[This discussion of Unipolar vs Bipolar has bee split off from this driver thread: viewtopic.php?p=22815 ]

Kwok, on my PhaltPrinter 3 I opted out of the supplied Easy-CNC board and will use instead the TB6560 based controller, like this one http://cgi.ebay.com/CNC-3-Axis-Stepper- ... 3890wt_969.
They are rated 3A and peak 3.5A. My steppers are rated 286oz/in and will be able to give all of it (not that it is needed with the printer, though ).

 

Yoram,

There are supposed to be lots of problems with that board. For example: http://www.cnczone.com/forums/general_e ... dated.html

 

Thanks, Kwok, I guess it's time to read up. :roll:

Edit: After reading the thread on the CNC Zone it's back to the drawing board. I guess I might as well tell Mark that I will take it all after-all !

 

Hey guys, just thought I would throw in my tid bit of knowledge. It really is motor dependent. For instance, if you look at the performance graphs for the vexta p266-02a, the steppers mark is/was selling, you will see the benefit of bi verses uni is highly voltage dependent. In fact, at 24vDc uni out performs bi by a lot after about 100 rpm. They don't show 36 volt, but I assume it's prolly between 24 and 48. That is about the same performance as uni on 24 volts. Now if you go with a 48 volt supply, you will see performance gains.
I am running uni drives I made with a 12 volt supply (regular pc pwr supply) with no problem at all. If I tried to run bi with the same supply, I would get terrible performance.

Bottom line? It is a whole system and needs to be addressed that way

 

After reading Bert's post I've checked the graph for the motor. I wonder if anyone know at what RPM these motors are typically running in the PP.

 

It depends on which PhlatPrinter and which axis. I did this real quick, so for what it is worth.

The MK1 had, I think 24 threads per inch on the Y. So 24 rpm to get 1" of travel/minute. So to get 40 ipm, the stepper is going 24x40= 960rpm.

The Y axis on the MK2, has something like a 1" pulley. Circumference would give 3.14"/rev. So to do 100 ipm, the motors are turning 100/3.14= 32rpm.

-Kwok

 

According to this they are better, then, running bipolar. I have some bigger motors and I couldn't find the graph for them but I can imagine it to be similar to the ones that come with the PP 3.

 

Well, if you are just looking at this one thing, it seems unipolar is better for the screw drive example. And bipolar is better for the belt example. But there must be more to it than that.

 

Hmm, I been thinking about this and it still not clear to me, if the conclusions I just gave is correct! Basically, the conclusion is that the graph showed how much torque the motor would produce for a given rpm. But that's just the motor. You haven't taken into account how much torque is needed to move the axis, itself.

 

Like everything else, it's all a compromise. The higher the motor rpm per ipm is the better the precision that we get but at the higher rpm the startup takes more time which is not so important on long cuts but on short cuts, like most of our cuts are, it does play a role.

Compromises, compromises. Decisions, decisions.

 

That's why unipolar might be interesting.
The engine torque at low speed are high and there is no problem if there is a small loss.
The problem is the loss of torque at higher speed and the engine stalls, just when the inertia has greater influence

In a CNC always want a higher final speed, this is achieved with sufficient torque, even with higher speed motor steep.

Another thing, when we look at the torque curve / spin, she was drawn to a particular drive and tension.
With another drive and another tension, the curve is different.

Hugs

 

This is all very fun to read. Kwok, you hit the nail on the head. How much torque do we NEED? I get that part of this thread is about drives in general, for any CNC machine. But to bring it in to something close to home, lets talk about phlatprinters. First of all, how much torque you needed is a function of several things. The biggest factor here is how fast you want to move AND how fast you want your accel/decel. I can't speak about the MK3, I have no first hand knowledge. Same with MK1

I have a MK2. on that machine the Y has to move the saddle (gantry) 26" max. My happy speed is 200ipm. I have read posts with much higer speeds, but I am happy at 200. I haven't weighed it, but lets assume the saddle weighs in at 5 lbs. because it rides on bearings, we will assume a startup friction of pretty much nothing. If my pully is 1" dia, that gives me 1/2" lever arm for the torque to act with. so how much torque do I need? Almost none, cuz there is nothing to react against movement but inertia. I don't know how to calculate this, but if I pull on it with a force gague, it is less than 2 oz to start it moving. I know I am getting at least 24 oz of torque at a full stall(this is not a good measurement, I don't have the equipment to actually measure the torque while running). This number comes from connecting a force gague to the slide and asking the motor to move. As I miss steps, I loose torque. So this is a worse case number! So I have at least a 12:1 safety factor to get it rolling.

Now what do I need to keep it rolling? This is a factor of what I am cutting. Using an endmill to cut foam, there is again almost no force there. After all, the 1/16 or 1/32 mills will snap! If you guys want, I will break one to measure just what it takes, but I can assure you it isn't much.

So now we are at stopping. Well if I just command it to stop, with no decel, How much torque does that take? Again, I don't really know, but it is less than my uni drive can handle.

So the bottom line for me is this: I don't need all the torque available at low end speed, I am VERY happy using a 12V supply (afterall, I already have it in the PC). I can tell you I have the full 127OZin of holding torque as specified in the chart. Thats with uni drive and 12 volts.

In fact, at 200 rpm, for 24 volt operation, the bi drive drops off from the uni. But I think thats pretty fast (3.1415*200=628 ipm). Up till that point, I am getting between 127 and ~85 oz-in torque. Thats Almost 6 pounds!! Way more than is needed.

So I have less expensive drives, and no money for power supply.

Sorry for the book. It just seems like everyone is throwing out stuff that is all true, but maybe not relevant? I am the guy who likes to look at what do we need, not what can we have. In a perfect world, we would use AC drives with encoders!

 

In a perfect world I would want more than that.

In the real world, servos are expensive and not everyone can have, there are many types of machines, and milling a 3D file can take many hours, any gain in time is welcome.

The Great achievement is to do a lot with little

Hugs

 

I agree!

This is a machine (PP) designed to be cost effective for folks to get in to this as a hobby. I am just throwing out ideas how to keep it that way. If we really want to go hog wild, lets throw on a 96Volt power supply @10A per stepper. But as I am sure you will agree, for this application that would be a huge waste of money. and again, if I needed that kind of muscle, encoders aren't that much. so servo would then take over MY design choice.

 

Oh, and we left out belt stretch! If we have too much torque and try to accel/decel too fast, the belts will give just a bit, not so much as you would see, but it may show up in the surface finish.

Please don't take any of this the wrong way, I am not at all opposed to bi drives, but there are lots of folks on here who may be trying to make decisions and may not have such knowledge as we do for making them. Then they may spend more money than needed.

 

NOT TAKINIG IT THE WRONG WAY??? I'm taking it the wrong way!!! Where were you before I bought my stuff???

 

Sorry Tiger, if it makes you feel any better, I own an easy-cnc bi-polar drive board. It wasn't until I blew the Y chip(due to stupid on my part) that I decided to start looking harder. After reading other peoples troubles, and not wanting to pay $10.00 just for a chip that comes loose in its socket, I decided to make my own.

I started by first looking at bi chips. they are out there, and not too pricy either. The problem for me was all of them I found have the heat path through the bottom. So that ment not only did I have to deal with smd devices, but also make a good thermal connection under the chip. Thats what drove me to start looking at uni drives.

All the above info is a result of me trying to decide what I actually needed VS what I thought I needed. I also wanted to get rid of the extra pwr supply. So its win win for me. I do still have a use for my eaasy board. I am going to figure a way to measure the torque while running at some point. When I do, I will do plenty of testing so I can have some experience with real numbers for my own graph.

BTW, they do run a bit warmer, but the shop is cold in the winter, so maybe free heat?

 

Mark, the info that you've posted I knew from before and was what made me want to go only with a bipolar driver. The info doesn't touch the subject of torque loss at higher rpm, something I was not aware of until it was brought up here.

 

That's some very good info from Mariss! I wonder what we should use as a supply voltage value for chopping drivers, like we use?

 

Would 'Series' than be bipolar and 'Half-coil' unipolar?

 

Yes, and here's why:

Unipolar, has to be 'half-coil' because that how it reverses the current direction. The center tap is always on and so you only need a transistor on each end (total 2 per axis) of the full coil. Then you switch power from one half coil to the other half coil. The direction of current will be different for each half, just because power is sourced from the center. So you still need the full coil, but you only use half of it at any given time.

Bipolar, on the other hand, uses a full bridge (4 transistor per axis). It can simply switch the direction of the current. So you can use the full coil, all the time. With bipolar, you can connect the half coils in "series" or "parallel".

You need 8 wires to use "parallel", so our stock motors can only be connected in "series" bipolar.

Here is a picture I found, from http://www.mechmate.com/forums/showthread.php?t=256 , explaining the wiring:
Gargoyle.cnc

Oh, and I just noticed it's also labeled in that graph pic Mark posted above.

-Kwok Attached files