In this last of the Encoder Woes series, (I can hear a collective sigh of relief), I would like to touch on the non-touchy subject of digital commutation. First, a basic explanation of analog commutation in a familiar "brushed" commutator type motor.
Take your hairdryer or cordless electric drill, well, don't take them anywhere, just think about them for a moment. The cordless electric drill is a fine example for our, ahem, study. I am sure you have seen the small sparks that are emitted from the rear end of the motor when the trigger is pulled and released. Those sparks are from the hard carbon brushes that contact the segmented copper commutator on the drill's motor rotor (spinning part).
These carbon brushes conduct DC voltage and current to the many winding coils on the rotor. When this happens, the rotor becomes magnetized. Since the non-spinning part of the motor (stator) has permanent magnets or wired electro-magnetic coils, it does not need a commutator to energize the "magnets". They are energized by direct wire connections or by their inherent magnetic structure.
When the DC voltage is transmitted to the different rotor windings as you pull the trigger, the magnetic coils quickly change polarity from positive to negative and back to positive again as it spins. The stator magnets stay magnetized in one orientation. This switching of polarity causes the rotor to be attracted to, and repelled from, the different stator magnets. Voila! rotation. The more you pull the trigger, the more voltage and current flow, making the rotor spin faster. When you release the trigger, the motor becomes a generator momentarily, and sends DC voltage and current back toward the battery. However, most modern cordless drills "shunt" this voltage to a power resistor. This acts as a brake, and dissipates the regenerated energy into the resistor and ultimately out, as heat.
O.K. so now, you know where those little sparks come from. (Carbon brushes rubbing quickly against a copper commutator.) Now, let's take your "brushless" AC Servo motor. You do have one? Of course you do.
The reason it is brushless is because it is an inside-out motor. The rotor in this case, needs no voltage and current supplied to it because it is made with permanent, (Rare Earth, Neodymium), magnets around its circumference. "So", says you, "how does it spin?" "Well," says I, that's where the encoder, (remember the encoder?) really shines. Unfortunately, more explanation becomes necessary. Sorry.
Not to confuse matters any more than they most likely are already, a Brushless AC Servo Motor is actually a Brushless DC Servo Motor, in that, the AC like sinusoidal wave forms sent to the motor are actually DC voltage that is made into sinusoidal waves by the Servo Amplifier. It is called PWM, or Pulse Width Modulation. It is made by the very fast switching of a three phase transistor (Darlington). Those three phases are connected to the three stator (outside case) windings of the motor.
So, think of those smooth flowing sine waves made by AC, and imagine a close up of the wave edge that looks like a set of rolling stairs going over a gentle hill and down into a deep valley and up and over another hill. From afar, it looks smooth, in fact, it is many on and off cycles of the Darlington transistor in increasing and decreasing magnitudes of voltage and frequency.
Now, how does the Servo amplifier know which of the three phase/s needs to be powered at any given time? Or, how does it commutate the voltage and current? Answer: The encoder. Quite simply, the exact position of the encoder relative to the stator, gives the Servo Amplifier its ability to "know" where the rotor is, and which stator phase or phases to give power to to hold position, rotate forward or reverse and what polarity to provide at any given microsecond.
This is why it is so critical not to disassemble the encoder from the motor. That is, unless you have some expensive equipment handy to realign it. The realignment process uses the encoder itself, the three phases of the stator, software, the amplifier, and possibly an oscilloscope to accomplish this exacting procedure. In other words, don't try this at home. Leave it to the professionals.
If your eyes are still not glazed over, congratulations! I think that is enough on this topic of encoders and their possible woes. Of course, my explanations are rather simple in nature, in hopes to make the ideas a bit more palatable. In fact, the processors and their algorithms necessary to accomplish the powerful, smooth running performance of a modern servo system are fairly complicated. Thankfully, most servo drive manufacturers have done the hard part for us all. They have, for the most part, made the systems "Plug-and-Play".
Until such time as I think of some other earth shatteringly important subject to wax poetic on...I remain your faithful servant. I think I'm tearing up...
Always, do the right thing...
Gaff
Friday, July 20, 2012
Tuesday, July 10, 2012
Encoder Woes Part II
My last post dealt with the possible failure of the servo system's encoder cable. Since it is most exposed to its environment, physical damage can happen quite easily, and will cause servo system encoder errors. The next most likely cause for Encoder Woes, is the encoder itself.
Many servo systems use optical encoders as the feedback device in their positioning and velocity closed loops. Other systems use Resolver based feedback. A resolver is a more simple device that uses sinusoidal wave forms to establish voltage references at certain positions of the resolver shaft. It is akin to a small motor, in that is has rotor and stator windings and their relative position to each other, as it spins, produces the sinusoidal wave. This "wave" is then use in an algorithm to output the feedback to servo control.
In this blog, we are concerned with an Encoder. Does it have a "pulse?" This device, the encoder, is a digital and mechanical optical feedback mechanism. Its primary component is a precision glass disk that is segmented about its outer edge with an opaque coating. This segmentation produces a number of "windows" that a light beam can shine through. The light beam is produced by an LED. This LED is mounted near the outside edge of the disk, and a receiver is mounted on the opposite side of the disk. As the disk turns, the light beam is alternately allowed to shine through the disk, and is then blocked by the opaque coating. This action produces "pulses" generated by the LED receiver module as the light energy from the LED is alternately "seen" and obscured. The number of "pulses" produced on each revolution of the encoder shaft is its resolution. There are multiple "tracks" of windows on most encoders. One such track is the index mark track. It produces only one pulse per revolution. However, the precise location of this pulse is used by the servo system to determine the encoder's (and Motor's rotor) position, relative to the motor.
Some optical encoders are low resolution, producing say, 360 pulses per revolution. Others are high resolution, these sometimes use special algorithms to produce over 1,000,000 pulses per revolution. A common resolution for servos is between 1000, and 10,000 pulses per revolution. Many servo systems have an additional algorithm that will multiply this resolution by four. This is a Quadrature circuit. With quadrature, a servo motor with 2,000 pulses per revolution (PPR), will deliver to the motion controller, 8,000 PPR. Increasing resolution, means increasing system accuracy.
Now that you have bit of general information on what an encoder does, lets concentrate on what happens when the encoder doesn't do its thing. Those encoder errors we spoke of in the last blog could be generated right at the encoder itself. A servo motor's encoder is attached directly to the end of the motor's rotor (the spinning part), and fixed to the motor's stator (motor frame).
Usually, the motor mounted encoder has some type of "industrial" connector that makes the connection to the encoder feedback cable. This "industrial" connector then connects somehow to the circuit board of the encoder inside the motor's encoder housing.
I hesitate to tell folks to open the back of a brushless servo motor. These precision motors and their encoders are assembled in "clean room" type environments and then sealed. In addition, special digital alignment tools are needed to "align" the encoder to the motor's stator (outer motor frame). If the encoder is loosened or moved in the slightest, the encoder must be realigned. Moving the encoder on a brushless servo motor, will produce system errors, and will result in your purchase of a new motor, or the expense of a servo motor repair.
Hopefully, that last paragraph instilled fear and dread in you. If you decide to open the back of a servo motor, to peek and poke around inside, don't say I didn't warn you. If you do decide to open it, your best tool will be your eyes. If at all possible, look but don't touch.
Remember, most encoder errors are due to faulty connections. Look carefully at the short interface between the "industrial" connector inside the motor and the encoder's connection header. Specifically, look at solder connections. More so, if the motor is attached to a machine with lots of vibration. That vibration can cause those solder connections to fatigue over time. The solder will not fail, it is usually the wire strands immediately adjacent to the solder that break. You might even get lucky and find the encoder's connection plug has worked its way loose. O.K., now you can use your big mitts to get in there and CAREFULLY reconnect the plug.
If all the wires prove firmly engaged and solidly connected, then perhaps the problem is with the encoder itself. Something inside the encoder has failed. Well, unless you have a very expensive alignment machine and motor specific software, you are faced with the fact that a new motor or motor repair is in your immediate future.
If you do have the afore mentioned tools and software, then why are you reading this? You probably know more about servo motors and encoders than I do. However, if you are still reading this, and would like to know why it's so darn important not to move an aligned encoder on a brushless servo motor...stay tuned for part three of Encoder Woes. Commutation explained. That's commutation, not communication.
Till next time...
Many servo systems use optical encoders as the feedback device in their positioning and velocity closed loops. Other systems use Resolver based feedback. A resolver is a more simple device that uses sinusoidal wave forms to establish voltage references at certain positions of the resolver shaft. It is akin to a small motor, in that is has rotor and stator windings and their relative position to each other, as it spins, produces the sinusoidal wave. This "wave" is then use in an algorithm to output the feedback to servo control.
In this blog, we are concerned with an Encoder. Does it have a "pulse?" This device, the encoder, is a digital and mechanical optical feedback mechanism. Its primary component is a precision glass disk that is segmented about its outer edge with an opaque coating. This segmentation produces a number of "windows" that a light beam can shine through. The light beam is produced by an LED. This LED is mounted near the outside edge of the disk, and a receiver is mounted on the opposite side of the disk. As the disk turns, the light beam is alternately allowed to shine through the disk, and is then blocked by the opaque coating. This action produces "pulses" generated by the LED receiver module as the light energy from the LED is alternately "seen" and obscured. The number of "pulses" produced on each revolution of the encoder shaft is its resolution. There are multiple "tracks" of windows on most encoders. One such track is the index mark track. It produces only one pulse per revolution. However, the precise location of this pulse is used by the servo system to determine the encoder's (and Motor's rotor) position, relative to the motor.
Some optical encoders are low resolution, producing say, 360 pulses per revolution. Others are high resolution, these sometimes use special algorithms to produce over 1,000,000 pulses per revolution. A common resolution for servos is between 1000, and 10,000 pulses per revolution. Many servo systems have an additional algorithm that will multiply this resolution by four. This is a Quadrature circuit. With quadrature, a servo motor with 2,000 pulses per revolution (PPR), will deliver to the motion controller, 8,000 PPR. Increasing resolution, means increasing system accuracy.
Now that you have bit of general information on what an encoder does, lets concentrate on what happens when the encoder doesn't do its thing. Those encoder errors we spoke of in the last blog could be generated right at the encoder itself. A servo motor's encoder is attached directly to the end of the motor's rotor (the spinning part), and fixed to the motor's stator (motor frame).
Usually, the motor mounted encoder has some type of "industrial" connector that makes the connection to the encoder feedback cable. This "industrial" connector then connects somehow to the circuit board of the encoder inside the motor's encoder housing.
I hesitate to tell folks to open the back of a brushless servo motor. These precision motors and their encoders are assembled in "clean room" type environments and then sealed. In addition, special digital alignment tools are needed to "align" the encoder to the motor's stator (outer motor frame). If the encoder is loosened or moved in the slightest, the encoder must be realigned. Moving the encoder on a brushless servo motor, will produce system errors, and will result in your purchase of a new motor, or the expense of a servo motor repair.
Hopefully, that last paragraph instilled fear and dread in you. If you decide to open the back of a servo motor, to peek and poke around inside, don't say I didn't warn you. If you do decide to open it, your best tool will be your eyes. If at all possible, look but don't touch.
Remember, most encoder errors are due to faulty connections. Look carefully at the short interface between the "industrial" connector inside the motor and the encoder's connection header. Specifically, look at solder connections. More so, if the motor is attached to a machine with lots of vibration. That vibration can cause those solder connections to fatigue over time. The solder will not fail, it is usually the wire strands immediately adjacent to the solder that break. You might even get lucky and find the encoder's connection plug has worked its way loose. O.K., now you can use your big mitts to get in there and CAREFULLY reconnect the plug.
If all the wires prove firmly engaged and solidly connected, then perhaps the problem is with the encoder itself. Something inside the encoder has failed. Well, unless you have a very expensive alignment machine and motor specific software, you are faced with the fact that a new motor or motor repair is in your immediate future.
If you do have the afore mentioned tools and software, then why are you reading this? You probably know more about servo motors and encoders than I do. However, if you are still reading this, and would like to know why it's so darn important not to move an aligned encoder on a brushless servo motor...stay tuned for part three of Encoder Woes. Commutation explained. That's commutation, not communication.
Till next time...
Tuesday, June 19, 2012
Encoder Errors
It's those A-Channel Lost, Hall Effect Errors, Excess Follower Errors, or Over Speed Errors that just make you crazy. Right? "Why me?" says you. Well, it's not you. It's most likely the signal transmission medium in the servo system. The encoder cable is the single most important, exposed component, in a servo's motion control system. Generally, it is sitting out there for the world to see, and gets stepped on, spilled upon, continuously bent, and thoroughly mistreated. Yet, the high frequency signals that course through it's many conductors are vitally important to the servo motion system's "Closed Loop" digital position and velocity information ("feedback"). With even one small conductor slightly damaged, the system is likely to generate a feedback error.
So, you've got an encoder error of some kind. Then, "Clean up your act!"
Try cleaning the pins and sockets on the cable and the connector of the motor. Spray a bit of contact cleaner onto and into the pins and sockets respectively, then, gently blow them dry with clean compressed air (wear safety glasses). Now, reconnect the cable, and try the system again. If you are fortunate, the servo control will be a "happy camper" again.
No Go on the clean scheme? Get out your Volt/Ohm/Multimeter...set it to stun...
If the cleaning idea is a bust, the problem may be an obvious or hidden break in the wires. Check the encoder cable for obvious damage such as cuts or crushing evidence. If all looks kosher, then the next step is to actually verify the conductivity of each conductor end to end, and the lack of conductivity of each conductor to the cable's shielding, back-shell and adjacent conductors*. You will need a pin-out drawing of your cable to fully ascertain the integrity of the conductors and terminations. (*Some cables have jumpers soldered between pins or sockets inside the military type connectors.) All servo manufacturers do things differently when it comes to Servo Amplifier to Motor connection. Hopefully, you have access to the specific component schematics, or pin-outs. If not, the World Wide Web is great resource. Start typing your search query.
In some cases, despite the seemingly passing grade of the cable after testing, it could still be the source of system errors. Only a known-good replacement cable can verify that. If, after testing and or replacement, the error/s are still hanging around, you will be staring the servo motor directly in the face. It, the servo motor, can cause those feedback errors too. More specifically, it would be the motor's integral encoder that could be the problem child.
That, is for the next entry...
So, you've got an encoder error of some kind. Then, "Clean up your act!"
Try cleaning the pins and sockets on the cable and the connector of the motor. Spray a bit of contact cleaner onto and into the pins and sockets respectively, then, gently blow them dry with clean compressed air (wear safety glasses). Now, reconnect the cable, and try the system again. If you are fortunate, the servo control will be a "happy camper" again.
No Go on the clean scheme? Get out your Volt/Ohm/Multimeter...set it to stun...
If the cleaning idea is a bust, the problem may be an obvious or hidden break in the wires. Check the encoder cable for obvious damage such as cuts or crushing evidence. If all looks kosher, then the next step is to actually verify the conductivity of each conductor end to end, and the lack of conductivity of each conductor to the cable's shielding, back-shell and adjacent conductors*. You will need a pin-out drawing of your cable to fully ascertain the integrity of the conductors and terminations. (*Some cables have jumpers soldered between pins or sockets inside the military type connectors.) All servo manufacturers do things differently when it comes to Servo Amplifier to Motor connection. Hopefully, you have access to the specific component schematics, or pin-outs. If not, the World Wide Web is great resource. Start typing your search query.
In some cases, despite the seemingly passing grade of the cable after testing, it could still be the source of system errors. Only a known-good replacement cable can verify that. If, after testing and or replacement, the error/s are still hanging around, you will be staring the servo motor directly in the face. It, the servo motor, can cause those feedback errors too. More specifically, it would be the motor's integral encoder that could be the problem child.
That, is for the next entry...
Subscribe to:
Posts (Atom)