Mesa 7i96S

Thanks, but...

I guess this is a general CNC question, since technically the 7i96s which this thread is about (and the spindle is commanded by) is working properly. ie, telling it to not run below a certain speed won't fix things since LinuxCNC via Mesa 7i96s isn't telling it to move, but it's moving anyway. And it will move even when disconnected from the control hardware.

But since you have experience:

Am I better off buying a new power supply, and try to get them to replace this one?

Or is there some kludge people do to, say, wire a resistor to ground to bleed off that "back current." Or is having that current there a sign of a bad asset that should be quarantined from the machine?

Thanks for your thought!!

A quick glance and I don't see where you mentioned what type of amplifier is controlling your spindle.

Old DC drives like KB and Minaric have an "IR Comp" pot and if that's the type of drive you have it's typical for them to output some current until you adjust the IR comp.

Newer drives it's pretty typical to have some sort of output enable on the amp. It sounds like you are not controlling the output enable by hooking it up to the mesa card rather it is constantly enabled by a jumper or maybe it just doesn't have an enable circuit.

What PCW was suggesting by "Disabling" the drive is doing just that. There are HAL pins that can be connected to outputs so that when you command the spindle on it will turn the output pin on. That output pin would be connected to your drive's "enable" input. Right now it sounds like you're just trying to turn the spindle on and off through the analog output which isn't altogether the safest way to do it anyway.

If your drive doesn't have some sort of enable you can probably do something with relays but I wouldn't recommend anything like that unless I knew exactly what you're trying to control.

Brilliant! I think I get it.

(And this brings us back to the Mesa 7i96s board topic, thank you!)

I'm using the PSU that it came with, those two wires going into the far left carry the 0-10V control signal from the Mesa card pins marked Spin - and Spinout.

And... it came hard wired with that short between the "Off/On" terminal.

You're so right for making my world safer too... when I pull that wire short off those terminals, the motor does indeed stop! Great.

The potential across those is 10V.

Except for the spindle encoder, I don't see a way to get a native 10V out of the Mesa 7i96s card.

I could use a logic converter to step up a 5V output pin, of which there are many.

Then in .hal, do something like:

And physically wire that wire to the 5v-10v logic converter and from there to the "On" terminal in the picture of the PSU.

Is that right?

The spindle analog wiring should remain the same
(SPIN+ from drives 10V potentiometer suppy)

VFD drives are very noisy and its best to not connect
anything but isolated circuits (like the 7I96S's analog out)
to the drive

I thought the outputs on a 7i96s are isolated and outputs 4 and 5 are "high isolation"? The manual specifically suggests outputs 4 and 5 for "noisy devices like VFD's and plasma torches"

Vector,
Never seen a "drive" like that, so no idea how noisy the thing is, VFD's do tend to be problematic but I'll tell ya what I think you would do ideally. You don't need a "10v supply", if there's 10v between terminals 1 and 2 of that power supply, then it's supplying 10v on one of those pins. Outs 4 and 5 are non polarized so pick one of those 2 on/off terminals on the power supply, let's say terminal 1 is the positive side. Run a wire from Terminal 1 of that power supply to OUT5A. Run a wire from OUT5B to terminal 2. (you could use OUT4 as well). Whenever you turn output 5 on those 2 terminals are connected and it should turn the Power Supply on.

I have no idea if those 2 terminals are meant to enable that power supply quickly, but they need to be meant for that purpose.

Not sure where you got that hal segment, but it doesn't jive with a 7i96s, I can't remember how I did my mill but I think you should be using the motion.spindle.N.on pin to turn a spindle on. No need to use the "not" component

Double check your hal file to make sure PnCconf didn't already setup that motion pin on a signal. if it did just connect that signal to the hm2 pin I have there.

As for the possible noise issue, well that's something else that may need to be addressed but I'd wait till PCW responds about that output's isolation

Thank you Blazini36! Wow.

I just spit out that hal code myself to try and get you guys while you might still be awake... based on documentation and what's in the pncconf generated file I have.

And you delivered! It does not have the spindle-enable nor motion.spindle.0.on pins.

Also thank you: without your help I would not have figured out that the OUT5 A/B and OUT4 A/B are connected when they are turned on.

Since you mentioned it being in the 7i96s documentation, I went back to it and discovered page 14 of the the manual which has this information you're sharing.

So for anyone who finds this thread in the future: be careful because while they connect, they can't take much current, and need a 1/4 amp fuse.

I'll be trying this out in the next few days and update this thread.

Thanks again!

"spindle-enable" is the signal, it's arbitrary and you can make it whatever you want but you have to create a signal to connect 2 pins.

I tend to write hal files like this because it's easier to follow

IIRC pcnconf tends to take common pins and connect them to signals and leave them hanging, I haven't used it in a long time though. IIRC it does this alot:
when you see that you just leave it and use pcnc_signal with the pin you need. 

Also thank you: without your help I would not have figured out that the OUT5 A/B and OUT4 A/B are connected when they are turned on.

Also, keep in mind the other outputs on the card work slightly different. They're polarized so it actually matters whether you connect the output common to output+ or output-.

 

I thought the outputs on a 7i96s are isolated and outputs 4 and 5 are "high isolation"? The manual specifically suggests outputs 4 and 5 for "noisy devices like VFD's and plasma torches"

That's correct. My suggestion was not to use an additional power supply for the drive enable (and its not needed
since the drive enable is just a contact closure type input)

Thanks again everyone!

Its been a few days before I could get back to this.

Progress includes wiring a fused line from out-5 to the "on" terminals of that spindle power supply...

... but I'm stuck on something else at the moment...

(keep in mind I'm a newb with this hardward, so I have my hand poised to smack my head...)

I have the home switches for x, y, z and a wired to the TB3 row: IN0-IN3.
They are wired in what I think of as "normally closed:" ie they have 5v until they are tripped, when they have 0V
I've checked with a voltmeter and sure enough, the voltage is there. Also the teeny led's by those terminals also are on.

the .hal file created by pncconf reads:

But when I crank up LinuxCNC, and put Hal Meter on the signals for home-x, y, etc, they read as False. (as do the corresponding pins)

I expected them to read True, or at least to change when I manually trip the switch.
I don't want to 'test' the homing sequence like this, since I'm pretty sure it will crash into itself. :-o

???

{Separately, I'm also getting the "error finishing read" issue... I'm studying the long thread by that name and I think my next move will be to backport to the 5.10 kernel (I'm currently on Bookworm, 6.x kernel) ... both JT and RodW indicate that's a fix.
Note: I have the hal-meter mystery before the error finishing read}