Z AXIS ISSUE. Any help would be appreciated!!

[pugzilla]

hey guys i have a weird problem with the vertical X axis. When using mach3 if i manually move the x axis by using the pg up key to go up it will move up but when i let the key go and press it again the movement is reversed and starts to go down. Same thing happens when i use the pg down key or when a print is loaded up. Any reason as to why this may be happening?

* I meant the z axis.

I tested the motor alone and that works. Checked the signal coming from the 3 axis board using an oscilloscope. Noticed that the signal going into one wire on the motor for the z axis wasn't being sent a regular square pulse like the other motors.

Is this a sign of a bad motor pulse due to a chip thats no good anymore on the 3 axis board?
All the other connections work normal.

Pics below of the signal sent to that wire. All other motor signals are fine.:
Wasnt using the bottom channel in the pics.
[

[johnoly99]

pugz, double check you have the motor windings going to the right places. It sounds like you might have the motor coils going to the wrong spot on the connector. It should be (if using our motors and board) from left to right on the connector when viewed from the wire side of it, red, blue, green, black. Red/Blue make one coil and Green/Black make the other. Check it out, and let us know.

[pugzilla]

Double checked the motor windings and they are wired correctly.

[johnoly99]

Hmm, have you done the jumper mod to the 3 axis board? If so, check all the jumpers are seated well. That's usually the case. If you havn't done that yet, def. need to do that. Here's a link

http://reprap.org/wiki/File:Skewfix2_Circuit_Mod.jpg

If you still have a problem, send me an email at our seemecnc gmail and I'll get you a new board sent out. We test all the boards before they ship, but could be a bad board?

[Fastest1]

Though it does appear you are much more electronically knowledgable than I, and also have the proper equipment for testing. My experiences with reversing of direction has always been a wiring issue with the stepper. That being said I wasnt overly impressed with the electronics/driver package. But I do have a few Gecko G540's and a Hobby Cnc Pro driver to compare against, performance wise. For the price and knowledge base the H1 is excellent.

[johnoly99]

Fastest, yes. It's hard to compare really isn't it? Yes, it takes a bit more fiddling to set up the driver boards (although not much, if this is your first experience with cnc or stepper motors, the learning curve is pretty much the same). The problem is, the Gecko costs' as much as the rest of the complete kit! And for that, you get a better piece yes, but is it needed? I've never burnt up my board or easydriver or anything, and have a TON of hours (over 500 easy) on the stock setup. Now, don't get me wrong. Steve and I both agree that we prob. wouldn't use these boards on a mill with a $5K spindle, or stick our fingers near a tool controlled with it, but for the printers, it's awesome.

[pugzilla]

thanks for the suggestions. I did do the mods on the 3 axis board since I'm using the easy driver setup. I double check my opto coupler jumpers. I figured it may be a possibility of a jumper. I checked all the signals from the board to the motors and they were all getting clean square wave pulses with electronically is what I expected except for that one signal on that motor. The funny thing is that the other 3 wires on the motor were receiving clean normal square wave pulses.

*I removed the zener diodes just like in the mod. Was wondering if removing them was necessary as we?

[drew]

I had the pretty much the same problem with my Z axis and it turned out that I had blown the B outputs of the Z TB6560 driver chip. I'd suggest checking the signal at the output of the TB6560.

[johnoly99]

Good point drew. Also guys, remember, it's not good to run a stepper driver chip without a motor connected, or even one coil, either way. The chips tend to burn up if they don't have coils to send the power to. Either way pugz, lemme know and we'll make it right.

[drew]

The datasheet on the TB6560 also says that they're VERY sensitive to static. I figure that's how I blew mine but I really don't know.

[K.Gerard]

Just to be clear about the risks of power on the controller - powering down includes both shutting off the power supply and waiting about 10 seconds for the board to pull all power to ground, and either shutting the PC off or disconnecting the "printer" cable. Make sure that all the LEDs on the controller board go dark before you do anything as the port supplies logic power to the board as long as the PC is on.

The X died on my TB6560 after about a half hour of testing (mostly calibration with Mach). I never could determine the cause. Scope showed nobody home on both A and B at the chip. All of the downstream components checked out OK (but it is a little problematic testing components on the board..) Since I was also having a terrible time finding a PC that could pass Mach's driver test (they all max out at about 5500 pulses/sec and have huge variances) I guess I sort of figured this was the least of my problems and just ordered a 4-Axis board (apparently shipped by slow junk from China). Meantime I'm scrounging MoBos trying to find one that will pass. Failing that I'll have to go to a motion control board (I include RAMPS and Sanguino in this group since they generate their own pulses).

[michaellatif]

(I include RAMPS and Sanguino in this group since they generate their own pulses).

I plan on building (when I get time) a Sanguino board and use my original 3 axis TB6560 board as the driver, as far as I can tell, that should work. I will just us a pololu driver for the extruder.
When you get started post up your results, I would be interested in what you come up with.

[K.Gerard]

Yes, that should work; and the sanguino will give you a clean, fast pulse on the input side; but I don't see why you wouldn't just drive everything through polulus The input side of the TB6560 boards is not well isolated so even though the sanguino is well buffered you might see some feedback or jitter.
I know you're using Repetier, and I'm trying to stick with Mach for now so if I do go with a motion controller I'll have to figure out ModBus. I've seen some folks are running sanguino through ModBus, but I really don't know much about it. I'm sort of waiting for the TCP/IP implementations to work out; I've contributed to some of the programming work on the sockets end in the open source community (one step away from the EMC2 folks) and it's more familiar to me. Either way I should be able to make Mach work with it.
I'll try to keep the group up to date, but meantime I'm still waiting for my junk to reach harbor.

KG

[johnoly99]

Hey guys, John here. Just wanted to drip in. You say sanguino has a fast and clean pulse stream. Doesnt it use the same 16 clock as arduino megas? I thought that thia was the limiting factor with ardy based control, as we can be MUCH faster with mach3.

[pugzilla]

Hey guys thanks for the tips.
It does appear that the output on the chip for the Z-axis burned out during calibration. John what can be done to correct this?

[K.Gerard]

(re-editied to clarify...)

John the Mega uses a 16 Mhz clock and the Sanguino uses a 20Mhz clock - assuming they're using the the main RC to seed an oscillator it may still be a limitation; I'm just diving into the RepRap electronics to see if I can get away from hardware problems on my hosts; and I don't know yet how they implement their pulse generators. My issue is that I'd like to see 25 Khz out of my port but I haven't gotten there yet (5.5Khz) - and all the mobs and ports I've tried are so unstable that the output is pretty much unusable. Mach is just software, so the problem is probably in the DSPs everyone is using to drive parallel ports these days (and the way the Mach pulse generator code works I suppose - but that's just a black box to me). Anyway, this is really off-topic for poor Pugzilla - I wish I had a better answer for him.

[johnoly99]

Quick thought, you arent using 64 bit OS are you?

[drew]

Hey guys thanks for the tips.
It does appear that the output on the chip for the Z-axis burned out during calibration. John what can be done to correct this?

For what it's worth I bought a replacement TB6560 CHIP from Mouser.com and replaced the bad one on my TB6560 BOARD. Not too bad a job if you're up to it and surprissingly it worked.

[K.Gerard]

I've been testing the PCs with a 32-bit version of Windows XP Professional. I'm an RTFM kind of guy

Having written hardware drivers with assembler in the dim past I can understand why the Mach guys don't relish the idea of rewriting their driver for 64-bit; but I get the impression from their forums that they're downright cranky about it.

I finally dragged up an HP XM600 workstation that I thought was too old - and it performs perfectly. Spent $6000 on it in the late 90's, and I guess it was worth it! My 4 axis board also arrived direct from China (covered with greasy fingerprints and with rust on the D shells); if it works - fine, if not I may take Drew's suggestion and just fix the old 3 axis.

[Polygonhell]

I think part of the issue is the Mach driver isn't signed, they'd have to start paying for signing if they were to produce a 64 bit driver, since 64bit versions of Win7 and later won't load unsigned drivers unless booted in test mode.
I run Mach to control my Mill and I'm just using the parallel port on an Atom board running XP pro 32bit, it has been more than adequate though I have an ethernet smoothstepper here which I keep meaning to install (mostly for the extra IO pins).
I'm using RAMPS for the printer, because I just didn't want another single purpose computer in my office, I use my laptop to upload the print to SDRamps, start the print and unplug the computer.