[Vegabond]

Hello, i use Nema 23 270 oz/in motors, when i use the machine at full speed, i can run 15mm/sec before the motor on the x-axis is missing step.
The power supply is 36V DC 9.7A 350W, i feed the driver board with 32.8 V.

You can see the pictures below of the axis in EMC2. The manual of the MDFLY card is there too.

What can i do to make the motors faster? do i need a bigger driver? what driver do you recommend me to buy? I look at the nice G540, but isn`t it something that are cheaper? i think the Geckodrive name is quite expensive


Hope someone can help me.


Rober.

[Vegabond]

Can someone help me?


Robert.

[slavkok]

Faster you want higher, voltage you need.
And acceleration should be set proper. Your acceleration is fast. If you have heavy machine that's no go.
at 1500 hz you might just hit midband resonance.
Try to go halfsteep and see if you have same speed.

bye

[Vegabond]

Try to go halfsteep and see if you have same speed.

bye

Thanx, but what is the difference between halfstep and 1/4 step?


Explain easly, im from Norway!


Robert.

[Mongkol]

Try this-->
http://www.rogersmachine.net/encoderinfo.html

[MrWild]

The G540 is expensive, but you are not buying the name. People run the G540 with 48v and their steppers fly. The G540 has extras that the cheap drives don't.

1. The G540 will take up to 50 volts. This is real world volts, not fantasy volts like the cheap 84~~ and 65~~ Chinese boards. Those boards claim 36v but if you use more than 30, they blow up.

2. Mid band resonance damping. The G540 will fly with no mid band problems,. It damps the resonance in the drive. You don't need to add dampeners to your mechanics.

3. micro step morphing. The G540 runs with micro steps at slow speeds for smooth operation, but at high speeds it switches to full step which is actually better for power at fast motor speeds. No worries, the drive reads the micro steps and keeps motor position at all speeds.

4. Motor tuning. G540 has a rheostat so you can tune each motor axis for smoothest operation.

5. The Gecko warranty. If you blow it up, they will exchange it for a new one.

I'm running 385 in oz steppers on a JGRO X/Y axis. I am using 1/2-10 Acme lead screws. With a 48v 7.8A power supply and a Gecko G540, I am hitting 150 inch per minute rapids. My Z is a 425 inoz stepper and it does 180 ipm rapids. All with no lost steps.

You aren't paying for a name, you are buying quality, smoothness, speed, power and great warranty service.

[irving2008]

The motor is losing steps because you are trying to accelerate the load and the screw against their inertia and that requires a level of dynamic torque which you are obviously exceeding - the static torque rating of the motor (1.8Nm) has little bearing on this. Dropping the acceleration will find a point where its not loosing steps, however you should be able to acheive 1000mm/s^2 on that motor at 33v unless your gantry/screw is unusually heavy or long or your axis is binding in some way.

What type and size of leadscrew do you have, what is the weight of your gantry and what linear rails are you using?

[Vegabond]

I have THK Linear Rail Actuator (4) SSR15 (2) 47" Travel on the x-axis. I actually don`t know how heavy the gantry is.

Here is some picture of the screw on the x-axis and the cnc.
The leadscrew on x-axis may be 16mm thick that i remember. the other specification you can see at my photos in the 1st post.

Thanx for the help! ( i need a g540)


Robert.

[irving2008]

well guessing at gantry weight of 30kg (which is probably more than it is) a 16mm screw turning at 15mm/sec (900mm/min) should be able to be accelerated at well over 1000mm/sec^2. At that speed and load you need about 0.5Nm of torque to accelerate without losing steps.

Since you are running at 180rpm on the motor at that speed the issues are not related to midband resonance because you are in the flat portion of the torque-curve - the corner speed of that motor at 33v is 625rpm where it will be generating around 1.3Nm torque.

So if you are losing steps there must be another reason and i wonder if its binding on the slides. Under acceleration that arrangement is going to put a twisting force on the linear carriers. How many SSR LM carriers have you go on each rail and how far apart are they?

[Vegabond]

Hello again, i have 2 carriers on each rail. From the outside to the outside of the carriers it is 135mm on the x-axis.

The y and z axis i can run very faster. (don`t remember how fast...)

Do you mean i need to try halfstep configuration?? Is not the precision goes poorer?

(sorry for my little bad english here ) thanx for the help!


Greetings from Robert.

[slavkok]

With this motors and your machine you should get at least 1000mm/min. As I have similar motors and only 24V I got over 1000mm/min and have heavyer machine with box ways
http://www.cnczone.com/forums/showthread.php?t=84773

There should be some problem. The question is where. Try to lose nut and check how easy you can move the table. If you have spring scale or dynamometer measure it. In my machine I need aprox 100n to move table and after that aprox 60n to sustain movment. With 50kg load on table I need aprox 300n force to start move it after that aprox 200n. So if you have lighter loads (you should have!) then problem can be in computer. You claim that you axis stall at 2400Hz(x4 microstep->600Steps/sec) and for me that's good point to have midband resonance problem. So what is?
Set your machine to 15mm/s (just use jog speed slider) and move machine. If machine move without problem then you set speed just little higher for example to 15.5mm/s .... and repeat that until you get motor stall. And now the question. Just before stall did you hear some strange noise? Vibration? If yes then you have resonance problem. If motors stall without vibration/noise you have some other problem. Did you run lattency test program and you "abused" computer as should. And do it long time! Some computer have problems. For example my computer is unusable in 1'st 10 minutes. I don't know what is wrong but got latency over 1000uS!! (1 milisecond!) after that period all things seems good. Jitter is under 10us and machine happy. I don't know what how/ happens but can be SMI related or maybe just stupid harddrive thermal recalibration. (If I don't abuse computer the bad think doesn't appear!) So check what happens before you purchase costly drivers.

[Vegabond]

Hello, thanx for the answere.

Here is a picture of the latency test, i run some program when the test is running.

I know this is too much, so maby i need a better computer instead?


Greetings from Robert.

[slavkok]

Depends what you want.

If you set base period to 80uS you can still get usable 1500 to 2000 steps per second out of that box. But you ned to be careful on timings. Check on manuals how to set up!. I do not have problem caused by jitter as I use phase drive. On Step/dir scheme you should take jutter into account!. Just search wiki as I know there are nice tutorials how to set up. And someone somewhere on wiki (probably) say's. DON'T GUESS PARAMETERS. Try to calculate it.

You should get at least 1500 steps out of that box. If not then something else is bad.

[irving2008]

Hello again, i have 2 carriers on each rail. From the outside to the outside of the carriers it is 135mm on the x-axis.

The y and z axis i can run very faster. (don`t remember how fast...)

Do you mean i need to try halfstep configuration?? Is not the precision goes poorer?

(sorry for my little bad english here ) thanx for the help!


Greetings from Robert.

I think the fact you can run Y and Z faster does, in my mind, indicate the problem is mechanical. Can you get faster speeds on X with a very slow acceleration setting?

[slavkok]

Can be mechanical or electrical.
As jitter is big the (incorrect)settings may be just on edge. And If X axis has longer cable to driver then this can be issue. (direction before step timing)

Of course can be mechanical too. Maybe stil just to high acceleration and... But in this type the motor usaly stall. If lost steps at begining of move they can run. So timing on reversals is critical on big latency system

[Vegabond]

I haven`t unmount the nut yet, but i have runned stepconf, and here is the maximum.

Velocity : 50mm/s
Acceleration 100mm/s2 (can run more, but i just place it on 100...)

When i use the machine, i run it at 500 accel. and 18mm/s All this on X-axis.

I will try to loose the nut tomorrow, maby i can find out something.

Thanx for the help!


Greetings from Robert.

[irving2008]

well 50mm/s is 3000mm/min and that sounds pretty good to me... which sugegsts to me that as you increase the acceleration the gantry, being top heavy, is twisting and causing the X-axis to bind and therefore you lose steps during accleration and deceleration..

is there an acceleration where the motor starts loosing steps at that 50mm/s^2?

[Vegabond]

I don`t get 50 in velocity for a long time, just some seconds. Im at 46 mm/s with 400mm/s2 in a few minutes.. But i don`t want to run at that, i don`t wan`t to loose step when i mill.

I try to have some liquid/grease (?) on the screws, but it didn`t help.

Tomorrow i will loose the nut, and try to move the gantry manually.


Robert.