PDA

View Full Version : Closed loop Step Servo Drive Driver 2HSS86H + 4.5N.M Motor + Encoder



george uk
21-01-2014, 12:09 AM
anyone used or have any opinion on these type, hybrid steppers with encoders

CNC Hybrid Closed loop Step Servo Drive Driver 2HSS86H + 4.5N.M Motor + Encoder | eBay (http://www.ebay.co.uk/itm/281237656097?ssPageName=STRK:MEWAX:IT&_trksid=p3984.m1423.l2649)

JAZZCNC
21-01-2014, 12:22 AM
Yep me but the smaller 2Nm they are brillant. See my video here. Leadshine Closed loop stepper - YouTube (http://www.youtube.com/watch?v=ZMZdCcLQc4M)

george uk
21-01-2014, 01:42 AM
They make so much sense, nice vid, it gives you a good idea of what they are capable of

Boyan Silyavski
21-01-2014, 02:09 AM
They are nice but the price+customs+PSU= the price of a good second hand servo. And they are steppers with their good and bad. I was in doubt for 2 weeks and at the end i bought AC servo motors +drives. Which i suppose for the same money will bring the machine to another level.

What i mean is though they seem nice, it seems better to me to go with normal steppers or go to servos.

JAZZCNC
21-01-2014, 02:44 AM
Which i suppose for the same money will bring the machine to another level.

Yes I agree.!!. . . . .It will take it to Another level of head ache.!


What i mean is though they seem nice, it seems better to me to go with normal steppers or go to servos.

Normal steppers don't even come close to these even with good drives and servo's offer nothing more other than speed and torque unless used with control that can close the loop correctly, infact they offer less if the loop isn't closed because they can only fault if a following error occurs not correct for positional loss.
If these drives/motors are sized correctly for the machine then they match servos for accurecy and absolutley blow them away for ease of use. No messing around with scopes and PID settings etc, just plug-in setup a few simple parameters and forget.

Servo's are great but they can also be a complete night mare and for DIY machine they can be more trouble than there worth with very little gain worth the trouble or cost.!!

george uk
21-01-2014, 11:03 AM
If these drives/motors are sized correctly for the machine then they match servos for accurecy and absolutley blow them away for ease of use.

Absolutly 100% agree there Jazz, that type of upgrade to steppers and drivers is a game changer.

Twinning an X, Accurate Z position over large time running jobs, 4th and 5th Axis positional holding. any CNCing that your likley to loose steps.

Jonathan
21-01-2014, 01:32 PM
Twinning an X, Accurate Z position over large time running jobs

If those are causing a problem, then it should be solved by setting up the system properly.


4th and 5th Axis positional holding. any CNCing that your likley to loose steps.

If you're likely to loose steps then you should choose a motor with a higher torque rating, since by the time the driver has corrected for any significant loss of position it could be too late.

Bear in mind that to make a stepper motor system closed loop, above a certain speed you don't need the encoder as the driver can monitor the motor back-emf and accurately infer the rotor position from that. For this reason, if you were to take the stepper motor from this driver and connect it to a driver which uses field oriented control (vector control), (e.g. 2M880N) the performance should be virtually identical above that speed. You can only gain anything below this speed, since whether the position is inferred from an encoder, or the back-emf, doesn't change the rest of the algorithm. That's the reason why the stall detect feature, e.g. on the AM882 driver, only works above 300rpm - it can't sense the position below that speed.
There will be slight differences, beyond the scope of this post, however these will be negligible compared to the gain you get from having a much wider selection of motor ratings with standard stepper systems - e.g. you'll notice the difference buying a 3Nm stepper instead of the 2Nm closed loop one as the closed loop one in general can't gain anything like 50% more torque from the motor.

There is potential for a gain in accuracy due to having the '1000 line' encoder on the stepper motor, however I'd argue that there are few DIY CNC systems which would benefit from correcting the position error that occurs when an torque is applied to an open-loop stepper motor, since it is so small compared to other inaccuracies in the system. Similarly, with an encoder the motor rotation can be 'smoother' at low speeds. Unless you're having big problems with resonance, this is unlikely to be an issue if the system is correctly designed.


Yes I agree.!!. . . . .It will take it to Another level of head ache.!
Normal steppers don't even come close to these even with good drives and servo's offer nothing more other than speed and torque unless used with control that can close the loop correctly, infact they offer less if the loop isn't closed because they can only fault if a following error occurs not correct for positional loss.

Servo motors are closed loop by definition. Whether you've got a closed loop driver on a stepper motor, or a more conventional motor, doesn't change the fact that the driver is continuously monitoring the following error. Following error and position loss are effectively the same thing - the closed loop control corrects for this error with both systems. With either system you should be able to select what happens when this error exceeds a certain threshold - i.e. loss of position.

Fundamentally, the only difference between this closed loop servo system and a standard servo motor is the number of poles and phases. Stepper motors generally have 50 poles and 2-phases, compared to around 2-6 poles and generally 3 phases for servo motors. That implies lower torque ripple, so smoother speed control, however any modern servo drive will use the field oriented control algorithm, which (if the bandwidth is sufficient) largely eliminates the torque ripple.

I certainly wont be buying any closed-loop stepper systems.

george uk
21-01-2014, 09:04 PM
hi Jonathan.


http://www.mycncuk.com/forums/images/misc/quote_icon.png Originally Posted by george uk http://www.mycncuk.com/forums/images/buttons/viewpost-right.png (http://www.mycncuk.com/forums/motor-drivers-controllers/7047-closed-loop-step-servo-drive-driver-2hss86h-4-5n-m-motor-encoder.html#post53729)Twinning an X, Accurate Z position over large time running jobs



If those are causing a problem, then it should be solved by setting up the system properly.

understood, i have been reading up on what experiences people are haveing using twin X, and i notice that a few were having problems homing them both, i just thought that these may be an easy way round that, but, yes defiantly, if people are lossing steps on X, its machine problems or bad design .

Z Axis, for what i want to do, thats more important, some of the jobs we are looking at have run times going over 5 hours plus, ( am doing 2 machines ), so it seems to be a reliable way of making sure any lost steps in Z dont compound themselves, as they would be corrected immediately.


You can only gain anything below this speed, since whether the position is inferred from an encoder, or the back-emf, doesn't change the rest of the algorithm. That's the reason why the stall detect feature, e.g. on the AM882 driver, only works above 300rpm - it can't sense the position below that speed.


Thankyou for this, i like information put this way, short sweet and direct,

Boyan Silyavski
21-01-2014, 09:16 PM
What i meant was that closed loop steppers are still steppers, low torque at high revolutions, which in my case meant i should have compromised with the gantry weight on the 2300x1300 machine i started building or buy a really big closed loop steppers, which in price surpassed second hand AC servo system. The other solution was to make all the gantry from aluminum, but as DIYers like me normally don't have access to machine 1 1/2 meters long pieces of aluminum, this would have raised again considerably the price.

I don't know if i am to obtain another level of headache. But maybe its the time to learn. Reading from the manual of the Samsung servos i have bought, the tuning seems quite simple in fact. If i drive them from the motion control in step mode. According to the manual you introduce some simple parameters, home your machine, push the button and the servos start some test moves and adjust to the machine itself. If that works, it certainly is a non brainer. If doesn't work, there is always something new to learn on the way. Plus i can control them in a couple of different ways also from the motion control board if desired so.

In short i prefer to be free with the design of the machine to be able to construct it the way i like, with the budget i like, to be heavy as necessary, and to drive it as its necessary :-) . As a DIYer i wanted a commercial machine for the money of DIY. That meant some second hand components.

JAZZCNC
21-01-2014, 11:11 PM
Servo motors are closed loop by definition. Whether you've got a closed loop driver on a stepper motor, or a more conventional motor, doesn't change the fact that the driver is continuously monitoring the following error. Following error and position loss are effectively the same thing - the closed loop control corrects for this error with both systems. With either system you should be able to select what happens when this error exceeds a certain threshold - i.e. loss of position.

So then by your definition just sticking a Servo motor and drive on a machine using Mach3 makes it closed loop.!!. . . . NOPE. . .Why not.? . . . .Because Closing the loop is the function of the Motion Controller not the Servo drive or Motor.

Again the Servo can only watch the following error report back to the motion controller and fault if goes past it's parameters. Positional correction is down to the Controller and if like Mach3 it doesn't have the abiilty to close the loop then position is not corrected.
The Closed loop Stepper system has built in controller to do the correction so it doesn't matter if the Controller sending the pulses doesn't close the loop position is still corrected AND just like a TRUE CLOSED LOOP servo system has the abilty to fault if outside set parameters.

Servo's have there place and like in Silyavski case with a very heavy gantry then I agree they are probably better suited.. . . .BUT . . My original case that for general DIY machines having or needing the luxury of CLOSED LOOP capabilty's still applies and they are much easier and cheaper than servo's plus can be just has accurate if correct size for the machine.

Now are they better than standard Stepper system then YES but only because they Close the Loop at the drive and in the case of the smaller 2Nm which use 3 phase motors with 1.2deg step angle makes for a very smooth running motor. They are no more powerful, faster or anything else and if you don't want the Luxury of Closed loop then with Very good drives then agree they are an expensive Luxury that standard steppers can match other than closing the loop and correcting positional loss.

For DIY slaved axis Stepper systems then IMO they are the DOGs Danglies offering the best all round solution with Minimal hassle of setup and high reliabilty against racking or positional loss.!!. . . . It just comes with a price tag.!!

Would I use them.? . . . . Yes but only on high end machine with Slaved Axis.!

Jonathan
22-01-2014, 12:58 AM
I think you're mixing up two different control loops. Any servo motor drive with position control uses feedback to maintain the position error below a specified limit. The closed loop stepper system is just another type of servo system, hence both 'standard' servo motors and 'closed loop stepper' systems correct for position error. How well they can correct is dependent on the motor ratings, but with either system if you exceed the motor torque rating the result is the same - loss of position (i.e. position error outside limit for some time).

Within the motor driver, the position command (step&dir) is converted to a speed reference using a PID (+possible other bits) controller, then into a torque reference using a PID (+other bits) controller, which is the input to the vector current control algorithm that uses a further two PI controllers to set the current within the motor. The speed measured from the encoder, or back-emf sensing, is fed into the PI controller and used to correct the position error. So you actually have several control loops inside one another. Sometimes people implement the position control loop inside the motion controller, and instead send speed&direction commands to the motor drive, but doing that is a separate can of worms. It is these PID controllers that are tuned when a servo motor is set up - the difference for the stepper system is they only come with a limited set of motors, which combined with the different characteristics of stepper motors means they get away without the user tuning the system for 'most applications'.

Again, this control system is essentially the same (except for the 'other bits') for all the types of motors commonly used for servos, including steppers, and can be implemented without an encoder so long as the motor speed is sufficient.


just sticking a Servo motor and drive on a machine using Mach3 makes it closed loop.!!. . . . NOPE. . .Why not.? . . . .Because Closing the loop is the function of the Motion Controller not the Servo drive or Motor.

The servo motor+drive, or closed loop stepper+drive are themselves closed loop, but the system as a whole arguably isn't as Mach3 doesn't know what the following error is. When the fault pin from the driver is connected to an input to Mach3, then Mach3 knows when the error exceeded the set threshold (i.e. the motor was operated outside its ratings for too long) and can act accordingly. You can set it up this way for servos, stepper based servos and stepper drivers with stall-detection.
Whether you close the position control loop within the motion controller (e.g. using LinuxCNC) or the driver is a separate debate to the one we're having now about the virtues of steppers and servo motors. I don't think that's worth discussing here, given that due to bandwidth limitations it's not viable using a parallel port which is probably the system most people reading this will use.

I'm not saying there are no advantages to using the closed-loop stepper system, I'm just confident that they're not tangible for the vast majority of machines made on this forum and in the case where they are, a standard servo motor system is likely offer the same performance at a lower cost.

JAZZCNC
22-01-2014, 02:35 AM
The servo motor+drive, or closed loop stepper+drive are themselves closed loop, but the system as a whole arguably isn't as Mach3 doesn't know what the following error is. When the fault pin from the driver is connected to an input to Mach3, then Mach3 knows when the error exceeded the set threshold (i.e. the motor was operated outside its ratings for too long) and can act accordingly. You can set it up this way for servos, stepper based servos and stepper drivers with stall-detection.


Yes I know all this and fully agree, I also admit I may have miss read what you meant but my initial point was that Servo's offer nothing more other than linear torque than these Closed loop Stepper systems do unless you close the Loop properly at the controller which like you say is another can of worms we both don't want to open.

Whether or not you can find servo's cheaper I think is a close call and I'd say not unless cheap chinese offerings and certainly not respected brand names Like Leadshine unless second hand. Then you also have the other can of worms of Pulse generation and what's needed to connect too and get the rated speed out of those servos, which makes those potentialy cheaper servo's suddenly not so cheap.!!

It's here where servo's don't match these drives for ease of use and setup which to your average DIY user with limitied knowledge and time is quite important. . . . They essentially just plug n play and offer just the same accuracy.!! . . . Which was my whole initial point.

Boyan Silyavski
22-01-2014, 08:39 AM
In the spirit of this constructive discussion i want to say something more. When looking for motors after i decided i will go with cheap Chinese servo. Then i saw a Leadshine AC 220v servo at ali axpress. As leadshine seemed at that time to be a better choice i was ready to buy until i read this review: 06/01/2009 - Servo Drive Review and Comparison (http://www.thecubestudio.com/ServoDriveReview.htm)

It seems the jump should be straight to good brand servos, as the cheaper servos have some inherent problems and lacking of their own.

What i wonder after i read the review for the Leadshine servo is about the closed loop stepper and its control. It seems quite good on paper, but i still couldn't find a review from a real life implementation on a machine. I mean, would it stand well to what it promises?

There is also one possible combination that i believe will be nice, better and maybe even cheaper. Closed loop stepper/ stepper with encoder/ + second hand motion control board with its brand stepper drives , usually in the form of tight fit daughter boards or in the same enclosure. That sounds better to me , if one is ready to jump in the head ache world :-)

george uk
24-01-2014, 01:01 PM
Again, this is a totaly out there question,

Have been reading of people using servos as spindles, can not quite get my head around the pros and cons. Am wandering if i large one of these closed loop stapper motors, might be usable as a spindle under certan curcumstances.

What am i missing ?

Jonathan
24-01-2014, 01:05 PM
What am i missing ?

Torque and speed.

george uk
24-01-2014, 01:35 PM
HI Jonathon

would that be the same problem with a servo as a spindle or just these type steppers. I was considering options for slower cutting speeds for stone, and i can not find a reasonably prices spindle that has the torque at lower speeds,

Jonathan
24-01-2014, 01:41 PM
Normally servo motors have a constant torque rating up to a particular speed - often 3000rpm. For the stepper motor system that speed will be lower, however you can't really make a decision without comparing the available systems. Work out what torque vs speed curve you require, then just look which type of motors gets it for the least money.

EddyCurrent
24-01-2014, 02:27 PM
A Vector drive will give full torque at zero speed, but with an asynchronous motor not a servo or stepper.
Have a read, look at the vector drive section.
http://www05.abb.com/global/scot/scot201.nsf/veritydisplay/771cd1d59737abfac1257b0f0046ac22/$file/TechnicalGuideBook_EN_3AFE64514482_RevG.pdf

Jonathan
24-01-2014, 04:08 PM
A Vector drive will give full torque at zero speed, but with an asynchronous motor not a servo or stepper.

Using vector control (also known as, FOC - field oriented control) also gets rated torque to zero speed with synchronous motors. Servo motors are generally permanent magnet synchronous motors and they will use FOC and hence can output rated torque down to zero speed, so long as the motor cooling is sufficient.
Stepper motors are less well defined in their ratings - with FOC they will output some value of torque between zero speed and some value, however this torque value isn't generally quoted in the datasheet, so it is unwise to make a general statement except that this value will be lower than the holding torque.

Boyan Silyavski
25-01-2014, 07:45 PM
Some servos i have seen go to 6000 rpm or even more, so 3000 is not the limiting speed, when speaking about servos. But the answer to your question is the people use them on mills, belt driven , not cnc for wood. Some use them belt geared also for bestial rotary axis, where at the same time the axis can perform lathe function.

Colin Barron
23-06-2014, 09:38 PM
I am looking at Yaskawa servos. There are plenty of second hand and new old stock sigma 2 drives on ebay. I have a 200w drive and motor for 170. The software is free and a program cable is about 20. I looked at other makes but found they were as expensive second as new, or support was not so good. Drive sgdh02aeoy, motor sgmah02aaa61doy.