Need Help!- Problem Losing Steps = (

[Brenck]

Hi everyone,

I have a small cnc router that I built, I use a unipolar driver. Powersupply of 40v. Everything was working just fine but now something is making it loose steps. Its loosing on Z a Y I dont think I noticed X loosing steps. I dont even know if its acctually losing steps or just going crazy, because like I had a 3d wood cut that I tried to cut and at the begining the machine did something that I believe it wasnt lost steps it seems that it took more steps then it was suppose to.... Its not mechanical because it was working before. With no problem. I'm gonna post pictures if someone have a clue tell me...

Thanks

[oko_didovo]

Originally Posted by Brenck Hi everyone, I have a small cnc router that I built, I use a unipolar driver. Powersupply of 40v. Everything was working just fine but now something is making it loose steps. Its loosing on Z a Y I dont think I noticed X loosing steps. I dont even know if its acctually losing steps or just going crazy, because like I had a 3d wood cut that I tried to cut and at the begining the machine did something that I believe it wasnt lost steps it seems that it took more steps then it was suppose to.... Its not mechanical because it was working before. With no problem. I'm gonna post pictures if someone have a clue tell me... Thanks

Hi,

Check how much current consumes every motor than add every amount to total sum. The total amount must be less than maximum current your power supply can deliver.

In most cases it is the only cause of losing steps.

I hope it will help.

[apache405]

Is the machine skipping steps under load only? Or is the machine skipping steps unloaded as well?

[Roger Hamlett]

Originally Posted by apache405 Is the machine skipping steps under load only? Or is the machine skipping steps unloaded as well?

Yes. This is the big question.
It is important to undertsand, that a system that 'works perfectly', just moving the head around without load, may well miss steps once the cutter is actually engaged in a material, as more torque is needed from the steppers. It also gets more complex, since on most drivers, the excess energy in the stepper coils, is 'fed back' into the power rails, so the current needed from the supply, will rise as load is applied, and the rail voltage may fall, making matters worse.

So you have two potential prroblems:
1) You may just be stepping too fast, for the load.
2) The controller and steppers may actually be able to cope with the step rate at this load, but can the supply deliver the current needed (first reply).

Further comment, Z, is typically the axis with the greatest load on a simple system. It is common to have to move this axis slower, than the X and Y especially when engaging/disengaging from a cut. If your routing program (for example), generates ramps in Y, as Z cuts are generated, then this would explain the missed steps being on these two axes, particularly if the real problem was '2'.

Best Wishes

[Brenck]

My transformer is a 31v 5A. Its more then enough for my 3 3A steppers. The steppers are strong enough, to push the way I setted up to, because it was working perfectly before at even higher speeds, working with aluminium and the object that it lost steps was on wood... The problem is not mechanical.

It dont loose that much steps that I can see when running without a work to do, its like if i make it draw a circle It will go perfectly but not because its not loosing steps without load but because it didnt have the chance or time to loose.. lol...

The only thing i did after it was working perfectly it was to wrap all the cables with thoose spiral cable organizer ( seen in picture )...

I wonder if that has something to deal with it...

I dunno if it can be ressonance, but i'm runnig on half step already...

[oko_didovo]

Originally Posted by Brenck My transformer is a 31v 5A. Its more then enough for my 3 3A steppers. The steppers are strong enough, to push the way I setted up to, because it was working perfectly before at even higher speeds, working with aluminium and the object that it lost steps was on wood... The problem is not mechanical. It dont loose that much steps that I can see when running without a work to do, its like if i make it draw a circle It will go perfectly but not because its not loosing steps without load but because it didnt have the chance or time to loose.. lol... The only thing i did after it was working perfectly it was to wrap all the cables with thoose spiral cable organizer ( seen in picture )... I wonder if that has something to deal with it... I dunno if it can be ressonance, but i'm runnig on half step already...

If all of your 3 steppers running at the same time at the max. load, (3 x 3A = 9A), your tranformer isn't enough for that purpose. In most cases, only two steppers running at the same time, (X-os, Y-os), so your transformer have to deliver 6A of current or more (9A) for a smooth work.

[apache405]

Was the last job you did in metal or another conductive material? Have you tried using some compressed air to clean out the box with the electronics in it? If a chip of metal is bridging a power rail or some other critical system line, then that might explain your problem. Also, have you tried removing the spiral wrap from the cabling?

[CNCRouterm]

Some time last century, when I ran an open loop stepper CNC router, there were several areas that tended to cause position errors.
1.) Accel/Deccel too high
2.) Overspeed, 600ipm on a really good day, but not reliable, 300ipm was about it for rapids.
3.) High resistance in the cut, heavy feeds/speeds.

Material is secondary, actual cutting forces are dependent upon tool geometry, chip load, tool size, depth of cut, full or partial (width of) cut etc. Cutting 1/8" aluminum on my current (AC servo 16000lb) machine is limited not by the machine but by the tools (usually a 3/8" single flute) chip load and efficient rpms. I work the machine harder running 1" particle board or MDF at 1200 ipm than 1/8" alum at 160 ipm.
Also, consider the possibility of mechanical binding, maybe something got "tweaked" a bit while you were changing things around. The old Digital router from 'way back' was built on an 8' wide by 18' long (yes, feet, not inches) table, which unwisely, was made primarily of mdf. Changes in environment, ie. humidity and temperature cause the table to move, which change the mechanical resistance, which is why we had to program the acc/dec, feed etc. parameters conservatively. Your project probably doesn't suffer such an extreme handicap as our mdf table, but consider the underlying principle as it might apply.

[shipmodeller]

Your problem can indeed be with the wrapping the cables. The stepper cables throw off a lot of rfi when switching. This can feed back into the stepping system and add steps or lose steps.

The best thing to ensure that doesn't happen is to isolate the power drive cables from the signal cables. The signal cables, those that tell the drivers to step, should be shielded cables and well grounded.

The drive cables should be of sufficient thickness to handle the current x 2. I used to have a similar problem till I redone my cables and grounded my system better.

Hopefully, this is all it is.

[Tom Kay]

Hi Brenck;

It's been about a century for me too, since I last played with my steppers, but I recall something that Maris (Mr. Gecko) once told me when I was having problems losing and gaining steps. Also, the fact that you have wrapped all your cables in spiral wrap means that they are close together, and that could, worst case, add problems.

Maris told me to use shielded cable with a braided drain wire to connect the printer port of the computer to the stepper drivers. Not sure if you use a printer port, or if everything's output from the usb now. (see, it's been a while).

I also connected one (and only one) end of the braided drain wire to the chassis ground of the computer. Maris said to connect only one end of the drain wire.

But then he asked me about the conductors in all the shielded cables, and I mentioned that I had used both of the wires in each cable to carry signals. He said "No, don't do that." and what he meant was that for each cable use only one conductor, and leave any others unused. They send garbage signals back and forth to each other, since they lie very close together with little insulation blocking unwanted signals. I think I was actually gaining steps, but sometimes it was hard to tell if I was gaining or losing.

I never actually had my motors hooked up to a mill or router, but I bench tested them under zero load. Zero load could also be part of my problem. As I commanded the system to simulate a circle cut, I watched the x motor, then the y motor. At the "quadrants" where one motor would really slow down and change directions, it would run rough, adding or losing steps.

So, first, check your main signal cables from the computer to the drivers. Is each cable, regardless of the number of available conductors, carrying only one signal? Is each grounding braided wire soldered to some sort of ground, and only at one end? Last, try taking the spiral wrap off your cables and see if that alone fixes the problem. If it does, it tells me that possibly some sort of communication was happening through the signal cables, into other signal cables.

If all of this doesn't work, then find a genie lamp, rub it, and hopefully Maris will pop out, since he really eats this stuff for breakfast, and excretes very useful knowledge of stepper motors and drivers. He da man!

Let us know how it all works out, Okie Dokie? I wish I'd completed my system, either a mill or a router, but one day ....

Cheers, Tom.

[lclore]

Principal reasons steppers "loose steps", aside from electrical integrity, are changing load or too rapid accel/decel for the existing load. It would be worth checking the linear rails/bearings for alignment (load change from binding). If no binding, then lower the accel/decel rates and see if problem disappears. If it does, gradually increase the accel/decel rate until positional integrity goes away to define your systems motion profile thresholds.

[gavin53]

One final option, especially for those reading this and contemplating a new system. In addition to slowing the acceleration (ramp up to speed), consider switching to bipolar drive of your steppers. This will double the amount of wire used in the motors (unipolar only uses half the coils at a time) and thus increase the torque available. Hardly anyone uses unipolar steppers any more, and bypolar stepper ICs (H-bridge) have gotten really cheap.