Motor supply ground vs Logic ground?

[Gashmore]

Putting the finishing touches on my driver enclosure. The system is based around 3 IM483 drivers, CNC4PC breakout board, charge pump and relay board. Motor power is a 600KVA transformer, rectifier bridge, caps etc delivering 42VDC. Logic power comes from a Power One 5/12V regulated supply.

The 5 and 12V grounds and all cable shields are tied to chasis ground but the motor supply ground is floating about 6V below chasis ground with the system idle. Probably higher with the motors running. What really has me bugged is that the IM483s have a 5V input to power the opto-isolators but no coresponding ground. The only ground is to the motor supply. I m afraid I will fry the boards.

Do I need to tie the motor supply negative to chassis ground?

[fyffe555]

Hi,

It depends on what version im483's you have.

The 8P2 and PLG drivers don't have a seperate P1 ground. The P1 34 pin option does have a dedicated ground for the input side including Step/direction and opto. IMS told me that this was to isolate the additional step clock out /direction out on P1 and that this was a seperate ground from that on the P2 (power side) connector, but a common ground for chained drivers.

I have both types and was advised by IMS that for the 8ps and plg types the opto supply (pin 4) is grounded by the Power side (P2) ground (pin 3). This works for me but all my grounds are tied to a common, including Power and logic and case. If your motor power ground wasn't at 6v under I'd say ground it to case and logic also. No straight answer but hope that helps.

EDIT: I've just found the Product manual for the IM483's (not later programmable or IE versions) It shows a minimum connection option where the step or direction switch is to ground with the +ve 5v to opto supply... If you need the manual I'll post it.

PS - Are you the same gashmore guy building a yacht in his back yard?

Andrew

[Al_The_Man]

The manual shows the motor power supply grounded and recomends that each amplifier NOT be daisy chained for both the + and common conductors from the DC supply, the 5v going into the IM483's is your 5v supply and is isolated by virtue of the opto's, this path is through the opto through your driver and to your 5v ground, so there should be no connection with your 5v supply and the amplifier apart from the fact that your 5v is grounded and the Motor supply would be grounded, which according to the manual is allowable.
Al.

[Gashmore]

Yep, that's me. Working on the interior and hatches and realized that I need to mortice in about 60 very complex shaped latches. A great excuse to build a new tool. 80% of this machine came from my junk box so I don't have a lot of cash in it.

I saw that diagram but still wonder. Looks like the +5V that powers the opto-isolators is grounded through the signal lines. I guess this makes sense as if there is a fault on the motor ground the opto-isolators would protect the rest of the system. Guess I will schlep the 'scope and function generator down to the shop tonight and see exactly what is going on.

[Al_The_Man]

The 5 and 12V grounds and all cable shields are tied to chasis ground but the motor supply ground is floating about 6V below chasis ground with the system idle. Probably higher with the motors running.
Do I need to tie the motor supply negative to chassis ground?

If you are using an electronic meter to measure the difference, it is in all probability a leakage reading and would not support any current flow, you could try grounding the DC supply using first a 100ohm resistor between common and ground and if the voltage collapses to any degree try a 10ohm, if this result in virtual collapse of the leakage reading, then it would be safe to actually ground the Motor DC common.
Al.

[fyffe555]

Glenn...

I've been following your build from your web page for a while. It's dangerous - almost enough to make me want to go do it. Almost - I realise I personally don't have what it takes to spend so long consistantly working at such a large task. So I bought one instead.

To add to something important Al said in his post; Various IMS doc's say that the drivers can be chained. What they mean is that some IM483's output step and direction signals so a second or n+1 drivers can slave from the first. BUT they're chained from the logic side only, the drivers motor power *must* be star wired and not chained. The IM drivers are apparently very efficient at dumping emf back to the supply and mid power chain drivers don't like it. Don't ask me how I know this.. In addition I'd recommend putting a cap locally across the driver power pins if you're running higher step rates.

If your driver is an 8pin P1 type then the driver logic is powered from the motor power side. The 5v opto supply does only drive the led side of the opto and nothing else, so grounding through step/direction is ok. The 34pin types have a opto side ground that has to be common wth other chained logic drivers.

Obvious stuff, but there's some consideration needed about how the signal lines are connnected to your PC since the signal lines are sinking ground for the opto's - I've a couple of drivers that want to sink 28ma and that's over the capacity of my laptop parallel port. Buffering the P Port is a requirement I think, which I think the cnc4pc breakout does?

I like these drivers and used them a lot, and have them on my mill and lathe at home and they've suffered a lot of abuse. They need a good heatsink though if you're running at 48v and 4a.

I have the older IM manual for these drives and can email it if needed. Can't attach it here 'cos its too big at 1195kb

Let us see the machine when you've got it running..

Andrew

[Al_The_Man]

To add to something important Al said in his post; Various IMS doc's say that the drivers can be chained.

Actually what I was refering to and generally it applies to all power supplies feeding servo's or steppers is that the motor power +DC and/or the commons and grounds should NOT be daisy chained from the supply to the drives.
All drives should have their own independant supply back to the PS.
Al.

[Gashmore]

These are the 8 pin models. All the motor power runs go back to a fuse panel at the power supply but the 5V side is more or less daisy chained. Found an article on the Gecko site that recommends a 470uF cap on each drive so I will do that too. The drivers are mounted on a monster heat sink sticking out the back of the enclosure. (Scrap box parts, remember?)

One other thing. How do you estimate back EMF? The IMS manual is sprinkled with red boxes warning never to exceed 48V including spikes and back EMF. I have a big cap on the supply, about 3x larger than the calculated requirement so hopefully it will absorb any transients but I am still wondering how close to 48V I can drive the system. I have a Staco variable transformer feeding the power supply so I can set the motor voltage just about anywhere I need to.

This is turning into a really satisfying project. I have been accumulating junk for 40 years and this is the first time I have had a chance justify it by using some of the really interesting stuff. I just hope it works when I finish.

[fyffe555]

I can't directly answer the Vemf question but here's what I've found with these drives, take it for what its worth;

Stepper motors are fairly effective two or four phase alternators. Vemf is proportional to the motors speed and the potential opposes the driving power source, so increasing the net Motor Power In voltage present. Vemf is only an issue for the driver when the motors are being driven. More so the faster the motors speed. With a sophisticated chopper driver changing frequency, rather than time based, the Vemf increases the duty cycle of the driver since the Vemf opposes the increase in current in the coil during on time and also aids decay of current on coil off.

To calculate the Vemf you'd need to know the rated Ke (back emf constant - volts/k steps) or Ker (back emf constant - volts/ radiens/second) for your motor. If you have that there are several methods and spice models out there but its basically a simple function of constant and speed. Obviously none of the surplus motors I've got have those constants available!

Without a constant it's down to measuring your own. Either chuck up the motor in a drill wired in the wiring scheme intended (bipolar parallel/series etc) and run it at a known speed measuring the suitably smoothed or rectified output. Or briefly measure across the driver power inputs while the driver and motor are running at the (known) max expected step rates. Obvious issues with both methods.

My powermaxII 274oz/in motors developed a measured 2.2v at 300rpm. With a 32vac supply rectified to 47.4vdc ( yeah I know the math doesn't work but thats what's produced) and a volt meter over the power in on the im483 I actually get 48.6vdc with the motor running at the same 300rpm which wasn't what was expected.

FYI I did some simple tests of power produced by the same setup in bipolar parallel changing supply Voltage and Amps. Changing (reducing) Amperage produced a significant reduction in torque. Changing (reducing) voltage made almost no change in produced torque but did reduce slightly the speed possible.

Moral of the story? Only worry about Vemf if you're running at faster motor speeds. If you want to account for it then use the manufacturers constant or measure your own. If you don't want to do that then simply set the V to a few volts under the 48vdc max, BUT make sure you set the maximum rated Amperage for your motor/driver combination if you want maximum torque.

Andrew

[fyffe555]

Actually what I was refering to and generally it applies to all power supplies feeding servo's or steppers is that the motor power +DC and/or the commons and grounds should NOT be daisy chained from the supply to the drives.
All drives should have their own independant supply back to the PS.
Al.

Thats what i said

[Gashmore]

The motors are Sanyo Step-Syns that I salvaged from a couple of old IBM DP printers so I don't think there is any way to find out the back EMF constant. I am going to put the 'scope on a phase output and gradually increase speed and voltage until I start seeing peaks close to 48V.

Being built mostly from parts I had on hand, this rig is far from optimally balanced. Everything is over sized with an eye to upgrading to bigger motors and a set of Geckos so I have way more amps available from the power supply than I will need so I am guessing that I won't see much voltage drop as the motors load up.

Not to contribute to thread drift but has anyone considered using the same power supplys and drivers on two different machines? Now that I have broken my vow to keep computers out of the shop I have been casting a perverted eye on my ZAY7045 mill.

[fyffe555]

Until I managed to EBay some more IMS drivers I shared PS and drivers with two machines. It worked ok, used a different TurboCNC INI file for each machine, Mach2 can have different profiles too I believe. Problems where the time taken to swap over and there was always something to fix - limit switch faulting or some such. Also different motors needed different limit set resistors on the drivers. I tried a switch but these drivers do like to have the resistor right on the driver. I used a connector instead - my drivers where the 34 type with pins connections.