ECN Forum Forums General Topics of the Trade General Discussion Area Electric motor rotation

Phasing was such a problem to get right that IBM just started putting a phasing jumper in the machines that cared (3p motors) and the ones that might self destruct had phase rotation detectors so they wouldn't even power up if it was wrong. (disk drives that didn't "unload" the heads)
Prior to that we were swapping the phasing in our disconnect box (or some other convenient place). It got so out of hand that there probably wasn't one machine in the country that had all the wires on the lugs they shipped on.
Sometimes we had machines that had it swapped a couple different places in a single machine.

Part of the problem was, do you look for clockwise rotation on the face of the plug or the face of the receptacle? IBM couldn't make up their mind about the from one plant to another so a brand new tape drive from Boulder would be phased differently than a disk drive from San Jose

We always had one particular motor we would look at to see if it was going the right way.

there was a particular Heidelberg sheet press that I had pulled a motor from....I was hoping someone else would be lucky enough to get to re-install it,
but I drew the short straw again .anyway I asked the operator to tell me if it was running right ...he was wrong, and the damage estimate was around $30,000.

luckily his co-workers told the plant manager that I was told the rotation was fine and I was off the hook.

ironically the rotation was on the flywheel behind a panel on the bottom of the press so it was my fault after all...

gfretwell,
In circumstances as you mentioned is it important to acheive the correct rotation the first time?

If the answer is yes, there is a way to do it. Install an encoder on the motor. You will need a sensing device to ensure it runs the direction you chose.
An encoder outputs a square wave relative to the input voltage. Thus while looking at an encoders output with a scope, one direction would be a positive wave and the other a negative wave. I use the word wave loosely as it is not a true wave. It is digital. But they do work, with an encoder and a device to accept the signal and report to the starting mechanism.

JValdes,
Most encoders I know of will output a square wave based on their supply voltage and loading. Dual, bi-phased outputs, usually 90 degrees offset, are required to sense direction. Please provide a link to some that act as you describe.
Joe

J, back in the olden days nothing was really hurt if the motors were running the wrong way. Things just didn't work and at a certain point you might start running hot because the blowers were not moving enough air but usually you couldn't get it doing enough to warm up.
That was when one of the things running was the hydraulic pump that made things go. It got serious when disk drives stopped "unloading" the heads. If you spin up a drive backward the heads will not "fly".
They put in a phase rotation detector that interrupted the power on relay circuit so you just got a red light. You would pull a plug, swap it 180 degrees and plug it back in. That swapped 2 phases.

www.baldor.com
www.wegelectric.com
www.bei.com ?
www.automationdirect.com

Joe...Check within these web sites for the encoder manuals.
Or just Google "encoders" on the web.
There is much info to examin....John

ps....all encoders produce a square wave directly proportional to the input voltage. So if 5 vdc is the source the output will be 5 volts. You see, an encoder does nothing but count revolutions. And is capable of direction selection as the encoder can output a negative or positve square wave depending on the shaft rotation.

Resolvers output a sign wave that also can produce voltage either negative or positive, but are analog, not digital.
I hope the web sites I have provided can be more specific and answer any questions you may have......John

gfretwell.... I am not familiar with your application. I was referring to electric motors and equipment. I am not literate in the inner workings of computers....John

The phase detection circuit is fairly simple. It just makes square waves from the phases and looks at phase A to make sure phase B gets there before Phase C with a simple digital trigger circuit. That way you know if it is safe long before you get the motors moving.

Back in the olden days "computers" (at least the I/O boxes)were really more akin to what you would see in industrial machines with hydraulics moving paper in printers and moving the heads in disk drives. That is where you had the 3 phase motors. Once electronics allowed stepper motors and voice coil head movement we didn't see as many big 3 phase motors but they were still used for disk spindles and such until they got smaller.
The encoders you speak of are used a lot. In the olden days they were magnetic pickups driving a square wave generator, later it was slotted wheels and photo detectors or hall effect devices. In boxes that moved cards or checks they used air sensors to see the card once they got too fast for switches. Optical sensors get dirty too fast.
I really miss the days when things dripped oil on the floor tho. That was when computers were interesting to work on.
The stuff was still pretty fast. A 1403 would print 1100 lines a minute (132 char per line) and that was 1960 technology.
Now it is just "cut open the box and plug in a new one"

OK John, I'll stick with my statement but explain it a little more. An encoder doesn't count anything or send out anything that varies other than the frequency of the output square wave. If you want to sense direction, you have to buy one with a second, quadrature output. This output, viewed individually, doesn't provide any additional information. However, if you were to positive edge trigger a scope on "A" output, and observe the "B" level at trigger,it would be one logic level for CW and the other for CCW rotation. It's a huge waste of an encoder to use one for just direction sensing. Feeding "A" to an edge triggered flip-flop and "B" to a D input is enough for that. But then it would still indicate the last direction of rotation when there is no rotation. So the circuit grows and the application is worthless for anything more than Science Fair. Now if you want to set up one-shots or other means of over/under frequency detection or feed a charge pump or PLL as part of a speed or phase locking servo, there are some apps there. More commonly, they function as part of a position servo system. Here the "A" would be the edge triggered clock of a ? bit counter and "B" would feed the count UP/DOWN input. The counter output reflects a position on a motion axis. On power up these systems tend to have to index to figure out where they're starting out. These circuits can be realized in TTL or older logic but are more likely to be part of a PLC program scanning digital inputs these days.
Joe

IBM managed to do phase detection on one chip. I don't know if they made it or bought it.

Greg,
If you mean one chip with no support components, I don't recall seeing one of those. But you do just mean for sensing phase rotation on the input rather than shaft rotation, don't you?

I've repaired a Time-Mark PLM within the last couple of years and the circuit was fairly simple. By the time I was repairing disk drives, they were all single phase units. I remember the 5KW transmitters I used to install or service having a PLM to make sure the blower motor wasn't running the wrong way. That would've sucked.
Joe