[Sray]

Can I get some opinions on the following two torque curves? To me the performance looks similar with some slight differences especially at the lower rpm range but at 1050 rpm they appear very close.



Although I will be running them at 60V these curves only show up to 48V.

One of these appears to be available in the US and can be delivered in less than a week where the other would ship from China with a much longer delivery time and cost considerably more. I would like to know if by looking at the curves you see anything that stands out between the two, performance-wise.

[JohnZ]

The top motor is rated at 3.0 MH while the bottom is 1.85. That being said, the top motor needs 69v to compared to 48v on the lower motor. So, if you ran both at 48v, the top motor is slightly handicapped at higher RPM, but if ran at 60v, the lower motor would run hotter. It's kind-of a toss-up between the two.

IIRC, the tried and true formula is square root of the motors MH rating times 32 equals the sweet spot in volts to feed the controller to maximize the speed while reducing the motors heat.

That formula has been around a long time and controllers can morph what is powering the motor to reduce heat, so you should be able to go above that a little bit.

For example the top motor.

Sqrt(3.0)*32=55

And the bottom motor

Sqrt(1.85)*32=43

So, the top motor needs 55/43 or about 1.3x higher voltage supplying the controller for a fair comparison between the two motors.

That means the top motor needs about 60v to compare with the bottom motor at 48v.

I hope that this helps.

John Z

Sent from my SM-G991U using Tapatalk

[JohnZ]

One additional item I notice on those two torque curve charts... the bottom chart, see how there is a jump in torque from 24 to 36v, but at 48v, the lines converge at 1050 rpm and the chart stopps there. This is probably the limit of the motor. If fed 60v, the motor may perform even worse due to overheating at the higher rpms... even at 48v, it is above the sweet spot for a motor having 1.85 MH, if I am understanding things right.

Sent from my SM-G991U using Tapatalk

[joeavaerage]

Hi,
that formula is nonsense....it may have been around for a while but it has no scientific validity.

To make a motor run fast use the highest possible voltage.....end of story.

Heating is caused by current NOT voltage.

A stepper driver will apply the full drive voltage to its stepper at the moment of switching or 'stepping', and keep it applied until the current builds up to
the motors rated current, usually no more than a few micro seconds. Thereafter the drive uses PWM to maintain that current, ie it uses rapid switching to
effectively reduce the voltage down to just a few volts required to maintain the current.

This means that on average the voltage applied to the motor is just a few volts, maybe 5V or less, despite the power supply input voltage being 60V or more.
The bottom line is that the input voltage has little to no role in heating whereas the current does.

The length of time taken for the current to build up to it rated value is dependent on the inductive time constant, the product of the winding resistance and
the inductance. The lower the inductive time constant the quicker the motor can step and therefore the faster it can run. The length of time taken for the current to build up
can be reduced by applying a higher voltage.....if you like the higher voltage 'forces' the current through the inductance.

To get the fastest possible motor use one with the lowest possible inductive time constant with the highest possible voltage. End of story...its not hard...its not
complicated....lowest inductance, highest voltage.

Pay the extra dollars for low inductance and drivers with the highest voltage rating and a beefy high voltage DC power supply and you'll have steppers that you'll use
for the next twenty years....try to save a few bucks and throw them out in a month.

Craig

[joeavaerage]

Hi,
I just got two servo kits from China, a 750W kit and a 400W kit, total weight maybe 15kg.

Took 4 days from Hong Kong on DHL to New Zealand, sure it cost $285 but I got what I paid for, and four days is less than UPS
takes to send something across the state let alone the country.

Craig

[JohnZ]

Joe, was that just for shipping?

With regards to the formula I used above...

I believe it was referenced earlier, but I have been going by the information provided by GeckoDrive. They have far more expertise than I will ever have on this subject.

https://www.geckodrive.com/support/s...or-basics.html

In section 6...

-----------

To figure out what the maximum power supply voltage should be, use the following formula with the motor’s inductance in millihenries (mH) used for the L value.

32 * SQRT (L) = VMAX

If you are using several different models of motors on the same power supply use the lowest inductance rating in the above formula. This will ensure that your motors will not overheat due to the voltage being too high.

------------

There must be a scientific and experimental reason behind this information.

This is at least what I based my recommendation off of.

Thanks, John Z

Sent from my SM-G991U using Tapatalk

[joeavaerage]

Hi,
I am well aware that Gecko published that formula, and its proven useful as a rule of thumb, but it is just that, a rule of thumb.

Use the highest possible voltage.....it doesn't get any simpler than that.

Craig

[dazp1976]

Although I will be running them at 60V these curves only show up to 48V.
One of these appears to be available in the US and can be delivered in less than a week where the other would ship from China with a much longer delivery time and cost considerably more. I would like to know if by looking at the curves you see anything that stands out between the two, performance-wise.

Because of the slightly lower inductance of the bottom one, it's curve is more consistent which makes it smoother running.
The top one has a good start and then drops a little like a stone.

Chances are you will NEVER get near 1000rpm on any stepper. Around the 800rpm mark is getting towards their limit if you don;t want any missed steps etc.
Feeding a lower inductance motor with higher voltage gives them more consistency and smoothness.
Higher inductance can also lead to more chance of resonance noise which in it self can make them stall.
In both cases the higher voltage of 60v+ gives them more reliability at higher speeds.

Hands down the 24HP39-5004S is the one to get. Don't be put off buying things from China if you're getting something good and worthwhile.
Longest wait I've ever had in the u.k. is 14 days via snail mail.
You don't really feel it at all and you always find something on the machine to fiddle with in the meantime.

Both motors will perform well on the higher 60+ voltage though. The top one may be less efficient and use a little bit more power (current) on acceleration so may run slightly hotter than the bottom one.

[dazp1976]

Heating is caused by current NOT voltage.
Craig

In a way you are right and wrong.
It can be caused by voltage. I've blown many brushless motors with voltage increase but that's mainly on RC vehicles.
Their rpm is relative to 1v = 1000Kv but you get different kv ranges so you can match rpm properly to the voltage you want to use for ttheir 'safe' rpm zone.
You can do high voltage on high kv and take them to 60,000rpm for speed records but they can melt in the process, even if geared to pull lower currents.

Steppers are different.
They're designed for 1000rpm and then stall so you can't really overrun them no matter what voltage you use.

[joeavaerage]

Hi,

In a way you are right and wrong.
It can be caused by voltage

You did not read or understand what I said 'The bottom line is that the input voltage has little to no role in heating whereas the current does.'. The operative word here is 'input' voltage.
That is because the input voltage (to the driver) IS NOT the voltage actually applied to the stepper.

Lets take OP's stepper for example:
1.85mH winding inductance
0.48 Ohm winding resistance
rated current 4.8A

At 4.8 A the winding will exhibit a voltage drop of 4.8 x 0.48 =2.304V This is the average voltage applied to the winding. Thus the power dissipated is:
P=V²/R
=2.304²/0.48
11.05W

Alternately you could calculate the power lost by the current:
P=I²x R
=4.8² x0.48
=11.05W

Both calculations are equivalent....but note that neither calculation involves the driver input voltage. The driver input voltage is responsible for forcing the current through the inductance initially,
but its cut down to 2.3 V by PWM shortly thereafter. It's that voltage which determines the power loss.

Craig

[Sray]

Tons of great info guys! Thanks for all the input.

I went ahead and ordered the 1.85mH motors from China. Had to order different couplings as well. Hopefully these motors will actually ship so I don't have to order motors again.

Thanks guys!

[Sray]

So I have finally got my CNC built. I have it communicating with LinuxCNC. I am in the process of setting up the ini/hal files and am needing some help with figuring out the correct microstep/current settings and the stepper timings.

Here are the components that I went with:

Mesa 7i76e (ethernet)
X&Y Stepper Motors (3) - Nema 24 (5.0 A)
Z Stepper Motor - Nema 23 (4.2 A)
Digital Stepper Driver - DM860T

My first question is about setting the dip switches on the drivers.
- For the current settings do I go by the Peak Current or the RMS current? Which current from the following list would I use for the 4.2A motor and for the 5.0A motors?

Peak Current RMS Current
2.40A 2.00A
3.08A 2.57A
3.77A 3.14A
4.45A 3.71A
5.14A 4.28A
5.83A 4.86A
6.52A 5.43A
7.20A 6.00A

My second question is about the Stepper Info section in LinuxCNC PNCConf setup. It is asking for the following settings. I have looked on the Stepper Drive Timing page (LinuxCNC Documentation Wiki: Stepper Drive Timing) for the DM860T but it is not listed. I was hoping someone may be able to help me with the correct timings for this driver?

Step Time
Step Space
Direction Hold
Direction Setup

Here is a link to the DM860T manual https://www.omc-stepperonline.com/download/DM860T.pdf

Thanks for any guidance!
Shawn

[rodw]

Start with:
setup and hold times to 10000.
step time and step space to 5000 (5 us)
The only reason to reduce these times is if you cannot reach the desired speed with 5 us/5us timings (100 KHz)

The Chiniese drives are usually quoted on Peak voltage so work on 3.77 amps and 4.45 Amps

I have run steppers at 25 x microstepping with Mesa hardware but I would not go that high when starting. Say use 10 x to start with.

we want things to move for you before trying any fancy stuff!

[joeavaerage]

Hi,
it common to run steppers at or near rated current. Running them over rated current risks failure due to overheating. Running them at less will mean they never
fully achieve the torque they should. Having said that they will run a little cooler and no harm .Therefore you can run them at lower currents without risk, and thats
what I'd suggest.

The current rating is about the acceptable heat load of the stepper as determined by the manufacturer and RMS current is directly related to heating...so use the RMS current.

I'd run the 5.0A motors at 4.86A RMS or even 4.28A RMS, at least to start with. If you think they are getting too hot, reduce the current a bit.
I'd run the 4.2A motor at 3.71A RMS, at least to start with.

Microstepping is all about smooth motion, astronomers first used microstepping to get smoother motion for their steppers on their telescopes. The higher the microstepping
the smoother the motion, but then you have to signal Steps ever faster, so there is a balance. 8, or 10 or 16 microsteps per full step is about right, it gives a pretty
smooth motion without unduly upping the signal rate into difficult territory. The truth is that micostepping is not hugely important, if you make a reasonable decision
when setting up you'll leave it alone thereafter, its not something that you have to tune, just set and forget.

8 microsteps = 1600 pulse/rev
10 microsteps=2000 pulse/rev
16 microsteps=3200 pulse/rev

Step Time
Step Space
Direction Hold
Direction Setup

These are to do with the timing of the signal sent to the drivers.
Usually the Step signal is a square wave of 50% duty cycle. So the Step Time would be equal to the Step Space. Say 5us each. That would result in a square wave
of 10us or 100kHz, which sounds about right.
The Setup and Hold times are the duration that the Dir signal present before the Step signals arrive and the time after the Step signals end. That will depend
on the drivers but about 5us each is about right.

Craig

[Sray]

Thanks guys for the info. I appreciate it.

As with a lot of the info I have found on building my CNC there is conflicting info. But it gives me a starting point and I now understand how to adjust the settings and what to keep an eye on.

Since we are on the topic of steps and direction settings I have a question about my ini file for LCNC. I have been scouring the forums for info on Homing and how to setup my two motor gantry. I have read a ton of ini/hal files and I noticed in most of the ini files that the STEP_SCALE setting is usually low, like around 150-400 (a lot of them are 200). Mine is 10159.9997. This was calculated by PNCConf. Is this number useful or even used? Should I change it?

Thanks!

[rodw]

The chapter in the Linuxcnc online docs called Homing configuration covers homing. Ignore anything else as it could be dated.

HOME_SEQUENCE sets the order of homing starting at 0. So maybe the Z axis is homed first to get it out of the road before moving.
Your two joint gantry axes should have the same number but to tell linuxcnc thay are joined together, make them negative.
To square the gantry, use HOME_OFFSET

Step scale is the number of pulses to travel one device unit. (mm or inch). I always calculate it by hand.For a 5mm pitch ballscrew, with 10 x microstepping with no reduction its simply 200 x 10 pulses per rev /5 mm/rev = 400 per mm ( x 25.4 = 10160 / inch). Does that sound like your setup?

[Sray]

Thanks Rod.

Yes that is exactly how I have mine set. So how can so many others be running fine with something like 200 or 400? I even saw one that was 80. It makes me think the number is arbitrary.

As for my HOME_SEQUENCE I currently have my dual motor Y axis set as -2. I have seen both -1 and -2 used. Does the number matter as long as it is a negative?

[rodw]

If you set your machine up with metric (mm) device units like I would in my country, it would be 400. Definitely not ariitrary unless you want arbitrary sized parts.
-1 or -2 won't matter but might alter the sequence joints are homed. There were some changes to this since I built my machine and while my setup is still valid, its not the default per the current documentation

[joeavaerage]

Hi,
in Mach there is a similar setting but its called 'Step per Unit'. It is by no means arbitary, but reflects the microstepping choice you have made and the mechanics of your
linear motion system.

I use servos and have programmed them (electronic gearing) to be 5000 pluses per rev. They are direct connected to 5mm pitch ballscrews, thus the 'Steps per Unit'
is 1000 (5000 steps/5mm), which is highly convenient. This also implies that the smallest movement I could make, being one step, is 0.001mm or 1um, there again a highly
convenient number.

Lets say you have a belt drive, say 5mm pitch between the teeth and a 30 tooth pulley (150mm diameterical circumference) direct onto a stepper with 8Xmicrostepping:
1600 steps/150mm=10.666666 Steps per mm. That would mean the the smallest move this machine could make , still being one step, is 0,09375mm.

So as you see the numbers can vary widely but are by no means random.

Craig

[Sray]

That makes sense guys! Thanks for the info.