage be proportional to the rpm? Or does it have to do with how fast you stop. Wouldn't you have to decelerate faster than you can accelerate to create a higher voltage?

Also, by lowering the accel and decel in the software, you're taking away one of the biggest advantages that servos have - very fast acceleration.

[Eurisko]

Originally Posted by kreutz I don't recommend to use a zener diode as a regenerative voltage dump load... Gecko has a diagram on their website for a simple voltage clamp circuit using a power resistor to dissipate excess energy, a lot more useful and safe solution.

kreutz,

I can't find it on the Gecko site.

I've spent hours googling shunt regulators.

Does anyone have a simple, effective, reusable circuit for this ?

[kreutz]

Originally Posted by Eurisko kreutz, I can't find it on the Gecko site. I've spent hours googling shunt regulators. Does anyone have a simple, effective, reusable circuit for this ?

Here is a document that used to be a Gecko's website (maybe it is still there, I did not check the website) and another document from Parker's website:

[kreutz]

Originally Posted by ger21 I'm assuming that you're saying that the motor will generate a higher voltage than the drive can handle? Correct me if I'm wrong here. I know that a servo's rpm is proportional to the supplied voltage. Supply a 90V motor with 70V and you'll get 7/9 of the motors rated rpm. So wouldn't the generated voltage be proportional to the rpm? Or does it have to do with how fast you stop. Wouldn't you have to decelerate faster than you can accelerate to create a higher voltage? Also, by lowering the accel and decel in the software, you're taking away one of the biggest advantages that servos have - very fast acceleration.

Look at this document, a basic explanation to sudden destruction of the power Mosfets during fast direction reversal or sudden deceleration at high speed.

It is possible to have a motor running over its maximum speed at a certain power supply voltage if "field weakening" is being employed on the drive, so a sudden quick deceleration could have similar results as sudden direction reversal at high speed can produce on other drives (if regenerative current is not controlled). "Field weakening" technique could be applied to Permanent Magnet DC motors (and BLDC/BLAC motors) including stepper motors, even when they don't have a field winding, if the manufacturer uses Field Oriented control on the drive.

The document refers to a possible worst case scenario on modified UHU drives (such a case has not happened so far), but the explanation applies to most of the servo drives used for Hobby nowadays.

[cjmerlincnc]

Hi Eurisko, Yes you make a good point. I never thought of the time a tool bury's it's head in something. I guess I'm going to take advice from those circuits (Thanks for posting them Kreutz). No Burnt crispy things for me then.

Originally Posted by ger21 I'm assuming that you're saying that the motor will generate a higher voltage than the drive can handle? Correct me if I'm wrong here. I know that a servo's rpm is proportional to the supplied voltage. Supply a 90V motor with 70V and you'll get 7/9 of the motors rated rpm. So wouldn't the generated voltage be proportional to the rpm? Or does it have to do with how fast you stop. Wouldn't you have to decelerate faster than you can accelerate to create a higher voltage? Also, by lowering the accel and decel in the software, you're taking away one of the biggest advantages that servos have - very fast acceleration.

One case I read was of a servo knee mill and the knee motor lost power and the weight of the knee drove the motor which over sped and destroyed the drive. Maybe a one in a million chance case but it could happen.

I'm putting a brake solenoid on mine just in case.





Thanks all.

[Eurisko]

kreutz,

Thanks for the links.

One last question...

Why would Gecko recommend a single zener diode for protection ?

[kreutz]

Originally Posted by Eurisko kreutz, Thanks for the links. One last question... Why would Gecko recommend a single zener diode for protection ?

No idea....