Author Topic: CNC Controller Repair (Any Ideas) Has Me Stumped!! (Read 10105 times)

KG7AMV · « **on:** October 17, 2014, 03:40:26 am »

My Project

I got a Second hand CNC lathe and it has a Microkinetics Controller that has a Failure on one of the inputs software says unexpected limit switch closure on M2- limit switch. (No Limit switch Hooked up) I Contacted the company!! The company wants $300.00 for a new board ouch or $150.00 to see if they can repair it. They have no real tech support or documents on the unit unless you want to give them $150.00 for engineering service Ouch! A schematic would be helpful! So this brings me to where I am at right now dusting off my tools from years of storage! Trying to diagnose it find out why it is giving the error.

To me it is not worth $150.00 to have it repaired or $300.00 for a new one! As a new breakout board that will allow me to run full featured software LinuxCNC or Mach3 is a fraction of the cost. Their software Turnmaster Pro is rather limited with functions would be nice if they allowed you to select what axis uses what motor just in case a controller has an issue you could switch it to talk to another move a few wires and have the machine running in a matter of min vs having it down for weeks on end waiting for a repair or replacement parts.

From the message the error seems to be on the M2- Limit Switch Input.

So I traced the M2- limit circuit down. The m2- limit pin goes from the HD50 thru RP5 on pin 15 out Pin 3 thru an opto coupler U21 pin 4 outputs pin 14 thru RP4 to the Xilinx XC95108.

The passive components RP4,5, and the Optocoupler seem to be working as they should. When power is on the pin on the XC95108 is at 5V high when I hook up a limit switch and trip it the pin goes Low to ground as it should. I cross checked it with a known working circuit M1- Limit and M2- limit seems to work as it should.

Another Circuit tied to the z Axis (Z- Jog) starts on Pin 11 of RP10 Passes thru RP9 Branched off thru a Cap to Ground then to the Xilinx The circuit seems to be the same as the X axis & Z + (Jog) in measurements.

Any Ideas? To Check before Busting out the Jtag this weekend to do a boundary Scan to see whats up?

Photo of the board I am diagnosing is attached I will try and get a better larger photo later on.

The only Info on the board can be found at Microkinetics MN400 http://www.microkinetics.com/index.php?page=mn400

kc9qvl · « **Reply #1 on:** October 19, 2014, 08:57:56 pm »

Dunno how much configuration there is in their software. Possible setup error. Looking for n.c. contact for the limit and has n.o. ?

KG7AMV · « **Reply #2 on:** October 28, 2014, 08:58:14 pm »

N/O switches. I do not think it is a software config when Z- is switched on H3 the second circuit it does nothing But Z+ Works that the other circuit tied to the Z axis.

KG7AMV · « **Reply #3 on:** October 28, 2014, 09:00:30 pm »

Got the Jtag out... Got the computer with the LPT port fired up! The computer will be used with my LPT breakout board so dual use! So maybe this weekend I will run some tests.

KG7AMV · « **Reply #4 on:** October 28, 2014, 09:36:10 pm »

Scored these drivers today... for cheap! Should be a simple fix prob the T6.3 Fuse and a set or so of IRF9540 & IRF540 transistors. I have a bunch of N & P Channel fet's from old CB Days! Could be a simple VR will find out when they arrive!

http://www.ebay.com/itm/251686632133

kc9qvl · « **Reply #5 on:** October 30, 2014, 05:33:42 am »

With those stepper drivers you could most like use a cheap lpt bob like a c-10. Then you could run mach3 or linuxcnc.
If you are into raspberry pi type microcontrollers. Check out the MachineKit. Linux cnc on a beaglebone black. Use this on my sherline micromill.

C-10 break out board
http://www.cnc4pc.com/Store/osc/product_info.php?cPath=33&products_id=45

MachineKit.
http://blog.machinekit.io/p/machinekit_16.html

KG7AMV · « **Reply #6 on:** October 30, 2014, 01:12:19 pm »

I got the NM400 hooked up to the Jtag and the Computer. Learning curve on the TopJTAG.com software! I Will install the Xilinx Software this afternoon. I was able to find the BSDL file on BSDL.info and get the probe software to talk and see live pin data! I was not able to get TopJTAG Flash Programmer configured but I was tired and it was near bed time so I gave up for the evening. Now to learn all the functions on the probe and programming software and label all the pins in the project.

From what the boundary scan showed the key pins seem to be functioning and it only detected the Xilinx in the chain then i had to load the bsdl file. I am not sure about the DSP ICE header never messed with DSP!

On the Drivers, I am Already wiring a drive rack that has three of them drivers in it to a bob to use mach3. I had first stepper movement last night. I have to rewire some relays in the rack to get them to work with the BOB correctly.

I Will add a switch as I want to keep the rack plug and play useable with the MN400 if I ever fix it.

Inside the Drive Rack.

kc9qvl · « **Reply #7 on:** October 30, 2014, 10:30:00 pm »

What size stepper motors are you driving? Those 3 drivers and transformers are beefy.

What are the 2 solid state relays for?

KG7AMV · « **Reply #8 on:** October 31, 2014, 11:27:28 am »

Nema 42 1,700 OZ and Nema 34 470 OZ

KG7AMV · « **Reply #9 on:** November 15, 2014, 02:23:23 am »

Have not had time to work on the Controller been organizing my parts bins. I did pick up 6 drivers on ebay for repair I got 2 working the day they arrived waiting on fets and fuses to fix the others. http://www.ebay.com/itm/251686632133

KG7AMV · « **Reply #10 on:** November 15, 2014, 04:18:18 am »

I cannot believe they charge $774 for a box with this poor wiring job... If I decide to give up on the project may send it in for a EEVblog tear-down.

Falcon69 · « **Reply #11 on:** November 15, 2014, 04:59:27 am »

I didn't read the whole thread, pretty much just the first post.

If you are receiving that message, my guess would be that your break out board uses optocouplers and it has burned up, making it act as normally open switch now, instead of normally closed, triggering the software that a limit has been switched. Simply replacing the chip should solve the problem.

Those motors you are driving may be too big for your application. Bigger doesn't mean better. This is covered over and over again on thousands of threads over at CNCzone.com. There are people who are getting better results with their nema 23 motors, then with their 960oz. nema 34 motors. It's just how there inductance and voltage rating are.

Give Ahren a call over at CNCrouterparts.com. He's a pretty cool guy and very helpful. I bout 5 times NEMA 34 motors from him to use on my machine. I am building a very big machine, 4x8 but the gantry is built differently and more then usual weight to it, so I'll be using a nema 34 motor to control that axis and the z axis as well. Ahren can set you on the right track.

In either case, for safety reasons, you want to make sure you use normally closed switches, and place them in series. That way, if one is triggered (causing it to become an open circuit), it triggers the software. If the switch is accidentally cut or disconnects, it triggers the software, as the software will see it as normally open then, and shut the machine down.

Falcon69 · « **Reply #12 on:** November 15, 2014, 05:14:30 am »

oh, another thing. what software are you using?

If you are using Mach3, there is a configuration in there for debounce settings. I know alot of people were having problems with the software getting false triggers on the limit switches. Playing with that setting seems to fix it.

I would say check your wiring see if there is something wrong from the breakout board to the switch, but it sounds like you have tested that directly with the onboard optocoupler and is working as intended.

Oh, and just glancing at the thread....

If you are using a laptop with an printer port (LPT), that's not going to work so well. People have been having a lot of trouble running a cnc with an old laptop using the printer port. Also, I have heard that on some computers running windows 98 or below (i.e. windows 95) is also creating some problems with Mach3 and the printer port. You may want to invest in an ethernet smoothstepper or a USB smoothstepper. Much much smoother machine running using one of those devices.

I will be using a smoothstepper with my machine. It plugs directly into my PMDX-126 breakout board I'll be using. I chose that breakout board because it can also have onboard control for spindle speed. I'll be using a 4.5kw spindle motor, watercooled, running off a hitachi VFD.

The C-10 breakout board that someone suggested is also a good break-out board, but correct me if I am wrong, doesn't the limit switches for those need to be pulled up to 5 volts, instead of sinking to ground?

I am not sure how big your machine is, but if you are using those big motors, I'm guessing your machine is at least a 4x8 ft. I was going to suggest the Gecko G540 breakout board which has onboard drivers for the motors, but it won't support NEMA 34 or 42 motors and i think only rated for up to 50 volt operation? Some of those motors run off 72 volts.

I really think you are overboard on your motor size. Need to talk to Ahren or some of the more experienced guys over on CNCzone.com. I thought I needed big motors like that for mine as well, turns out that 960oz NEMA 34 motors was all i needed. I didn't need the NEMA 42 like i thought i did.

KG7AMV · « **Reply #13 on:** November 15, 2014, 06:37:05 am »

This is on a 12X36 lathe.. I am switching over to Mach3 and ditching the Microkinetics Software!

KG7AMV · « **Reply #14 on:** November 15, 2014, 09:17:11 pm »

Ugghhhh China! Got a E-mail they saying they have sent me IRF540N's and IRF9540N's for the Drivers not sure if the N will replace the non N and not give issues we will see in about a week. If not only out a couple bucks wasted!

replacing some broken LED's today and going to pull all 8 fet's 4 IRF540's and 4 IRF9540's even the good ones to be ready for replacement's for rebuild.

The diffreance i see is the IRF540 has a resistance of 0.077 and the IRF540N is 0.044 ohms not sure it would make much difference?

rs20 · « **Reply #15 on:** November 15, 2014, 09:53:51 pm »

Quote from: KG7AMV on November 15, 2014, 09:17:11 pm

The diffreance i see is the IRF540 has a resistance of 0.077 and the IRF540N is 0.044 ohms not sure it would make much difference?

Lower resistance is better, if anything. The only other thing to check is that those resistances are at the same Vgs; and also check the gate charge, gate capacitance etc -- if that's similar/lower as well, then you're just getting an all-around superior part.

Marc M. · « **Reply #16 on:** November 16, 2014, 01:56:52 pm »

Steve,

What are you using for the spindle drive? I've got a Monarch 10ee on the back burner that I'm converting to CNC, mainly to enable it to do metric threading. Most of the stuff I do is 1 off and it's quicker to just sketch a print on a napkin and turn the handles than go thru the process of creating/debugging a program. Plus there's the wasted stock and tooling when things go wrong

. I've got a 3000 rpm - 3.5 Kw Mitsubishi servo motor and drive for the spindle. I was planning on using servo motors/drives from the same series for the X and Z, but they are $$$ on Fleabay. Going with steppers on the X and Z would save a lot of money but sacrifice some accuracy and possibly add complexity to the project. My day job is repairing CNC machine centers (and ancillary equipment) in a medium sized machine shop so I'm comfortable with servo drives. I haven't really played with steppers at all which also steers me towards going all servo. Since you're using steppers on your axis' I'm curious how you plan on driving the spindle.

Marc -

KG7AMV · « **Reply #17 on:** November 16, 2014, 10:25:53 pm »

I am just going to use a VDF and 3 phase motor. Changing controller board to a PMDX-126 with ethernet smooth stepper and placing this controller on the back burner for a while or till I get my old hot air system found and unboxed i have it some place!

I am still going to use the drivers and power supply's I was thinking of using a servo for the spindle but the costs exceed my budget so it is VFD for sure.

Last night I Pulled all 48 Fet's from the 6 Drivers I got for repair... Either lack of QC or they just used 4 mfg was a mix fet's per driver and an odd ball SFP9540 on 1 of the channels! But that board looked like it was reworked and poor rework job solder job was not a pro job at all for sure! The pins on one of the FEP16DT's was broken and soldered? New ones on order!

Brands of Fets.

International Rectifier
Fairchild
H (Unknown MFG)
ST

I ordered all International Rectifier IRF9540 and IRF540's for replacements Digging deeper have a couple Bad +12V 1 Watt Zener diodes on each board so since they are tied to the fets replacing them all. Found a few Burnt TIP30C's also.

I looked at the data sheets and from what was in it the N's that they are sending the spec's are similar and lower then the non N.

I think I am also going to mill out some heat-sinks for the drivers??? Not sure!

Falcon69 · « **Reply #18 on:** November 18, 2014, 05:11:09 am »

Heatsinks for the drivers is a MUST. Those drivers will get very warm under full load operation. I am using Gecko G203 drivers. I'll be using heatsinks with them.

on those boards you are trying to fix, I think you need to ask yourself two questions:

1. what caused the fets and other things to fry? and
2. what will keep the new fets and other replaced components from the same fate?

KG7AMV · « **Reply #19 on:** November 18, 2014, 05:40:52 am »

Quote from: Falcon69 on November 18, 2014, 05:11:09 am

1. what caused the fets and other things to fry? and
2. what will keep the new fets and other replaced components from the same fate?

Not sure I got the drivers 2nd hand on ebay the seller said they were driven hard.

The photo below is how they are mounted in the rack. For sure will need heat sinks all 3 drivers in this rack work!

Not sure why the mfg did not put heat sinks on the drivers that aluminum angle they are mounted to does get warm and has holes drilled to mount a heat sink! I will be rearranging them and stuffing 2 more in the box when i do that i will mill out some heat sinks.

KG7AMV · « **Reply #20 on:** January 25, 2015, 10:30:00 pm »

Small Video of the Lathe.

The driver on the Z axis was one i rebuilt with parts I had on hand.

Here is a Video of the Homing Should Work now.

Still have not got to deep in to the NM400 Controller yet? but will dive in to it in the future. I went with a Ethernet Smooth Stepper + A PMDX-126 Controller. And Rewired the Entire Drive Rack.

I mounted the drivers to the side of the case to use the case as a heat sink.

deephaven · « **Reply #21 on:** January 25, 2015, 10:45:39 pm »

Link to video doesn't work. This should: http://www.stevenrhine.com/wp-content/uploads/2015/01/lathe-homing.mp4

KG7AMV · « **Reply #22 on:** April 24, 2015, 05:31:26 am »

About ready to pull the MN400 back out of the bin and do some more looking. I ordered a few PS2801-4 optos from china been 45 days no receipt! They are resending because they cannot find the tracking info. Glad I put a sticky note on the board!

KG7AMV · « **Reply #23 on:** July 09, 2015, 07:27:01 am »

Wow, My PS2801-4's will be here next week I think.

KG7AMV · « **Reply #24 on:** July 14, 2015, 02:52:33 pm »

Love it when the mail woman delivers the goods a rail of PS2801-4's time to dig the board out of the closet and get to work.


EEVblog Main Site	EEVblog on Youtube	EEVblog on Twitter	EEVblog on Facebook	EEVblog on Odysee

Author Topic: CNC Controller Repair (Any Ideas) Has Me Stumped!! (Read 10105 times)

Share me