China NEMA17 Stepper Motor Torque Rating Scam

the common china 4.8kg.cm (50N.cm) rated stepper motors are just scam. i have 17HS5415 and 17HS8401 for 3d printer design purpose that cant keep up with my heavy design big printer. so i got curious and tested it. it turned out the steppers can only hold comfortably around 1kg.cm of torque and failed at 2kg.cm. my setup is simplest 1cm radius (r) coupler connected to the stepper and a string lifting a weight (m, 1-2kg) from the floor, the torque should be Fr = m.g.r. no need mentioning on the electronic side, its working as it should and i think i'm good on that side (EN pin LO to enable holding torque setup). anyone with an experience to comform this is a myth or truth? in my case this is the truth. in reality, less than 2kg.cm of holding torque only. any suggestion for bigger torque or true 5kg.cm (but same dimension or size) stepper motor at reasonable price? thanks...

Edited 3 time(s). Last edit at 07/18/2015 03:01PM by shafri.

Did you use the rated current? Was microstepping on or off?

i used A4988 stepper driver as is controlled by arduino uno. microstepping is at full 16th microsteps. speed is at lowest possibly imagined, or no step command at all, just driver enabled ON.

Edited 2 time(s). Last edit at 07/18/2015 03:08PM by shafri.

How do you know how much current was in the motor?

i dont know, i cant measure with convenience, i have a dc hall effect cc-65 hantek current probe, last time i measured the reading were jumping all over up and down in my dso so i cant be sure. but i'm assuming the A4988 is capable to supply full current. A4988 is rated at 2A, the motor is rated at 1.8A. so.. ??? maybe its also possible the culprit is in the A4988 driver module but i felt its less likely...

to add value to the claim, here is a attached picture of my setup. light brown string still attached to the motor. whats not shown is my 2 x 1kg dumbell weight. sorry i need to go to bed i hope i may have promising answer tomorrow. thanks and good night...

Motor torque specs are based on current in the motor's coils. Your test is meaningless until you know that the current in the motor matches the motor manufacturer's spec.

The only way you can know the current in the motor is to measure the current. Don't assume that measuring the reference voltage of the driver (or the voltage across the motor coil) will provide an accurate indication of the motor current.

Edited 1 time(s). Last edit at 07/18/2015 03:55PM by the_digital_dentist.

The A4988 can't supply 1.8A unless it is very well cooled. This requires a large PCB. If you were using a stepstick-type driver module, I think there is zero chance that it could supply 1.8A.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

ok, maybe the ready made driver is the culprit. any suggestion for home made powerful driver? i'm guessing this test is not as easy as connecting a battery to the motor input is it? can i connect the battery to one winding to hold the rated torque? using V=IR, measure the coil's R and provide the necessary V to get the 1.8A? correct? wont i burn the coil winding in trying such a test? thanks for input guys...

i measured the coil "static" resistance is 2.39ohm, to get 1.8A through the coil i need around 5V power supply, is it a safe bet to statically connect 5V to one of the coil to redo the holding torque test? i'm waiting reply before burning anything. i dont mind burning something if it can be get in the shop next door, but this one is 2-4 weeks lead time order all the way from china, so its an expensive job timely wise, thanks.

Why are you so concerned about testing the motor torque?

You mentioned a "heavy design big" printer. I suspect you're doing what so many others do- you buy a cheap motor first and build the printer around it and fail to understand why it doesn't work the way you want it to. The normal way to design a machine is to define the moving masses and performance requirements and then select a motor and its control electronics to meet those requirements. If you don't select the motor to do the required task, and simply use whatever is at hand (or worse, whatever is available cheaply) you may have to reduce your expectations of the machine's final performance.

Fortunately, you can easily figure out the required torque based on the performance requirements and data such as moving mass, etc., by using some of the easy to use motor sizing tools at web sites such as Oriental Motor. You may have to make a few guesses at some of the parameters (coefficients of friction, etc.) it asks for, but the returned torque value should be in the ball park. Then you buy a motor that provides a bit more torque than it says is required, provide a suitable driver, and your machine will work as expected.

If you have a heavy mechanism, the pololu stepper drivers are unlikely to be able to drive the motor in a satisfactory manner. You might look into DSP based drivers that will handle more current and allow finer microstepping to smooth the motion and reduce vibration and noise.

i built a heavy but calculated design with mechanical counterweighing mechanism. based on the advertised 50N.cm spec, the motor should be able to lift my design no problem. but since it failed to do so, i conducted the test. so the 17HS8401 + A4988 combo wont reach the expected spec... i'm thinking of building a diy driver for it... thats the plan B, if its still fail, i still have plan C and D. i posted because i want to make sure the validity of my claim... and i want to know if anybody have done to the edge of its claimed spec limit... and i'll try as much as possible to use the stepper that everybody used, easily available in the ebay, cheap and very well developed and supported in this community. wish me luck...

plan C: since i used 5:1 step up gear in my original design to increase speed, i still have room to reduce the gearing to reduce the torque needed. reprint some assembly rigs and get along with slower Z axis motion...
plan D: non standard setup motor... geared down dc motor with diy rotary encoder to maintain motor positioning... but thats more electronic circuit and mcu programming design work. i hope i will not reach this plan.

i measured holding current (70% peak) is 0.4A, so peak current for A4988 is only limited at around 0.57A, 1/3.5 of its rating. so yeah i guess now its the driver who is bottlenecked assuming linear proportion, the holding torque should be 5/3.5 = 1.42kg.cm very close to what i got of the stepper motor. the stepper rating should be ok, so for now my first claim is busted.

Edited 2 time(s). Last edit at 07/19/2015 08:05AM by shafri.

All the little stepper driver modules I've seen that look like the ones you have, have a small pot to adjust the current. You should be able to get more than 0.57A from them. What is the power supply voltage that you're using for this test?

I use a pair of these drivers with 32V power supplies to drive the X and Y axis NEMA-23 motors in my printer.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

According to microstepping-myths-and-realities

Microsteps/full step Holding Torque/Microstep

1 100.00%
2 70.71%
4 38.27%
8 19.51%
16 9.80%
32 4.91%


Suggesting that your 1kg.cm is quite high really.
-a

Quote
the_digital_dentist
All the little stepper driver modules I've seen that look like the ones you have, have a small pot to adjust the current. You should be able to get more than 0.57A from them. What is the power supply voltage that you're using for this test?
I use a pair of these drivers with 32V power supplies to drive the X and Y axis NEMA-23 motors in my printer.

i managed to burn one of the module earlier by just slightly adjusting it, yet torque improvement is almost nonexistence from hand twist resistance feel test. i used 12V power supply. thanks for the link... but $36 per stepper is quite expensive for my taste since i have 8 steppers in my design and at least 4 of them (Z axes) require high torque. i try cheaper option first before anything, thanks.

Quote
alan richard
According to microstepping-myths-and-realities
Suggesting that your 1kg.cm is quite high really.
-a

thanks for the link. to be frank, i'm quite hesistant to believe whats written on paper (stepper's spec in ebay), but i can use that as initial design info when the items in delivery. but when something fishy like what i've experienced, i will conduct "practical" or "empirical" test, the result, thats what i will believe 100% and use in the next redesign. i tested both static rotor EN on (holding torque only) and very slow rotation at 16th microsteps, the stepper can hold on the air or lift up 1kg load happily at 1cm radius... since this is torque priority, i will setup anything possible to maximize it, even if i have to use full step only and reduce the speed. but iirc i did test full step vs 16th microstep earlier, torque strengths were almost identical from "hand twist resistance feel test".

Edited 1 time(s). Last edit at 07/19/2015 11:31AM by shafri.

I don't think you've experienced anything fishy. I think you're misunderstanding a lot of things all at once. Your test techniques, too, leave a lot to be desired.

Motor torque required to do the job is proportional to the load (which can have a very complicated definition). Motor torque produced is proportional to coil current. Both the motor and its driver have to be up to the task at hand or the machine will not perform the way you intend. If the motor can't produce enough torque to perform as needed when operated at its rated current, you need a higher torque motor. Assuming you have a motor with an adequate torque spec to do the job required, if the stepper driver you are trying to use can't deliver the current the motor needs, you have to switch to a higher current motor driver (such as the parts I linked or other driver that can deliver more current).

As you have discovered, destroying those little stepper modules is very easy when you try to adjust the current setting pot. You can short out one of the other contacts on the board or even break the little adjuster off the pot. You have to be extremely careful with them. The module should allow you to set the current up to about 2A (if you can keep it cool). There is no rotational stop on the little pot, so when you adjust it, current may jump from a lot to very little (or vice versa) very suddenly, depending on the rotational position. The torque will jump (as will the driver IC temperature) along with the current in such a case. This video shows how to adjust the current to set the thing up for operation: https://www.youtube.com/watch?v=89BHS9hfSUk I would not use this to test the motor manufacturer's torque specs- for that you need to measure the current in the coils instead of relying on the reference voltage on the driver module. Measure the current with a DMM, not a hall effect probe for an oscilloscope, and forget about finger tests for anything but checking to see if the motor is energized.

thanks for the youtube link, good info...

since its concluded that it is the A4988 module that is bottlenecked, i then read carefully the datasheet and made some modification to the module. now the stall (hold) current is 0.8A per phase (set at 1A Imax limit) and can hold and slowly lift a 2kg.cm torque, but only marginal, a slight jerk will result in miss steps. since the current is now doubled, i expect the torque will too at around 3kg.cm, but in reality its only 2kg.cm, well its better than nothing. also note from alan saying 16th microstep will have only 9.8% torque, this is not true, infact its the other way around. i made the test, full step and 16th microstep. full step setup cannot hold (let alone lift) the 2kg.cm load, but 16th microstep is apparently stronger as i've mentioned above. so i will stick with 16th microstep setup.

the damaged A4988 module i mentioned earlier now i believe because i try to adjust the Vref while the circuit is ON and the steppers was rotating. this time i adjust it during idle (ENABLED off) using the DMM method similar to in pololu youtube linked by dentist. and my module is not using Vref = 2 X Imax, its Vref = 1.6 X Imax (Rsense 0.2ohm) reading the datasheet and probing passive components revealed all this truth specific to my module. the component in the module pcb also only limit the trimming to 1A (trimpot at dead end) so i mod it by changing a resistor divider value on it so the module is trimmable up to 2A, i'll cross my finger before adjusting to the max. at current 1A setting, the A4988 chip heated up in few seconds, not so serious its controllable, with abundant cooling i believe close to 2A current per phase is possible, wish me luck.

Edited 3 time(s). Last edit at 07/20/2015 08:09AM by shafri.

Quote
alan richard
According to microstepping-myths-and-realities
-a

You have misunderstood the link. Marginal torque is not holding torque. Microstepping has no material impact on stepper motor holding torque.

You should not be testing a stepper motor using an a4988 driver. Put a current source across the coils.

Note that holding torque has almost no relation to torque at speed. The torque in a high inductance motor falls off extremely fast.

Edited 1 time(s). Last edit at 07/20/2015 11:22AM by 691175002.

latest report... after modding and trimming the A4988 module very close to its designed limit (i maxed it to 1.8A peak (1.28A rms)), decent heatsink and fan, 16th microsteps configuration. here's the latest tabulated "my sloppy poorman method" torque spec (give or take):

Arms, Apeak, torque (kg.cm)
0.4A, 0.56A , 1
0.7A, 1.0A, 2
1A, 1.5A, 3

Arms is also the holding torque DC current (70% peak current during stepper running). i didnt test torque capability at 1.28Arms (1.8Apeak) because my setup can only bear barely 3kg of weight hanged to it, the string occasionally snapped and i'm afraid i may damage the attached flexible coupler on the stepper if i go further. although i'm currently testing it for continual operation at zero load heat test to see if the h-bridge mosfets built into the A4988 can survive for prolonged continuous usage (with proper heatsink and fan), after about half an hour its still running, i believe they can survive hopefully. it also seems the A4988 is limiting the current at the region of the 1.8A, maybe due to its max 2A spec limitation safety or something, so increasing Vref further seems to have no effect on the supplied current. as i said, its just a rough not really accurate result, but i believe its very close. guesstimating at 1.8-2Apeak setting, the stepper should be able to hold 4kg.cm torque. i also believe the indicated 1.8A stepper rating spec is rms value per phase (2.5Amax) since i feel the stepper is just a bit warmer at 1.8Amax operation. and further extrapolating the result, at its max spec 2.5Apeak, the torque capability should be really the claimed 5kg.cm. so i think i myself proved that my first claim is busted. there's no scam about the stepper spec. the problem was the underdesigned A4988 circuit module that i bought (i believe due to safety reason and normal medium size usage for muggles who not able to muck around). thanks for the discussion that led me to this result... FWIW...

Edited 2 time(s). Last edit at 07/25/2015 01:18PM by shafri.