[Streetslab]

I have built a big CNC machine with 4 Nema 34 motors. It's 1.70 meters per 1.50, and the gantry is moving via 2 synchronized motors. All running via Linux CNC.
It's a Longs Motor Kit with DM860A on 60volts and 3.6 amps, and the motors are rated for 11 nm of torque, and a very basic breakout board, which i think it's the problem.
Whenever i push the machine over a given feed rate, say around 70%, one of the synced motors jams.
If i run slowly then it's all fine.
Useless to say it's too slow and i would really love to go faster, especially now that i've got the new spindle. Will a Mesa card solve my problem?
If yes, what's the best/fastest one?

Many thanks.

[alan_3301]

Hard to say for sure, but you are probably running out of torque at high speeds.
A lot of factors will influence top speed, top of the list is drive ratio, and stepper voltage.
If you can look up your stepper motor and find the inductance value rated in mH, then you can figure optimum voltage, I'm guessing yours are around 22mH, so they need around 150V max.

Classic case of mismatched parts on a conversion.
If your sherline has 1/4 - 20 leadscrews like most, then you would be best running around 280 oz-in motors @ 48V
These would run 2-4 times faster than your 1600 oz-in motors, with the same resolution.

Hi guys. Sorry to pop in here and ask another question, but it's somehow related, especially for the Mesa cards.
I have built a big CNC machine with 4 Nema 34 motors. It's 1.70 meters per 1.50, and the gantry is moving via 2 synchronized motors. All running via Linux CNC.
It's a Longs Motor Kit with DM860A on 60volts and 3.6 amps, and the motors are rated for 11 nm of torque, and a very basic breakout board, which i think it's the problem.
Whenever i push the machine over a given feed rate, say around 70%, one of the synced motors jams.
If i run slowly then it's all fine.
Useless to say it's too slow and i would really love to go faster, especially now that i've got the new spindle. Will a Mesa card solve my problem?
If yes, what's the best/fastest one?

Many thanks.

[Streetslab]

Hi Alan. Thanks for the answer. Here's the label of one of my motors. Cannot really get the number you asked.
Attachment 275824
Unfortunately i was in doubt on the size and motor type, so i went for the big ones.
I'm on direct drive on a 20 mm trapezoidal leadscrew. I'm using aluminium couplers and everything is aligned properly.
I'm running at 1600 steps per rev, set like so:
SW5 SW6 SW7 SW8
ON OFF ON ON

[alan_3301]

20mm diameter leadscrew (how would you fit that on a sherline)
or is it 20mm travel per turn

[Streetslab]

Sorry for not having specified. It's 20mm diameter. The motor's shaft is 14mm. Found some nice 14 to 20 couplers on aliexpress.
The travel is 4mm per turn.

[alan_3301]

ok I at 60v on that motor I would guess a max reliable speed of 200 rpms?
are you able to jog at 800mm/min or 31 inches per minute?

[Streetslab]

I'm pretty sure i can. I was filming a solution i'm trying for a dirt cheap cable management solution (scrap used cleaned bike chains get for zero bucks on a bike store) I was able to run the whole Y axis yesterday in less than a minute, and that's roughly 105 cm. I will double check later.
Are you about to say that a speed improvement is not possible in the actual conditions? And that it's not a breakout board issue?
If yes, what should be changed in my setup to cut / mill faster?

Thanks a lot for your help.

[alan_3301]

Sorry I thought you had a sherline..
Why don't you start a new thread and post some pics of your machine.
anyways, I would get smaller motors, or change the screws to around 20mm per turn to take advantage of the motors torque, but you lose resolution.
It all depends on what your requirements arel

[ger21]

You have a combination of big, high inductance, slow motors, and leadscrews that are very inefficient, and require fairly high rpm's to get good speed.

On a machine that size, you should ideally be using screws with at least 10mm pitch, which would double your speed. ballscrews would be about 3x more efficient than your leadscrews, which should give you even more speed.

Another member here built a machine with the same motors you have, and found them to be very slow. He switched Gecko drives and 400oz motors, and increased his speed by 5x.

[Streetslab]

You have a combination of big, high inductance, slow motors, and leadscrews that are very inefficient, and require fairly high rpm's to get good speed.

On a machine that size, you should ideally be using screws with at least 10mm pitch, which would double your speed. ballscrews would be about 3x more efficient than your leadscrews, which should give you even more speed.

Another member here built a machine with the same motors you have, and found them to be very slow. He switched Gecko drives and 400oz motors, and increased his speed by 5x.

Oh my...
Well, the price to pay for lack of experience. I like my big and powerful motors though, i think it's mainly a problem of leadscrew pitch, which i will eventually upgrade after i have improved the machine in all of it's parts and made some money out of it.
That Mesa PCI card though, is a very tempting purchase. I will actually be in San Francisco next mont, could pay a visit to our friends at Mesa.

Thanks for the advices guys. This forum is great and you are all very nice and helpful.

[ger21]

Well, the price to pay for lack of experience.

Your not the first to pay this price.

I like my big and powerful motors though, i think it's mainly a problem of leadscrew pitch

It's both, but I'd say that the motors are a bigger issue.

[Streetslab]

Thanks Gerry.
One thing.
How a smaller motor with less torque, can be more efficient than a bigger one?
Do they have a different torque curve?
My understanding is that stepper motors will give the most of their torque at low revs, and the more rev will go up, the more torque will go down (a lot).
So if i pick a smaller motor and make it run faster, won't things get even worse?
For sure i'm missing some bits here, but i don't know which ones. Would you be so kind to explain?

Thanks a lot.

[ger21]

Typically, the larger the motor, the faster it's torque drops off.
And the higher inductance,, again, the faster the torque drops off.
low inductance, high current rated motors can spin at much faster speeds than big, high inductance motors.

So, a 400oz motor might have 200oz of torque at 1000 rpm.
A 1200oz motor might have 200oz of torque at 300 rpm, and not enough to even be usable at 500 rpm.

It's both, but I'd say that the motors are a bigger issue.

[Streetslab]

Typically, the larger the motor, the faster it's torque drops off.
And the higher inductance,, again, the faster the torque drops off.
low inductance, high current rated motors can spin at much faster speeds than big, high inductance motors.

So, a 400oz motor might have 200oz of torque at 1000 rpm.
A 1200oz motor might have 200oz of torque at 300 rpm, and not enough to even be usable at 500 rpm.

It's both, but I'd say that the motors are a bigger issue.

Now it's 100% clear. I will also update my drivers / steppers combination then.

Thanks a lot.

[James Newton]

ger21 is generally a very knowledgeable guy and I would trust his advice... I would recommend trying things in this order:

1. Change the lead screw first. That may be enough.

2. Upgrade the power supply and drivers and keep the motors just to see if you can't get those motors to move. The main thing is to raise the drive voltage. More voltage = more speed. A couple of gecko-drives at 80 volts may just do it. Use one driver per motor, and feed them both from the same step pulse. Don't try using one driver for both motors (not sure if you were doing that)

3. THEN upgrade the motors.

In the end, it's no loss because you probably had to do all that anyway.

[Streetslab]

ger21 is generally a very knowledgeable guy and I would trust his advice... I would recommend trying things in this order:

1. Change the lead screw first. That may be enough.

2. Upgrade the power supply and drivers and keep the motors just to see if you can't get those motors to move. The main thing is to raise the drive voltage. More voltage = more speed. A couple of gecko-drives at 80 volts may just do it. Use one driver per motor, and feed them both from the same step pulse. Don't try using one driver for both motors (not sure if you were doing that)

3. THEN upgrade the motors.

In the end, it's no loss because you probably had to do all that anyway.

Hi James, thanks for your bit of advice as well. I saw that Gerry is present in a lot of 3ds and very helpful and knowledgeable.
A thing that you wrote lighted a bulb in my head. When you say:" feed them both from the same step pulse" you mean just to double the wires coming out from the same pulse pin out to the second motor on the axis???
Because what i'm doing is synchronizing them via software in the .hal file that is then read by LinuxCNC. Is there a better/preferable way to do it?

[James Newton]

Hi James, thanks for your bit of advice as well. I saw that Gerry is present in a lot of 3ds and very helpful and knowledgeable.
A thing that you wrote lighted a bulb in my head. When you say:" feed them both from the same step pulse" you mean just to double the wires coming out from the same pulse pin out to the second motor on the axis???
Because what i'm doing is synchronizing them via software in the .hal file that is then read by LinuxCNC. Is there a better/preferable way to do it?

Ah. I didn't know LinuxCNC could do that. Yeah, that is better. Sharing the step pulse would have the advantage of not being susceptible to LinuxCNC being overwhelmed or having a bug, but that is VERY unlikely. Sharing the step pulse has the disadvantage that the driver for the port pin might not be strong enough to drive two driver input, which is more likely... although still pretty unlikely in general.

[wendtmk]

Hi James, thanks for your bit of advice as well. I saw that Gerry is present in a lot of 3ds and very helpful and knowledgeable.
A thing that you wrote lighted a bulb in my head. When you say:" feed them both from the same step pulse" you mean just to double the wires coming out from the same pulse pin out to the second motor on the axis???
Because what i'm doing is synchronizing them via software in the .hal file that is then read by LinuxCNC. Is there a better/preferable way to do it?

The best way to do it is to slave another axis to the first which is easily done with LinuxCNC, and is the way you currently have it set up. I have two axes slaved, X and A on my saw beveler, to move my gantry down the X axis. You do have to reverse the A direction, since it's mounted opposite from the X. Let the magic of software rather than a physical connection do the work for you.

[Streetslab]

The best way to do it is to slave another axis to the first which is easily done with LinuxCNC, and is the way you currently have it set up. I have two axes slaved, X and A on my saw beveler, to move my gantry down the X axis. You do have to reverse the A direction, since it's mounted opposite from the X. Let the magic of software rather than a physical connection do the work for you.

Hi Wend. Thanks. That is exactly how my setup is working. Now i'm wiring some microswitches for the limits and changing the skates bearings enclosed in some pieces of corrugated tube with hose clamps (LOL, but it works just fine) with some 20mm bearings plus housings i've just received today.
Then my machine will finally be ready to mill some stuff