Machsupport Forum
Mach Discussion => General Mach Discussion => Topic started by: Dan13 on May 03, 2010, 08:05:24 AM
-
Hi,
Does anybody know what is a Sine/Cosine Encoder? My Ultra5000 AC servo drive manual says that it supports both TTL encoders and Sine/Cosine encoders. Do they mean by Sine/Cosine encoder that it supports a resolver feedback?
Thanks,
Dan
-
Nope, its the intelligent encoders that are fitted to some of the Allen Bradley motors. I have them on the MPL motors I use on the spindle of the mill and Lathe.
Do a search for Sick Stegmann Hiperface.
Hood
-
I see, Hood. So no way it would work with a resolver if I connected one? Sorry for the ignorance, but they both seem to output Sine/Cosine wave signal. But I am not sure about the commutation...
Is that Hiperface a kind of motor feedback device? Not really clear from what I've found here http://www.stegmann.com/product/servo/index.html Are you suggesting me to replace the resolver on the motor with one of these?
Thanks,
Dan
-
Are you suggesting me to replace the resolver on the motor with one of these?
Thanks,
Dan
Nope, simply saying to you that the Sine Cosine inputs are for a Sine/Cosine encoder. You can try and hook up the resolver if you wish, if it works I would be very interested but from reading the manuals for the Ultra 3000 drives and the UltraWare software it seems to me it is not for a resolver.
Hood
-
Yes, the manuals don't mention anything about resolvers. I will try and let you know. It will either work or I absolutely misunderstand the operation of these to feedback devices.
The motor doesn't have any commutation signals, so have no idea how is the commutation done with this motor...? Even with a resolver it must have some hall signals to commutate, doesn't it?
Dan
-
The commutation with the intelligent encoders is via a serial signal (Data + and Data -) not a clue about the resolvers I am afraid.
Hood
-
Daniel
The very last line on this page would seem to confirm what I thought.
http://rockwellautomation-test.custhelp.com/ci/fattach/get/27879/1247153771
Hood
-
Thanks Hood. Bad news then :(
But how do you always manage to find this documents... I thought I've searched through their whole database and haven't found any clue about resolvers.
I think I will have to find an encoder then to replace the resolver. It's not going to be easy - the motor is 13,000RPM and I haven't seen encoders that work at these speeds.
Daniel
-
The max operating speed of the SRM intelligent encoders on the MPL motors is 12,000rpm so even that falls short.
Not sure if one of the resolver to encoder boards would handle that kind of speed and then would the drive handle it is another problem.
Hood
-
Hood,
I will probably won't need speeds above 10,000rpm (or may be even 8,000rpm), just wanted to do it the right way and put an encoder that can handle the whole range in case that at some point I will need it. The closest I've found so far were 8,000RPM so the 12,000RPM is sure better!
So you're saying that any motor of the MPL series has an encoder rated to 12,000rpm? Think the MPL are rated up to 5000RPM, so they used quite nice encoders on them. Do you know what make are the encoders? Model?
As far as I understand the motor poles number should have no affect on selecting a commutating encoder, but for some reason Q-Phase encoders do specify the pole count of the motor on their encoders...?
Daniel
-
Some of the MPL motors are rated speed of 8,000 rpm
The problem with the intelligent encoders that are on the MPL motors (or some, as some have normal encoders) is that they communicate with the drives and contain the info about the motor they are fitted to. In other words you would need to use them on the motor they came off I think.
They are Sick Stegmann SRM encoders or at least some are because I think there are also SRS ones on some motors.
Hood
-
So it makes things even more complicated :(
Any idea about the possible relationship between the motor poles number and encoder commutation? As far as I know you simply put the number of poles in the drive and it makes the calculations of the waves based on it. Have no clue what does it have to do with the encoder itself?
Daniel
-
Are you referring to the intelligent encoders? If you are then I presume that is where the programmaed part of the encoder comes in as it will contain the info about the poles. Never done much reading up on it though so its just an assumption.
One of the MPL motors I bought to get the normal encoder off had a very tidy encoder on it that looked like it would possibly be capable of speed but afraid I cant even remember what make/model it was so cant look up the specs. Normally the AB motors had Renco encoders on them but looked at their site and dont see any max speed info there :(
Hood
-
I was referring to these:
http://www.quantumdev.com/products/optical_encoders/qd145.html
See that they differentiate the models by the number of poles. Max speed for these is only 8,000RPM though :(
Daniel
-
Not sure I understand your question, these seem to be normal incremental encoders and as such you need the commutation channels on them to match your motors pole pairs. Or are you meaning something else?
Found some info on encoders saying 30,000rpm but not sure if they are any use or not as have not really read it yet. http://www.moog.com/literature/MCG/opticalencoders.pdf
Oh and if this is for a spindle you may not even need commutation signals, well not sure about the Ultra 5000's but the 3000's can do an auto sense for commutation startup which on a spindle I dont think would be an issue.
Hood
-
Basically a resolver produces a sin/cos signal, they require analogue detection circuitry to work, I have some Allen Bradly drives that are resolver only, the encoder signal is extrapolated from this in the drive for output to a controller, this method has been done for some time now for AC sinusoidal commutation, this is opposed to BLDC commutation which typically uses three commutation tracks on the encoder disk, used to be known as Hall effect due to 3 descrete hall effect devices used.
There are current sinusoidal drives that use an initial marker signal and then produce the commutation from the encoder pulses.
BTW, Since Allen Bradley bought out Electro-Craft, their motor and drives have principally been Electro-Craft, often the same part number.
If you replace an encoder on a motor to use as AC or BLDC, the pole count and pole position has to be known or detected and the encoder lined up with the pole position.
Someone mentioned Renco, they have now been taken over by Heidenhain and their product line has been trimmed quite a bit.
Nosmo.
-
Not sure I understand your question, these seem to be normal incremental encoders and as such you need the commutation channels on them to match your motors pole pairs. Or are you meaning something else?
That's what I had thought - you put a commutating encoder on a motor and its 3 hall signals have to match the motor poles. I hadn't realized that the NUMBER of poles had any affect - still same three hall signals from the encoder. But from that link it seems that an encoder is specific to a motor with the NUMBER of poles it was designed for. Am I missing something here? Are the signals from an encoder on a 4 pole motor different from those of an encoder on a 6 pole motor?
Found some info on encoders saying 30,000rpm but not sure if they are any use or not as have not really read it yet. http://www.moog.com/literature/MCG/opticalencoders.pdf
Yes, these look good. They are not commutating though.
Oh and if this is for a spindle you may not even need commutation signals, well not sure about the Ultra 5000's but the 3000's can do an auto sense for commutation startup which on a spindle I dont think would be an issue.
Hood
Ah... Hadn't noticed this detail in the manual before. This is perfect! The motor IS meant for a spindle indeed. Thanks for pointing me to this. So I wonder now what if I fed the Sin/Cos channels from the resolver to the the Ultra5000 while the resolver Ref is being excited by a DC voltage.. ??? Does it make any sense? In second thought, seems it wouldn't work at low speeds because the resolver wouldn't be able to induce the Sin/Cos when the magnetic field changes slowly. But with a high frequency excitation I think the drive "will be confused"... Can you tell me how does the Sin/Cos encoder work? Is it the same phasing like normal, square wave signal, encoder?
Daniel
-
Thanks for the info Nosmo. Does it make the AB drives BLDC commutation?? They do expect hall signals (the Ultra series).
Daniel
-
Are the signals from an encoder on a 4 pole motor different from those of an encoder on a 6 pole motor?
Never looked deeply into it but I would assume that it is the spacing of the three signals that would let your drive know the commutation startup. So if it was two pole the signals would be 180 out, 3 pole 120 out, 4 pole 90.
Yes, these look good. They are not commutating though.
If using self sensing commutation wont be required.
Ah... Hadn't noticed this detail in the manual before. This is perfect! The motor IS meant for a spindle indeed. Thanks for pointing me to this. So I wonder now what if I fed the Sin/Cos channels from the resolver to the the Ultra5000 while the resolver Ref is being excited by a DC voltage.. ??? Does it make any sense? In second thought, seems it wouldn't work at low speeds because the resolver wouldn't be able to induce the Sin/Cos when the magnetic field changes slowly. But with a high frequency excitation I think the drive "will be confused"... Can you tell me how does the Sin/Cos encoder work? Is it the same phasing like normal, square wave signal, encoder?
Daniel
Still dont think the resolver will work or I would imagine AB would mention it, then again only one way to find out ;)
Thanks for the info Nosmo. Does it make the AB drives BLDC commutation?? They do expect hall signals (the Ultra series).
Daniel
As far as I am aware BLDC and AC servo motors are one in the same but it is the drives that are different in the way they do the commutation. The Ultra drives are AC.
Hood
-
Never looked deeply into it but I would assume that it is the spacing of the three signals that would let your drive know the commutation startup. So if it was two pole the signals would be 180 out, 3 pole 120 out, 4 pole 90.
I am afraid this is not correct as there are always 3 signals and they are equally spread along the 360 degrees of one revolution, meaning there would always be a 120 degrees phase shift between them. Also, brushless motors are essentially 3-phase motors so the control signal is always 3 phase 120 degrees apart and the commutation signal must match. You can also see this in the Ultra manuals where they specify the signals' phasing - nothing mentioned about the motor poles number. As I see it, the number of poles only dictates the control signal frequency to achieve a given speed and the commutation essentially follows that frequency.
Still dont think the resolver will work or I would imagine AB would mention it, then again only one way to find out ;)
Yep... will do some reading on Sin/Cos encoders and try to understand their operation. Hopefully the Sin/Cos signal of these encoders and those of a resolver are the same and AB say their Ultra drives don't support resolvers merely because their drives don't have an AC excitation signal required for resolvers. In this case an external signal might be used I assume.
As far as I am aware BLDC and AC servo motors are one in the same but it is the drives that are different in the way they do the commutation. The Ultra drives are AC.
Hood
Yes, I also thought the Ultra drives were AC, but if I understand what Nosmo says, then since they are using hall feedback they are BLDC?? :-\
Daniel
-
I am afraid this is not correct as there are always 3 signals and they are equally spread along the 360 degrees of one revolution, meaning there would always be a 120 degrees phase shift between them.
Yes thinking about it thats correct, will have to look into it when I get a chance.
The commutation signals are only used at startup as far as I am aware so once the motor is moving they are not used, some of the S type motors (similar to H) had hall signals rather than encoders with commutation tracks, again will have to read up on that.
I remember reading an article a few years back which showed the difference between BLDC and AC drives and it had some nice animated graphics showing how the poles were energised in either case, will have to see if I can find it again.
Going back to your motor, IF the Ultra 5000's support Self Sensing as the Ultra 3000's do, then even one of the cheap encoders from US Digital should work for you. You probably would need to make a line driver as well, but no big deal in doing that :)
Hood
-
Have a look at this http://www.renco.com/uploads/media/106803_113.pdf
There are 4 tracks for commutation on that disc but whether its 8 pole or 4 pole I am not sure, back to work now and tonight I will try and do some reading, unless someone chimes in and saves me the bother ;)
Hood
-
Ha.. ha.. This is the exact same paper you once showed me ;) I don't know, still think the number of poles should make no difference.
Yes, as I've mentioned in the previous post the Ultra 5000 does support self sensing as per the manual, so using a US digital encoder sounds like a good solution. But what do you mean by a line driver?
Daniel
-
Ha.. ha.. This is the exact same paper you once showed me ;) I don't know, still think the number of poles should make no difference.
I can assure you it does,a few years ago I tried to fit an encoder from a H series motor (6 pole) To a 1326 motor(4 pole) and it didnt work very welll ;D
Yes, as I've mentioned in the previous post the Ultra 5000 does support self sensing as per the manual, so using a US digital encoder sounds like a good solution. But what do you mean by a line driver?
Daniel
Line driver is to turn the single ended outputs of the US Digital encoder into Differential outputs. In fact I think US Digital sell them for under $10 so probably just as well buying from them. I made up my own for the Step/Dir outputs as I wanted it to plug directly into the BOB eliminating wires.
Hood
-
Ah... I see. Saw, they also have encoders with differential outputs already. But the speed seems to be the problem with these. Also, most of their encoders don't have an Index channel.
Daniel
-
Ah yes, forgot about the turbo speed :( Just have to keep an eye out on eBay and check the specs if anything comes up.
Hood
-
Here are some good resources I've found on feedback devices:
1. This one explains about Sin/Cos encoders operation:
http://www.sensorsmag.com/sensors/position-presence-proximity/understanding-integrated-hall-effect-rotary-encoders-1254?page_id=1
Apparently this type of encoders is not optical but they rather use hall sensors. As can be seen the resulting Sin/Cos signals have 90 degrees phase shift and their magnitude is not dependant on the shaft angle. This is why a servo drive that can use Sin/Cos encoder, can't use a resolver instead.
2. A very good document explaining resolvers operation:
http://www.admotec.com/TT02.pdf
3. Another one on resolvers:
http://www.amci.com/tutorials/tutorials-what-is-resolver.asp
4. This one covers several feedback devices and has some good information about resolvers:
http://www.motor-solutions.com/datasheets/servo_motor_feedback_devices.pdf
5. Another one on various feedback devices:
http://www.allegromicro.com/en/Products/Design/compumot/a39a44.pdf
So looks like I'll have to replace the resolver with an encoder if I want to use the motor with the Ultra 5000 drive.
Daniel
-
Ah yes, forgot about the turbo speed :( Just have to keep an eye out on eBay and check the specs if anything comes up.
Hood
Yep... guess that is what I'll have to do. The speed seems to be the biggest problem.
Daniel
-
The typical resolver has 3 windings, one is for exciting the rotor and the other two detect the sin/cos signal, the biggest problem with mixing drive is that the excitation frequency and magnitude is different between makes.
The output is two sine waves 90deg apart.
Older types had a permanent magnet rotor.
The Renco commutation encoders have 3 tracks and have to be matched to the pole count of the motor, Renco end theirs with a /x the x being the pole count in pairs, IOW /4 would be for a 8 pole motor.
When fitting one of these encoders the proper track has to be aligned with the appropriate pole, this is usually done with a double beam scope.
Here is a animated link that explains BLDC.
http://users.tinyworld.co.uk/flecc/4-pole-bldc-motor031102.swf
Nosmo.
-
Thanks Nosmo. Still can't see what would be the difference between a commutating encoder designed for an 8 pole motor and a one designed for a 6 pole motor, for instance?
Daniel
-
If you mean the hall equivalent tracks, then for an 8 pole motor there is 4 electrical cycles for every mechanical cycle, for a 6 pole it would be 3 for every mechanical.
IOW, The 8 pole repeats 4 times/rev.
Nosmo.
-
OK, I think I understand it now :)
Thanks,
Daniel
-
Lots of reading to do there Daniel ;D
Nosmoking
Thanks for the info, the animation I have seen before but the one I was thinking of was different so will have a search as it made it even clearer for people like me ;D
Hood
-
Yes, Hood, a lot of info there, but it worth the time to read :)
Daniel
-
Daniel
Yes found a lot of other good info last night myself,especially one that tells you the way you drive the brushless servo determines whether its AC or DC. Basically DC drives two coils at one point in time where AC drives all three. Said the increase in torque from the same motor was 28% more by driving with AC.
Also did a bit of searching for encoders and came across this
http://cgi.ebay.co.uk/ws/eBayISAPI.dll?ViewItem&item=230450656966&ssPageName=STRK:MEWAX:IT
Not able to find the exact specs for it and not really sure what the guy means by single channel, I am presuming he means single ended as I dont think you can get such a thing as a single channel encoder. As said couldnt find that exact model but the series seems to be capable of 12,000rpm continuous speed. Might be worth an email to see if he has any more info.
Hood
-
Hood,
Thanks for link. Only information I've been able to find on it is this
http://sensor.baumerelectric.com/productnavigator/downloads/Produkte/PDF/Datenblatt/Winkel_und_Positionsmesssysteme/en_E_1DG_BDK.pdf
You've probably seen it already. Looks like the same encoder just a newer version. So looks like it does rated at 12,000rpm, but it is only 100PPR :( so I am not sure how will the motor perform at low speeds. Also it is 10-30VDC so would require some interface.
Daniel
-
The guy says its 5v but as you say the numbers seem to suggest 24v, would presume the docs are correct.
100 ppr isnt going to be great, thats for sure, especially if you are intending to turn at low speeds but at least you know that Baumer seem to rate their encoders for high speeds so might be worth doing a search on US eBay.
Hood