whats the matter!

[John S]

Torque curves are useless.
Sounds a far reaching statement but lets break it down.

For a start under what testing criteria was the motor under when the curves were produced as regards the driver ? This is never published and different drivers will give different results.

Some of the charts compare torque to PPS [ pulses per second ] and some compare torque to revs. There is no standard to any of this, for a start what voltage was used? Even a different of 10 volts will make all the difference to the output.

Add to that these motors are on a continual improvement path and what ships in one container isn't guaranteed to be in the next container even though they have the same part numbers.
How can you guarantee that they even put the right motor with the right data sheet ?

99% of people at home shop user level has no way of ever checking these statements. I have literally spent hours and hundreds of pounds trying out different motor / driver combinations on various applications and just because it works for me it may not work for the next guy if he's got a slightly different setup / design.

I have seem and had large 34 motors what should have oodles of torque crawling away only to perform brilliantly when the motor has been swapped for a smaller higher speed type 23.

Compare the torque curves on these and it doesn't make any sense, that 34 has 3 times the torque of the 23.

The secret with steppers is voltage, you must have 20 to 25 times the motor voltage to get anything like performance out of them.

Many of the so called high torque motors are high voltage low amps like 5.4 volts at 2.1 amps, this puts that motor expecting 108 to 135 volts, something that's not obtainable with current drives.

Now the smaller motor rated at say 2.7 volts and 4.2 amps needs 54 to 67 volts, obtainable with the larger more expensive drives and workable with the lower voltage drives but not at peak torque.

Factor in the worse axis like the Z with it's weight and you now have a whole load of figures that even a mathematician can't work out.

It's no secret that ARC's X3 CNC kit that later morphed into the Sieg KX3 took over a year to just run all the tests to get the right motor / driver / power supply combination.

There is also a move to used switched mode supplies as these are cheap and easily imported. These are not really suited to CNC motor operation as they do not have the capacitance to hold the voltage stabe and still keep the power required on tap.

A simple power supply rated at 60 volts for something like an X3 requires a capacitor of 100 volts working and a minimum of 22,000 uf to be rated correctly. This component will cost about £30 to £35 on it's own so you can see that switched mode supplies must cut corners to arrive at the price they do.

John S.

[ptjw7uk]

Thank you JohnS for stating what I have slowly realised is the truth, its a black art form.
So what are we hsm's to do, we do not have the resources to test and must rely on those who have gone before but we bocome stuck if we change to what we assume to be for the better one part of the system only to find it isnt quite right!

Anyone know where I can get some reasonable priced encoders in the UK(only got one arm left and no leg!)

Peter

[Robin Hewitt]

Many of the so called high torque motors are high voltage low amps like 5.4 volts at 2.1 amps, this puts that motor expecting 108 to 135 volts, something that's not obtainable with current drives.

I hope I'm getting that and more with Gary's "plug it straight in to the mains" drivers, but you can't switch the current below 3.18 Amps rms

The first worry with them was the heat, but I'm getting used to that, doesn't seem to worry the motors.

New problem is the noise, set at 800 steps per rev they make it clear I'm not letting them get anywhere near the sine wave they want to do. Working on it

Stepper perfection is a never ending series of upgrades :whistling:

[John S]

Robin,
We really need Gary to reply on this, I have only ever had one test unit of these to look at and the one we had was just an 80 volt driver with an in built power module.

Gary's may be different, after all there are many driver manufacturers out there although a lot share the same technology.
The one I had was rated at 240 volts but turned out to be designed for a 220 volt power supply, as our mains here can easily get to 265 volts at peak it didn't last long :nope:

[Robin Hewitt]

The one I had was rated at 240 volts but turned out to be designed for a 220 volt power supply, as our mains here can easily get to 265 volts at peak it didn't last long :nope:

This one protects itself, if it goes over-volts it switches off and needs a reboot. You have to watch the fault output. The manual clearly states, "can be connected directly with single-phase 220V AC power to save the cost of a transformer".

[John S]

Yes I read that but we are not on 220 volts like a lot of countries, so it it defaults on one axis whilst cutting and trips you wreck the job ?

.

[Robin Hewitt]

so it it defaults on one axis whilst cutting and trips you wreck the job ?

No, surprisingly enough, so long as it doesn't expire on a G0, they've thought of that :naughty:

It tells you before it fails so you can stop everyone, then powers up to the same phase settings it was on when it failed.

[John S]

No, surprisingly enough, so long as it doesn't expire on a G0, they've thought of that :naughty:

It tells you before it fails so you can stop everyone, then powers up to the same phase settings it was on when it failed.

So that relies on you being there ?
Not a lot of use to me then, My big CNC is running all day today, last half hour I've been down Asda buying a new kettle but it's still churning out parts.

.

[Tom]

I've wondered about this too. We buy motors based on charts and curves, but the manufacturers don't declare anything of the test method.
How tricky would it be to build a simple test rig? Budget ~£100?

Stepper -> direct coupling -> load
The load could be something like a motor with variable resistance across it.

I'm thinking: set a stepper, drive, and power supply combo running at a known RPM from mach or EMC2. Measure the voltage and current through the load, gradually increasing the load (by removing resistors in small steps), until the stepper stalls. Plot the data point, then increase the RPM and repeat.

Obviously the results would only be valid for the stepper/driver/power supply combination tested. And the results between different test rigs (if more than one were built) would not necessarily be directly comparable because of efficiency differences of the load motors. It would allow you to check the effect of a power supply voltage increase, or a new motor type, or a new driver, or moving to microstepping, or etc etc etc...

How much work is there to get to the point where the results might be comparable between rigs? Would defining a "standard" dc motor type be good enough (eg choose an appropriate RC motor P/N from a popular manufacturer)?

Just thinking out loud...

[John S]

Tom,
Good idea but still relies on "Is the motor in the box the same one the spec sheet is for ? "

And to be honest working with the Chinese I'm very sceptical.
Out of a box of 40 rotary tables that all had the exact same run out error according to the spec sheet? every one ?