# uniform angular motion using a 200 steps motor

V

#### Vittorio Mattiolo

We need to turn a round platform ( diameter =3D mm.100, height =3D mm 50 = and weight =3D 300g) by using a stepping motor having 200 steps. We
need to have a constant speed of 600 turns/minute and a very uniform motion.

We drive the motor by a microstepping controller by Toshiba.

We measure the motion uniformity by means of a laser mounted on the platform. When the laser beam hits a photodiode installed in a fixed position, a counter start counting clocks with a period of 160 nsec.
When the laser beam pass again on the above photodiode the counter will stop.

Reading the counter value we note a difference from two reading. This difference is randomly variable from 3 periods till 40 periods that is not acceptable for our needs. We think this is due to the jerky motion of the stepping motor.

It is possible to obtain a uniform motion (difference within 5 clocks) from a stepping motor?

The platform is accurately balanced.

We thank everybody can give us some hints.

Vittorio Mattiolo

T

#### Tom Gianni

If your stepper is open-loop, you may need to go closed-loop stepper or brushless servo. Even with microstepping, a stepper is still somewhat jerky
because the control assumes the motor torque-angle curve is perfectly sinusoidal (no harmonics), which it is not. Also, make sure the step rate does not coincide with the any motor-load resonant frequencies, as this would excite the elec-mech system.

J

#### Johan Bengtsson

My calculations give that the error is between 0.5% and 6.4%.
By the way, where is the limit?

A stepping motor will normally not miss this at all, provided that you:
2. the input pulses driving the motor have the right frequency

I don't think it is the stepping that is the problem, because the steps occur at the same position every turn, and with microstepping this problem should be even smaller This could be a problem for you application, but for the
measurment those movement will not give that variation.

Even if if should be a problem it should be at most one step (ie at most 0.5%)

I think the problem is located in one of the above reasons, my guess is that your input frequency isn't stable enough.

By the way, have you checked the tolerance of your reference clock?

I would suggest the following:
do the same measurement again, but have two clocks connected
both clocks should be started and stopped in the same way as you describe, but the new clock should be counting input pulses to the stepping controller.

If the new clock don't vary (or at most vary by 0.5%) but the old one does then either your input frequency or your reference clock are not accurate enough.

If both clocks vary and the variations appear quite synconised then your input clock and your reference clock probably are good enough, and my guess is an overload problem (or some problem with the measuring method)

If both clocks vary and the variations are not following each other then you probably have more than one problem

Hope this gives you some ideas.

Johan Bengtsson
----------------------------------------
P&L, Innovation in training
Box 252, S-281 23 H{ssleholm SWEDEN
Tel: +46 451 49 460, Fax: +46 451 89 833
E-mail: [email protected]
Internet: http://www.pol.se/

B

#### Brandon Ellis

To avoid over speculation, I want to make sure I understand your testing procedure. Here are a few questions:

1)Is the clock (which is activated by the laser / photo-diode) running independently of the stepper indexer?
2)Are you doing your measurements after the stepper has reached it's speed, or is it accelerating?
3)What is the micro-stepping resolution setting of the drive?
4)What is the maximum step pulse frequency of the indexer?
5) what model of Toshiba drive are you using?

Just let me know so I can help,

Brandon S. Ellis
Sales Manager
Robotic Control Group
700 S. Illinois Ave.
Suite A104
Oak Ridge, TN 37830
Tel: (865) 425-0301, Ext. 160
Fax: (865) 425-0268
http://www.roboticcontrol.com

G

#### Guy H. Looney

My interpretation of your application (which may be wrong) is that you aren't microstepping. You said you have a motor w/200 steps. You also indicate that you have a microstepping controller by Toshiba. I would consider the controller to be the device that sends out the step & direction
signals. Is this correct? If so, then you haven't addressed what type of amplifier you have.

Neither the motor nor the controller determine the resolution of the system. They both contribute, but the amplifier is the ultimate factor in determining whether or not you are microstepping. Most stepper motors on the market today are made w/ the same concept: 2 phases & 50 "teeth".

There are exceptions to this of course (Oriental makes a 5 phase motor). A 2 phase motor with current either fully on or off allows for 4 states/rev:
(4 states/rev) x (50 teeth) = 200 steps/rev.

The amplifier determines how many steps/revolution there are. Full stepping amplifiers only allow for full current switching (thus my previous calculation). More advanced amplifiers allow for intermediate current switching via MOSFETs or other such devices. Here are common "titles" associated with drive resolutions:

Full stepping = 200 steps / rev
Half stepping = 400 steps / rev
Mini stepping = 1000 - 4000 steps / rev
Microstepping = 8000 steps / rev & higher

25,000 steps / rev is probably the most common resolution associated with microstepping, but I've seen drives with well over 250,000 steps / rev.

I'll ask the rest of the list serv in advance not to flood me with responses about mini stepping is actually this many steps/rev or microstepping is
actually this many steps/rev. The exact number of steps / rev associated with a particular "title" isn't the point of this email. The point is I'm trying to address whether or not it's a drive resolution problem.

If you'll notice in Johan's email he indicated that 1 step was an error of 0.5% (1/200). That calculation is based on 200 steps/rev which is not microstepping. Microstepping is definitely better than full stepping when smoothness is an issue. The more steps per rev you have, the less ripple (torque & velocity) you will see. I can assure you that smooth motion is not a problem for steppers....they're used in the semi-conductor industry all the time. There are not many applications around that require smoother motion.

My advice is to check you're drive resolution and see how many steps/rev you have. If it is only 200, try to increase your drive resolution.
Remember that your controller will have to put out more steps/second with higher resolutions, so make sure it has a frequency high enough to handle
higher resolutions.

Regards,
Guy
Guy H. Looney
Motion Control Engineer

A.C.E. Systems, LLC
170 Medearis Drive
Old Hickory, TN 37138
email: [email protected]
website: www.acesystemsllc.com
work: (615) 754-2378
fax: (425) 944-5017
cell: (615) 330-0044

B

#### Bill Sturm

There are many possible sources of error here. The photodiodes may not be perfectly spaced. You may be getting some electronic or software timing

If you want to focus on the stepper motor, then see if there is a spec for step to step accuracy. Look for a good quality motor with highly accurate step spacing.

Someone already mentioned the need for a stable clock source.

You should be able to get very good speed regulation with a stepper motor. I am not sure that you could do as well with a closed loop servo, unless it was very well tuned.

You might try a five phase stepper, if your budget will allow. They naturally have 1000 steps per revolution. Nyden, Berger-Lahr, and maybe Oriental motor sell these.

Bill Sturm

R

#### Rufus

> We need to turn a round platform ( diameter = mm.100, height = mm 50 > and weight = 300g) by using a stepping motor having 200 steps. We > need to have a constant speed of 600 turns/minute and a very uniform > motion. By my math, that's 600 * 200 / 60 = 2000 steps/second = 500 ns / step > We drive the motor by a microstepping controller by Toshiba. > > We measure the motion uniformity by means of a laser mounted on the > platform. When the laser beam hits a photodiode installed in a fixed > position, a counter start counting clocks with a period of 160 nsec. > When the laser beam pass again on the above photodiode the counter will > stop. At 10 turns per second or 100 Ms/turn you should get a count value of 100000/160 or 625 counts > Reading the counter value we note a difference from two reading. This > difference is randomly variable from 3 periods till 40 periods that is > not acceptable for our needs. We think this is due to the jerky motion > of the stepping motor. 3 periods * 160 ns = 480 ns (about 1 step) This one is understandable. Depending on when the laser signal is relative to the pulse train, you could have a variance of plus or minus a period. 40 periods * 160 ns = 6400 ns, almost 13 steps! or 13/200 * 360 = 23 degrees! I can't see how this is possible. Are you getting counts both above and below 625 by this error? i.e. count values of 585 to 665? > It is possible to obtain a uniform motion ( difference within 5 clocks) > from a stepping motor? If a stepper motor is good for nothing else, it is good for a repeatable motion of a constant load. There must be some flaw in the measurement, or perhaps you are losing pulses, but then you'd see your counts all above 625. If it was just jitter, you should be able to count 10 consecutive rotations and be pretty close to 6250 counts.

C

#### Curt Wilson

Rufus -- Sorry, you made a basic mistake in your units. > > We need to turn a round platform ( diameter = mm.100, height = mm 50 > > and weight = 300g) by using a stepping motor having 200 steps. We > > need to have a constant speed of 600 turns/minute and a very uniform > > motion. > > By my math, that's 600 * 200 / 60 = 2000 steps/second = 500 ns / step > Hold on here! This is 500 micrsoceconds per (full) step, not 500 nanoseconds! > > We drive the motor by a microstepping controller by Toshiba. > > > > We measure the motion uniformity by means of a laser mounted on the > > platform. When the laser beam hits a photodiode installed in a fixed > > position, a counter start counting clocks with a period of 160 nsec. > > When the laser beam pass again on the above photodiode the counter will > > stop. > > At 10 turns per second or 100 Ms/turn you should get a count value of 100000/160 or 625 counts > 100 milliseconds is 100,000,000 nanoseconds, so the value should be 100,000,000/160=625,000 > > Reading the counter value we note a difference from two reading. This > > difference is randomly variable from 3 periods till 40 periods that is > > not acceptable for our needs. We think this is due to the jerky motion > > of the stepping motor. > > 3 periods * 160 ns = 480 ns (about 1 step) This is about 1/1000 of a full step (but what is the microstepping resolution?). > This one is understandable. Depending on when the laser signal is relative to the pulse train, you could have a variance of plus or minus a period. > > 40 periods * 160 ns = 6400 ns, almost 13 steps! > or 13/200 * 360 = 23 degrees! > > I can't see how this is possible. It is really just about 1/80 of a full step, that's how. If he is using a 125 microstep-per-step (25,000 microstep-per-rev) drive, this is between 1 and 2 microsteps. > Curt Wilson Delta Tau Data Systems

V

#### Vittorio Mattiolo

Thanks for your reply. Sorry, but 2000 steps/sec correspond to a period of 500 microsec = 500.000 nsec. At 600 turns/minute, we have 625.000 counts and the counts between two steps are 625.000/200 = 3.125 steps. Thank you again Vittorio Mattiolo

S