24vdc power supplies

craig_avanzar · Dec 7, 2007

Hey guys. I've been having an intermitant com's problem between a Motoman robot controller and a Mitsu FX3u-48M via CC Link model FX2n-16 CCL-M. Its actually 3 robot controllers, a master/slave set-up for arc welding and the third is a spot weld. I have looked numerous times at the program including with the integrater comapany and with the robot support person and they just point fingers at each other. I scoped the com's signals of this cell and compared them to the identicle cell next to it and the wave pattern was all over the place at the problem cell and relatively even and square at the others. I've double checked that all shielded coms wires are free and clear from any high voltage thinking induction. I talked to another guy with the integrater and he suggested that I check to see if the 0v at both power supplies were using the same ground for both the plc/CC Link and the robot. I want to try a test and connect a wire from each to a common ground. Does anyone see a problem with this test? Any other suggestions that might cause the difference in signal conditions?

Stuck in S.A.

mordred · Dec 7, 2007

single ground point is the better way to go otherwise you can create ground loops also signal grounding should only be done on one end for the same reason. I would also check the CC_link instqllation guidelines manual I beleive they use terminating resistors of 120 ohm to balance it at 60 ohm but not positive on those numbers as we just recently recieved our first CC-link device so I'm still studying it

MartB · Dec 7, 2007

I have to agree with Modred - single ground point is the best way to go, and in my experience, this will cure most robot comm's issues.

craig_avanzar · Dec 7, 2007

• In the CC-Link system, the terminal resistor required varies depending on the cable used.

From the manual regarding resistors.

- When a dedicated CC-Link cable is used: 110

Ω, 1/2 W (brown, brown and brown)

- When a dedicated high-performance CC-Link cable is used: 130 Ω, 1/2 W (brown, orange and brown.

mordred · Dec 7, 2007

heres the CC-link installation guidelines to help you out

craig_avanzar · Dec 7, 2007

Thanks alot Mordred. After looking at this section:

4.3

Radiated - Induced Noise

In order to prevent induced noise, keep communication cable as far away (electrically and

mechanically) from other cables (power line, etc.) as possible. It is recommended that the communication cable be kept a minimum distance of

≥ 100mm from any power line. Avoid communication cable setup within the same conduit or set in parallel with any external power wiring (120VAC). Avoid communication cable setup on a control panel where high

voltage devices are mounted. If necessary, attach a surge absorber on devices that radiate noise. Do not connect non

CC-Link compatible products (lightning arrestor, etc) with CC-Link communication line. This causes digital noise reflection and/or attenuation and communication failures. If it becomes necessary to cross the CC-Link communication cable and power cables, cross the cables at right angles to each other to reduce any radiated or induced noise.

I went back and double checked inside the robot panel and some brainchild had ziptied it to the 480v incoming supply power. I don't know if this is my problem or not but I'm sure it didn't help.
Thnks for the info as I'll add it to my Mitsi folder as reference in the future.

mordred · Dec 10, 2007

also check all terminations as one bad connection can lead to the type of signals yoiur seeing on the scope

Lancie1 · Dec 10, 2007

I went back and double checked inside the robot panel and some brainchild had ziptied it to the 480v incoming supply power.

In which case, I would check the resistance-to-ground for the 480 volt power also. It is possible that there is a very poor 480 volt ground, and the stray AC noise is following your comm ground instead, which would be another noise source for your signals.

craig_avanzar · Dec 11, 2007

Ok, I think I have checked, checked and double checked all terminations, added grounds to the shields(they are connected internally also), and verified the resistance to ground at the line voltage for min/max is 27.74 on the min and infinity at the max and average. The enclosure did not have any grounds attached and were relying on grounding points from the 110v supply to the plc. We added an enclosure ground and attached the shield to it. We are now in the monitoring mode. This is not a new system as it has been in full production for a little over a year. It has a twin that sits right beside it without any problems and a resonably stable square wave pattern.
Thanks to all for your help so far.

Hovering in S.A.

Lancie1 · Dec 12, 2007

...and verified the resistance to ground at the line voltage for min/max is 27.74 on the min and infinity at the max and average.

I am sorry but I do not understand what you mean by this. Resistance to ground is usually measured by driving a test rod at least 25' away and then measuring resistance between ground-point-in-question and the test rod. It should be 25 ohms or less. If it is more, then it can be said that you have a "poor" grounding system.

What type of building grounding do you have? The old buried loop-and-rod system? Or are you lucky enough to have a Ufer system?

craig_avanzar · Dec 12, 2007

You are absolutly right. What a brain F**t. I used to work for a contractor doing sub-station installations and instead of checking ohms he would meg it using each ground rod point to point before adding the grid. Forgive my blue screen on brain. I'm not sure what type exactly you call it. I was told that there are 3 ground rods sunk and that the building is attached and each rod is connected to each other. There is no grid as it is a manufacturing facility. This was asked because they are shutting down Christmas eve to work on there grounding(facilities maint.) Is it probable that their grounding (bad neutral to ground bond?) is the catalyst thats causing me headaches? I know almost anything is possible but there is alot of the same type of equipment in use and only one of ours is flaky.
Thanks Lancie1,
Still learning in S.A.

mordred · Dec 12, 2007

Would not a check of neutral to ground also reveal poor grounding? On 3 wire 3 phase it should create an inbalance? or am I incorrect on that?

Lancie1 · Dec 12, 2007

...that there are 3 ground rods sunk and that the building is attached and each rod is connected to each other.

I know there are sandy parts of Texas where 3 rods in a loop would not be enough to achieve a 25 ohm path-to-ground.

keithkyll · Dec 12, 2007

craig_avanzar said:
... I scoped the com's signals of this cell and compared them to the identicle cell next to it and the wave pattern was all over the place at the problem cell and relatively even and square at the others...

Identify the noise source. That's what needs a ground. Subtract the square wave first, so you can see the noise. Can you scope the line unpowered?
Can you scope the ground?
If it was a SCR or phase angle controller, you might see noise spikes at 60 Hz, for example.
Crude method for finding a noise source is an AM radio. Noise gets louder near the source.

mordred · Dec 12, 2007

Another lengthy method is to power each device one at a time and check the power and noise levels on each device in your cell disconnect main power to the cell and check the incoming power sort of the same procedure as finding a problem ground short on circuits with multiple devices make sure your constantly checking for changes on your original checked circuit.

24vdc power supplies

craig_avanzar

Member

mordred

Member

MartB

Lifetime Supporting Member

craig_avanzar

Member

mordred

Member

craig_avanzar

Member

mordred

Member

Lancie1

Lifetime Supporting Member

craig_avanzar

Member

Lancie1

Lifetime Supporting Member

craig_avanzar

Member

mordred

Member

Lancie1

Lifetime Supporting Member

keithkyll

Member

mordred

Member

Similar Topics