I do want to add in that I'm using drivers that are very popular (yes, I know, price and all) and am having a hard time believing that I have 8 bad drivers.
I've also used 3 different BoB's from 2 different sources and have a hard time believing I have 3 bad BoB's (although admittedly I may have burned up the first two but I doubt it based on the knowledge I gleaned while figuring things out).

I KNOW my cable is good now. I know I have a full 5V power from the PC cuz it's hard wired now. There's still the possibility of the 3.3v signal coming from the PC causing problems. I have a parallel port expansion card on the way with the oscilloscope.

And, yes, once I get the whole thing figured out I will buy a beefier power supply. I don't have an unlimited budget so I have to take baby steps. Once everything is built, I will upgrade the components as this will, hopefully, be a money-making venture in the long run.

I think I'm going to connect an LED to the direction pin to see if it's switching when the motor is hiccupping. It's not definitive but could tell me something til the scope arrives.

when I say it appears to back up, it can back up 10 steps and then flip forward again.
ULP! That is NOT what I was talking about!
It suggests the Dir line is flapping around. Can you stick a LED on that line and watch it while you move the X axis slowly (F=10 mm/sec)? You may need to attach it at different places, like out of the PP, into the BoB, out of the BoB, into the driver.

Yeah, I'm confused too. Time to tinker. I do remember one time when everything was utterly confusing. It turned out to be a problem inside a connector. Fwiiw. I had to take the covers off the connector to find it.

Just a thought, which may or may not have any effect: could you put a 1.5 kohm pull-up resistor on every motor output at the input to the BoB? 1.5 kohm, 1/8 W or 1/16 W resistors.

Cheers
Roger
(More fun than a midnight movie.)

Just a FWIW, it really, really, really is not a good idea to export power from the PC - it likes to live in its own little world inside its steel box A decent 5v 1A SM PSU is a couple of ££ at most, you should have the BOB and drivers powered by that and use the PP to run the opto-couplers on the BOB inputs alone, no need for a USB then, USB sucks.

It does sound like you have issues with the DIR signals and if you do then you will have issues with the STEP signals too which could = your problems.

I had a bad BOB, a good one too, it would work fine on a 9v battery (built in 5v regulator) but not a PSU - tore my hair out for weeks on my first plasma with this, in the end they replaced it and away we went - this was from a good UK supplier, not eBay etc, it was not a cheap BOB - just shows that even good stuff can be bad out of the box.

The fact you have ordered a 'scope shows serious intent here, but if it were me, I would likely have binned the bob and started again on a different make - it really is the king-pin in the system.

Can you not go ethernet UC300eth etc?

Ethernet rocks for machine control as its noise immune and reliable.

I do agree with Dave about going ethernet (UC300 or ESS), except that I am curious as to why the current system misbehaves this badly. Chuckle - that is of course an academic interest!

It may be the BoB, but I still have some money on connectors and the PS distribution. I would definitely like to see a separate external PS for the BoB - common earth to the PC of course.

Cheers
Roger

 - common earth to the PC of course.

I disagree here I would NOT common anything to the PC - just let the parallel port operate the opto-isolators on the BOB inputs under their own power, the BOB should not care how or where the opto LED's get their juice only that they go on and off correctly (3.3v/5v etc) The BOB circuitry should have its own PSU.

Don't forget PC's are rarely earthed - they usually float just above supply earth/ground by a small resistor in the PSU, grounding the PC case is a death-wish on a plasma machine I know this is not plasma BUT what goes good for the worst process in CNC (plasma) is generally bullet-proof for anything else so it's a sound example.

Isolation from PC to machine is vital - this is why they fit opto-isolators on BOB's

I looked up the specs on my computer today and was shocked that it only has a 200w power supply. What a piece of junk.

I was reminded of a problem I had with a computer about 12 years ago. Whenever I went to print I would have to turn off a peripheral or the computer would lock up solid. The company had a dozen of these computers and all of them died within a year or so. All were bad power supplies.

This makes me wonder if it's as simple as that. Maybe the 3.3v is enough for my BoB but the 3.3 is tapped when more than one signal is sent. Worth a thought anyhow.

Tonight, since I don't have the oscilloscope yet, I will connect leds and maybe pullup/down resistors to see if anything jumps out.
I can also, probably, dig up a 5v power supply and isolate it from the host computer.

OK, earthing. A topic all in itself!
My PC is thoroughly earthed. Australian Standards require that, and they are very good standards.
My whole system is also earthed to the same Earth point - every bit of it.

Electrical isolation between PC and CNC for all signal lines - YES. This can be optical, which is very good and I use a lot of it, but remember that Ethernet devices are also isolated from the Ethernet cable: they are transformer coupled.

So - full signal isolation, PS grounding, CNC frame grounding, and lots of screening everywhere (screens to ground).

As for the PC power supply - sigh. You will have to monitor for a bit. External PSs are always recommended.

Cheers
Roger

Ha, I saw "earthling".

So, I tried the LED test. The LED doesn't blink as if it's getting a cruddy signal.

I thought I'd try some other things while I'm waiting for the Oscope and haven't found a 5v supply yet.

First. I get 4.8v at the signal pins to the driver. It appears consistent via the LED and the multimeter which, admittedly, are both flawed in their functions.
Then I thought I'd go into the motor tuning and play with the velocities to see what happens when.

Here's where it got fun.

I set the velocity at "1", but the program saved the velocity a hair over "1", and was able to single step the motor perfectly. It did, however, make noises while holding certain steps.

Next I thought I'd set the velocity to 50. When cycling the motor it was a VERY loud motor.

When I set the velocity to 45 the motor hops around. It's back and forth all over the place.
At 100 the motor runs smooth.

So, something causes the motor to freak out around the 45 steps per second mark.

I found a 5v power supply and disconnected the PS from the computer.
I common grounded (earthed) everything.

No change in the results.

This is compelling.