[Jack000]

this isn't for a cnc machine, but pretty similar..

So for our final project in undergrad engineering, we're building a 3d bike simulator. The idea is that you strap your bike to a lift mechanism that emulates the outdoors - complete with hills, bumps and water, for people who can't go outside for whatever reason.

here's some of my design sketches:


and a 3d environment that I built:


we're having some trouble selecting the motors and screws.. the problem is that we need a lot of speed to emulate hills and such, I'm thinking 30cm/second. At those speeds we either need a strong motor with a high-lead ball screw, or a stepper motor that can hold its torque over 10 000 pulses per second.

I found this motor torque graph from keling: http://www.kelinginc.net/KL23H284-35-4BT.pdf

but I'm a bit suspicious of its claims since most other torque curves I've seen has one less zero on the bottom.

any ideas?

[jsheerin]

The driver's set to 1/8 stepping, so 80000 pulses would get you 10000 full steps from the motor I believe.

What is your budget and how much force do you need at 700ipm?

[irving2008]

this isn't for a cnc machine, but pretty similar..

So for our final project in undergrad engineering, we're building a 3d bike simulator. The idea is that you strap your bike to a lift mechanism that emulates the outdoors - complete with hills, bumps and water, for people who can't go outside for whatever reason.

here's some of my design sketches:


and a 3d environment that I built:


we're having some trouble selecting the motors and screws.. the problem is that we need a lot of speed to emulate hills and such, I'm thinking 30cm/second. At those speeds we either need a strong motor with a high-lead ball screw, or a stepper motor that can hold its torque over 10 000 pulses per second.

I found this motor torque graph from keling: http://www.kelinginc.net/KL23H284-35-4BT.pdf

but I'm a bit suspicious of its claims since most other torque curves I've seen has one less zero on the bottom.

any ideas?

Funnily enough I have been working on the idea for a bike trainer with a similar capability (but not the fancy graphics!), although i was thinking of a hydraulic ram. So far I have concentrated on designing and acquiring parts for the roller mechanism, the rear to front drive drive and the variable load, most of which are mapped out although I have not yet cut any metal.

How do you get to 30cm/sec and what forces/torque have you estimated? Maybe your hills are a lot steeper than the ones I ride up

[Jack000]

30cm is the approximate length of travel, I just ball-parked that figure - it seems that to get any good simulation of rough roads it would have to be pretty fast. We're hoping that it would be fast enough to simulate bumpy ground like gravel and rocks.

here are my numbers:
ball screw lead (linear distance per turn): 5mm
max velocity: 30cm/second
step angle per pulse: 1.8
load inertia: 50kg

to travel 30 cm, the screw needs to turn 60 times, which corresponds to 60*(360/1.8) = 12 000 "full step pulses"

the load will be approximately 50kg, with springs to help it fight gravity. Using a 5mm/turn ball screw and 1 Nm motor I should be able to get around 1000 N of force, or 2Gs of acceleration at 30cm/second. This is provided that the motor can actually deliver 1Nm of torque at 30cm/s.

I'm getting the feeling though that steppers just don't go at that speed.

are there any good full-step drivers out there? we'd like to keep things as cheap as possible, and precision isn't really an issue in our case.

[jsheerin]

You could look at Keling's 960oz*in N34 frame stepper with a Gecko drive and geared up with the screw spinning about 5.5x faster than the motor.

Or you could go with one of Keling's DC servo's. The larger N34 size would work pretty well if you geared it up just slightly and you'd have excess force in reserve for momentary usage if you ponied up for a 20A power supply. I'd go with a Gecko G320X for that. This is what I'm looking at for my router's z axes, but I'll probably limit acceleration to 1g to have more force left over for cutting.

[Jack000]

would a belt or chain drive system work better? We don't really need the mechanical advantage that comes with a screw. I'm just worried that a belt wouldn't be able to handle the forces involved.

[ger21]

A belt would be fine, but you'll need a gear reduction. I think you'll find that you do need the mechanical advantage.

[Jack000]

a little update.

We've bought THK slides, a 5mm lead ball screw and a bunch of unistrut, as well as all the motors and electrical components. The problem right now is how to attach the motor shaft to the ball screw. The motor is 1/4" and the ball screw shaft is 8mm, and I can't find any couplers for it.

would a compression type coupler still work ok if drilled out to 8mm? I'm thinking of buying the flexible aluminum coupler on this page and just drilling it out: http://www.kelinginc.net/CNCPackage.html

here are my cad designs and a photo of the prototype:


bike tilt mechanism (little cylinders are rubber skate bushings):



that's my foot!

[YZF]

Shouldn't this be your homework? :-)

You need to figure out the speed (e.g. rpm, steps/s) and torque (N-m) requirements and then browse different motor vendors for the right motor. Look at the speed/torque graph and use a safety factor of 2. Any gearing changes you make will move you on that curve. You're trading off speed and force. I would pick up a point around the "knee" in the step motor graph (e.g. 10,000pps in your motor), figure out the gearing needed to get the linear speed, and see if you have your required torque margin. You can use a gear/belt reduction between the motor and the screw at that point (for example).

The primary mechanism that is reducing the torque as you go faster is back emf. One way to overcome this is to use a higher voltage supply driver. Most stepper vendors will show you torque curves at different supply voltages. You'll find that you need either higher current or higher voltage to get that higher torque (surprise).

This is a good reference on stepper motors:
http://www.cs.uiowa.edu/~jones/step/

Very generally speaking a 3-phase brushless servo motor will have better high speed performance characteristics than a stepper motor.

[Jack000]

the thread title is a bit misleading... but I don't know how to change it.

actually we already bought the motor. It's a nema 23 stepper from keling. It turns out that the smaller motor has a shallower torque curve, which coupled with a 5mm screw gives us about 10cm/s. It's not as fast as I'd have liked, but larger motors and drivers would have blown up our budget, and we had to make some compromises.

Right now the problem is couplers!

[Jack000]

ok.. another question.

I got the 48v power supply from keling, but it turns out to be +36v and -12v from earth ground. I asked the gecko guys if this is fine for use with the g251, and they said:

This power supply may cause shorting to the G251. You need a different 48V power supply.

is this true? Has anyone else used these supplies with the gecko drives?

[YZF]

ok.. another question.

I got the 48v power supply from keling, but it turns out to be +36v and -12v from earth ground. I asked the gecko guys if this is fine for use with the g251, and they said:



is this true? Has anyone else used these supplies with the gecko drives?

I'm not familiar with this particular drive but if all the inputs are optically isolated you should be able to use the power supply. If the inputs are not isolated and the input grounds are connected to the DC bus voltage grounds then you'll be shorting -12V with ground. So what you don't want to do is connect two inputs that are connected internally to -12V and GND at the same time. That would be like shorting your car battery.

Also look at the current ratings for the -12V and 36V. It is quite common to see supplies where the -12V is rated much lower than the positive 36V. If this is the case the 48V you get between -12V and +36V will be limited by the current rating on the -12V.

You may want to consider running off 36V if the drive supports it - check the manual. Unfortunately it will further reduce your maximum speed.

[Jack000]

there is a separate signal ground on the g251, but I'm not sure if that's connected internally to the power ground.

The keling website said "KL-350-48 48V/7.3A" so I assume I can get 7.3A from it total.

[YZF]

there is a separate signal ground on the g251, but I'm not sure if that's connected internally to the power ground.

The keling website said "KL-350-48 48V/7.3A" so I assume I can get 7.3A from it total.

I took a look at the g251 and it doesn't look like it's optically isolated. I couldn't find your exact supply on keling's web site and their specs for the others seem fairly sketchy. I can't say for sure but it sounds like it won't work.

[YZF]

ok.. another question.
I got the 48v power supply from keling, but it turns out to be +36v and -12v

By the way, how did you figure out it's -12v and +36V? It may simply be floating relative to the earth ground. Do you have three outputs? (0, -12V and 48V?)

[automationtechinc]

I took a look at the g251 and it doesn't look like it's optically isolated. I couldn't find your exact supply on keling's web site and their specs for the others seem fairly sketchy. I can't say for sure but it sounds like it won't work.

You can use the KL-350-48 48V/7.3A
with G251 driver.
http://www.kelinginc.net/SwitchingPowerSupply.html

[YZF]

Ah. So I guess it *is* a 48V supply. Not a 36V/-12V supply. It's really hard to figure out the specs for this model on your web site.

[Jack000]

I measured it relative to the earth pin on the power supply.

The problem is that our logic stuff gives 0-5v relative to earth, and the power and signal grounds on the g251 are connected.

Would we have to put our logic circuits on the same power supply then, so everything shares a common ground?

[YZF]

I measured it relative to the earth pin on the power supply.

The problem is that our logic stuff gives 0-5v relative to earth, and the power and signal grounds on the g251 are connected.

Would we have to put our logic circuits on the same power supply then, so everything shares a common ground?

The problem with measuring to the "earth" pin is that there could be a very large series resistance and you wouldn't see it. It's not necessarily meaningful. V- is most likely not connected to the earth in the power supply, i.e. the DC supply is floating.

[Jack000]

just thought I'd share an update on our project.

here's a video (I'm the chinese guy at the end of the video):



and some screenshots:




we ended up just tying the power supply neutral to earth ground. I think we were supposed to do that in the first place.

we also blew a g251, rather spectacularly - there was a bright flash and it burned up pretty bad. Had to disassemble the control box for my cnc so we can get a replacement in time for the demo.. (nuts)

does anyone know if steppers can lose steps silently (do they always make a screeching sound?). We're seeing some error accumulation over time, but I'm not sure if it's due to the stepper stalling (it sounds fine) or our control system not working right.