BIMA 300 CNC Center

The reason that i like my job, is because it never gets boring. Well, frankly speaking, it gets boring sometimes, but usually there is always something new to discover. I am a CNC technician, and luckily, i work with companies like IMA, Weinig, Schelling, which all have only high-end technology. I decided to upload some of my work because first of all, i think that it will be interesting for people who own such machines, and second it could be my future CV :D. I am not going to post EVERYTHING i do for it has no sense, only the problems that i enjoyed solving.

My colleague in front of the electrical cabinet

The customer reported that, sometimes, when he presses the power button to activate the machine, the NUM controller fails. Whenever this happens, he gets no error message in the computer (UI) screen. The LEDs on the NUM communication module do not indicate a problem, but the "Serial Link Fault" LED on the power module and the axis drives are turned on.

My experience tells me that, the word "sometimes" will give us a hard time. And this is normal, because if something is destroyed, then the problem appears all the time, and with a quick reverse engineering the problem can be easily found. But if the problem appears and disappears, this is the worst case scenario. We visited the customer many times, but the problem did not appear. Until finally, he called us and told us that this very morning, he was unable to turn on the machine, no matter how hard he tries.

This is what we faced when we went to the customer (click to enlarge):

The red LEDs are the "Serial Link Fault"	This is the topology of the electrical cabinet, regarding the NUM controller

Normally, we started by checking all the wires on the NUM. We took out each and every wire, cleaned it with spray and plugged it in again. But the problem still occurred. Then i tried to reproduce the error, by removing wires while the CNC was in operation. I did this for all wires. Whenever a wire was unplugged, the system was forced to fail, but every time we got different error messages on the UI. Remember that the problem that the machine has, did not send an error on the UI. So this is not a wire problem.

Then, we took out the first module of the NUM (which is responsible for the communication and has also the CPU), to see what' going inside and cleaned it with dry air. There is a trick to the whole operation: There is a rechargeable battery 3.6V, which is used to maintain the system parameters of the machine, whenever the controller is not on power. This battery must NOT be unplugged from the motherboard no matter what, because if it does, all parameters will be lost, and it is not fun passing the parameters again.

The first module is on the surgery table	It is very well made, after all, NUM is a top brand (and top-price)	Be careful with the battery for bad things happen if it is unplugged

Afterwards, we put it back into the cabinet and re-tested the machine. Guess what: the problem was still there. So we wear again our thinking cups... The customer said that the problem only occurs when the machine is in static mode, or when the power key is pressed, but never when the machine is in operation. A headache. We also tested the earth grounding cables, as i know that if the grounding of a machine is incorrect, crazy things happen. To be sure, we connected a new wire directly o factory's ground, yet the problem still occurred.

Another strange thing: Moving the battery outside the controller did not cause the problem. But moving it inside, the problem appeared. So i thought that the problem was on the charging PCB, because the battery was in direct contact with it. It was about time to take the controller on the surgery table again... This time, we checked and double checked the charging PCB for cold solder or other problems, but nothing suspicious could be found. So, we put the controller back in position again.

As always, i always slap-test electrical and electronic equipment. I call it "slap-test", whenever i slap slightly (or not so slightly) something to see if an error occurs. I did slap-test all the modules many times, but the problem did not occur. But something crazy happened: My colleague drove mad, and took the battery out and started moving nervously saying "you see? it is not the damn batter". And then, he left it on the controller with force! APOCALYPSE II: The error occurred when the battery touched the controller: Next logic though: The battery leaks to the controller chassis. So we connected a wire to the chassis, and started touching it on the battery... No problem at all. That was crazy...

And then, i slap-tested the controller again, not with my hand, but with a screwdriver. That's it! The error occurred every time i was hitting the controller with the screwdriver. Strangely, no matter how hard i was hitting it with my hand, it did not fail. But now i was sure that the problem comes from high-frequency vibrations. To discover exactly which connection was bad in the controller, we made a table inside the cabinet, and placed the controller on the table wide open, while the machine was still in operation:

And then, i tried hitting (slightly) the parts on the PCBs one by one with the screqdriver. APOCALYPSE III! The problem occurred when i hit the CPU! Actually, i only had to touch the CPU for the system to fail. It was obvious that the problem was on the CPU chip itself. Relief. So, again, we took the controller on the surgery table. This time, we took out the motherboard to check it for cold soldering or other problems. Again, the battery charger along with the battery had to be always connected on the motherboard. No problem was found in a first glance, so we took out the CPU, cleaned the base with spray, and put it back again:

This time, we took out the motherboard from the chassis	The CPU chip was removed from its base	And the base was cleaned with oil-less contact cleaner	There was no thermal paste between the chip and the heatsink...
... so we added some and put it back together	We put the CPU back on the motherboard...	...and test-run with the controller open... No problem at all !!!	That's me testing the wires once more.

So what really happened is this: When the machine is armed (when the operator presses the power button), two large load relays are armed simultaneously. These relays, cause high-frequency vibration on the back side of the cabinet, where the NUM controller is fixed, and this vibration caused the system failure. Most probably, a pin from the CPU had bad good contact with the CPU base.

So what we've learned from this? Slap test with the hand, and if nothing happens, use a screwdriver :D

Name   Email (shall not be published)   Website Notify me of new posts via email Write your comments below: BEFORE you post a comment:You are welcome to comment for corrections and suggestions on this page. But if you have questions please use the forum instead to post it. Thank you.