Got my harness rigged up this morning, but ironing out some issues at the moment. Busted out the oscilloscope to verify my pinouts are the same as yours, and they are with one exception:

 My X step pin is 21 (vs. 20), and the Y step pin is 20 (vs. 21). This is easily fixed by swapping the outputs in EMC.

 I'm getting the proper direction outputs to the control box, but both X and Y are fixed in one direction, and move in that direction regardless of what direction you toggle it to go. Z works fine. Will update when I have a solution.

 Cheers,

Joel

I've changed the diagram to correct the error with the X and Y step pins.  Good catch.

Got my harness working! Swapped around the 20/21 pins so that I can use the Sherline pinout without modification. This fixed my direction issue as well, since before it was swapping direction on the axis that wasn't running.

I took a different approach on the E-stop and mill enable pins. On my machine, pin 22 is the mill enable. I wired pin 1 on the male to pin 22 on the female, and assigned pin 1 as an amplifier enable pin in EMC. It works out that when you select the 'Toggle Machine On' button in EMC/Axis, the machine turns on; hitting the software e-stop turns the machine back off.

I also wired up the e-stop on the machine, from pin 10 on the male to pin 23 on the female, and assigning pin 10 in EMC as an e-stop in. Tried it out and seemed to work great for me. No issues with the clock pulsing (I had also noticed the clock pulse on pin 17 when Z was being jogged, didn't think much of it till I saw your post).

Only other change is my harness still uses pin 25 on the male to pin 7 on female for ground like your original one did.

Keep up the good work on this!

-Joel

SAFETY ISSUE:

If you wired your cable like me, with the mill enable pin wired to some output from the parallel port, there is a serious issue. When the computer is off or booting up, power is applied to the mill enable pin (and spindle) if the controller box is turned on.

From what I can tell,  pin 25 (gnd) shorts to all the pins when power is applied somewhere.

Be careful for now with this. Will post new info when I figure out a solution.

 

-Joel

Joel,

I encountered this issue with my first harness.  In that harness I had tied pins 17 & 22 as well as 24 & 25 together.  As you have seen, this causes the mill to "come alive" when it shouldn't.  My solution is to run all signals from the PC to the controller, without any loopbacks.  This is incorporated in the design above (version 2.0).  

-- Shannon

Shannon,

I've got no loops in my harness... pin 1 becomes energized on the parallel port side when power is applied to one of the other ports (not sure which one). This only occurs when the computer is off, once linux is up and running the problem is solved.

Fairly confused at this point, so let me know if you've got any insight (some transistor is activated??).

-Joel

Joel,

Very strange.  Evidently your parallel port ends up in a different state than mine does.  I may well end up making a PCB board to prevent this condition and enable reliable reading of the switches.  I went into this expecting to make a AVR based circuit to handle the stepper timing.  Couldn't believe my luck when the parallel port was up to the task.  How ironic that I may have to rely on a MPU to sort out the handshaking.  

If I design such a circuit I need to first crack control of the spindle speed.  Such a circuit would need to be conprehensive to be worth doing.

A back of the napkin drawing would include IO for the following:

1. Cover open
2. Emergency Stop
3. X / Y / Z limits, one for each direction
4. X / Y / Z home
5. Manual / Computer spindle control

Ideally it would interface with three parallel ports, or scale back gracefully for two and one port configurations.  Thoughts on this?

I've moved the cable to a newer computer, and my runaway mill issue is gone. The old parallel port seems to be the issue in that equation, like you said.

You future plans look good, but I think you might be able to implement it all with a single parallel port. If the X/Y/Z switches are all wired in parallel, you can minimize it down to a single input port. If the program hit a switch, the mill will stop as desired and the operator can then determine from where it stopped which axis was in question. For homing, I think EMC can run a simple homing script that moves one axis at a time until an input is tripped, in this case your single limit input. The safety stops such as cover open and e-stop could be combined into one as well if you were short on inputs.

That being said, additional parallel ports will definelty add additional capability for future items (5 axis plus a robot, anyone?).

Looks like a good project that not short on challenges. By the end you'll have a setup far superior to the initial setup, and for pennies on the dollar compared to what LMC would have charged you for it.

Cheers,

-Joel

Joel,

I took your advice and wired combination limit + home switches on the mill.  Everything is working nearly perfectly.  The major unresolved issue is filtering out the pulses on estop and the cover open switches. 

I have advice on how to filter this out using a cap.  I had tried using a cap as a noise filter, but wired it in series.  I should have wired it parallel to ground.  

The other issue involves the "runaway spindle" when the computer is on, but EMC isn't running.  You saw this behavior on a paticular parallel port, and not another.  I have seen this issue once.  Last night my spindle came on when EMC wasn't running.  

All that should be needed is an XOR gate, the parallel port pins seem to all be at the same / nearly the same voltage.  The XOR gate will only work if two pins are NOT at the same voltage.  This should fix the problem.

Once I iron these remaining issues out I'll make a few PCB boards incorporating all of these features.  I'll also post the schematics here.  

If anyone wants a board, this is the time to let me know.

Email me: shannon dot haworth@gmail dot com.

 

I have now just got the mill working however I do not know the stepper timing, gearing, or any other info needed to operate the machine properly. The steppers seem to be taking a step but then waiting for a period. If anyone is prepared to share this information I can be contacted at rexoregan20@yahoo dot com dot au or on this thread http://www.linuxcnc.org/component/option,com_kunena/Itemid,20/func,view/catid,16/id,3501/lang,english/.

 

Thanks Rex.

Rex,

I will e-mail you.

Cheers,

Shannon

 

Ok now I have the steppers moving more reliably however they make a lot of noise and are not moving smoothly.

 

Any ideas?

Could the fact that I have just cut the cable be the problem? I have a spare so if it is that shouldn't be too hard to fix 

Rex,

There are many reasons why you might not have smooth stepper movement.  Thse incldue:

* Trying to move the steppers too fast (I generally stick to 30 ipm or below).

* Mechanical resistance at the leadscrews.

* Check the wire connections on the stepper.  I had some pull loose and it wasn't obvious.

* Incorrect settings in emc.  (Note the sample settings are for inches, you may need to convert to metric).

Several users have reported success.  Others have used a cut-off cable as well.  I doubt this is the issue.

Keep us posted,

Shannon

What is the revision of the board inside your controller box? could that be the problem? mine have no guard on the fan at the back...

 

Cheers, Rex 

Hey Shannon,

Thanks for posting this "GOLD" info.

One quick question, do you know if this cable will work with the spectralight lathe under this pin configuration?

It does in fact look to have the same control box as the mills.

As a Lightmachines newbie, any info on this matter would be most helpful

Thanks,

Jim jensen 

wow, no input from anyone....what a world,what a world

Jim,

 

Sorry.  There are not a lot of comments left here, so I hardly check.   These plans should work on the lathe also.  Have you tried it?

 

Shannon

 

I have the Spectralight driver box running from Mach3 and a Pentium 4 desktop parallel port.  My cable wiring is almost identical to Shannon's above.  I did add a 3-component charge-pump circuit in the cable so the enable signal to pin 22 would only be active when the program was running.  Made pin 1 a charge-pump output rather than an enable.

In order to get the Spectralight box to respond to the Mach3 5-15 microsecond-long step pulses, I had to add three pullup resistors to the opto-couplers inside the box.  I don't know if that is necessary on all of the Spectralight boxes, but it was on mine.  I also added a pullup to another pin to supply +5 to the charge-pump circuit.  I reverse-engineered almost all of the Spectralight box to a logic/wiring diagram.  The cable diagram is included.  My drawing is in MS Visio, but I can save and send it as a .jpg.  Email me at heyvan@verizon.net  I'd rather email it than post it, because it is still a work in progress

The Lathe will run just like the mill, but direction signal levels might be reversed.  For the Lathe, only the 15-pin and the spindle power get connected.  The 9-pin cable from the Lathe DOES NOT connect to the box - at least I don't believe so.  It is for spindle speed-control only, and it is intended to connect to the special Spectralight controller.  I intend to reverse-engineer spindle speed control on the Lathe eventually, but it might be a while.

My E-stop mushroom button works, but I don't have the door and limit switches working yet.

Thanks for your comments on the lathe s/u.

I went ahead and gave up after hours of frustration.

Iam going the hobby cnc route with driver board and stronger steppers.Far too much time lost on this adapter cable...hope it works for everyone else. 

Who needs a black control box??  

Thanks all,

 jim 

might as well remove all my comments then...sorry to step on your feet, sissy.

Jim,

Good luck with the hobby cnc board.  I've looked long and hard at that board and it looks like a great solution.  If I was going the upgrade route I would also upgrade the steppers while I was at it.  I can't justify the expense given that the mill is one of many hobbies.  I would love to know how the project turns out. 

Sorry the cable didn't work for you.  

Shannon

Thank you for posting this info.  I have the Spectralight mill and even an old controller card box.  I built the harness and connected dirctely from the computer to the controller. I'm not sure what I'm doing wrong but it seems there isn't enough power for the steppers to work.  They groan a little and occasionaly turn in little jerks.  I know the steppers work because when I hotwire them from the spindle power to the steppers, they seem to work fine. I've tried two different computers and two different controllers with the same result each time.

 

Any suggestions? 

Hey Jim,
       I have used the hobby cnc boards. They work just fine however, they use chopper stepper drivers which can induce spurious signals if you do not plan your wiring carefully. Run your limit switches separately in shielded cables apart from your motor drivers. Keep your cables short use star wiring techniques. Oh thank you for the black box wiring diagram I will give it a try.
Sincerely,
         Willem

Will the SpectraLight software milling programs still work using this new type of interface?

Hello,

 I am trying to help maintain three old ProLight machines for my stepson's Robotics class.  I believe these are relatives of the SpectraLight machines and use the same interface board and black box.  The school provides zero funding, so there is no chance of buying new hardware or software.  Recently the mill started acting up, so I may have a dying black box or ISA interface board.  If anyone out there has leftover boards, parts for a black box, schematics, anything we could use to troubleshoot and repair the old hardware, please let me know, email scooperman at vigproducts dot com. I can't afford to pay a lot.  In the meantime I am going to try collecting all of the cable/interface information posted here.  Thanks, John.