Here's You Sign #98 - Pennsylvania USA

[danilom]

Don't kill UVW ! if you need a dead motor then mount the switch on phase inputs to VFD

[pblackburn]

I am assuming to minimize the risk of switching off the vfd prior to it spinning down and possibly frying the vfd.

[Surfcnc]

Here is an in depth discussion - http://www.practicalmachinist.com/vb...er-vfd-102341/

Short answer as a stand alone means of cutting power to the spindle switching the feed wires to the spindle = not a good idea.
The longer answer is basically what Danilo says, then you can put a switch or contactor between the VFD and the spindle if you want after that.

Regards
Ross

[zumergido]

I thought that de vfd have some electronic protection for power cuts. mine its a cheapi cheapi ls vfd and says power low for a while when i turn of the machine. i will add the contactor to the spindle for shure after reading this.

[pblackburn]

A vfd does have protection in case of total power loss but there is the possibility of problems with opening the uvw when it is decelerating or running. However this may happen it, I have never seen it. I have seen two legs open on a motor before and only have the vfd go into protection mode. As with most things, it depends on the manufacturer and quality of the vfd. I know AB, Woods(now Vacon) have this protection but I do not know about Hitachi, Delta or some of the others used with these spindles. If in doubt contact the manufacturer technical department and they will let you know for sure.

[pblackburn]

Another Electro-Mechanical tech I work with, his brother was working on a vfd controlled motor. When someone hit the reset he lost three fingers and part of his hand. Some things you can not replace or wish later on that you took more time or precautions. Obviously my router had a kill switch on it that I always used. I will let you know what Hitachi advises on their drives. I have seen disconnects placed before and after the vfd. Is it a good practice, as Ross and Danilo have both said, No. But sometimes it needed.

[pblackburn]

For the time being I think I will break the drive enable circuit. You can break the output of a vfd at idle. The problem comes if someone does it when it is running or decelerating. If I broke the enable but had deceleration setup the output is still energized. Since I am the only one using the machine, I did not see this as a problem but what if.... Someone else may not understand how this works and could possibly damage the unit. At the very least I would have to disable drive then break the output if and only if the drive was set to coast to stop. Not very feasible.

[danilom]

When you command a VFD to stop, it ramps down or breaks in seconds
when you kill the enable while its spinning, the spindle will keep free rotation until it looses its inertia, same with breaking the drive circuit.

So in emergency its better to use a VFD STOP input so it breaks and not freewheel.

[pblackburn]

Alright, talked to Hitachi about my drive. It is a WJ200. It has a the capability for a safety circuit to be added. It uses GS1 and GS2. From the manual..."By pressing the emergency stop button, the current to GS1 and GS2 is shut off, and the inverter output is shut off. By this, motor is free-running. This behavior is according to the stop category 0 defined in EN60204."

Since this is the behavior I am looking for this will be the direction of approach. You have to use a reset to enable the drive. Since this is to be used at a tool change the deceleration has already occurred so the free running aspect is not an issue.

[pblackburn]

Have all the parts except the spindle mount and one relay for the alarm on the water flow. Machinist is busy. Well I can start the wiring and try to figure out the modbus and mach3 settings. Probably will not have the mount for another week but I do have a nice mount for a PC7518 now.

[pblackburn]

Got the cable track laid out again to fit the new hoses and vfd wire. Mounted the box for the vfd. Used an old industrial computer case and modified the back. Put aluminum on for the back plate then a one inch space then aluminum mounting plate. Used clear plexi for the cover with a cut out for access to the buttons but not the covers. Have two fans for additional air flow if needed. With the amount of dust that will be reduced from not having the massive fan from the pc 7518, the dust collector should work even better. The vfd wire is routed to the new enclosure and the water sensing wire is run to the main panel for the alarm.

[pblackburn]

Wiring almost complete. Need to update my wiring schematic to reflect changes. Still waiting on the spindle mount. One part is ready but the other two are not. I am not sure which will be easier. To program from the keypad or using the software. I have always done by keypad but have never messed with modbus. This will be a learning experience I am sure.

[pblackburn]

First smoke test complete. No Smoke left out.

[pblackburn]

Have tested the spindle now manually. Have some stray voltage floating on the case of the motor when not mounted. Dissipates to zero when mounted. Modbus is giving me a fit. I keep getting Timeout Received. Ordered a different cable to try with the rs232 - rs485 converter. Will keep you all posted and post when and if I figure the modbus out for the WJ200. Not sure it I can send control word, have seen bits here and there about only one control at a time. May have to use brain instead of script editing as I originally planned but first I have to get it to handshake.

[danilom]

Check the Slave ID or if the modbus is disabled on drive, also decrease UPDATE Frequency in Modbus setup to 10hz

[pblackburn]

Danilo,

I am assuming the slave id is the modbus address of the drive, is that correct? I did try to decrease the update freq but returned it to mach original setting of 50. I will move it back down. The cable I had was short so the length of the SP+ and SN- were long. I saw on machwiki that windows does not like 2 wire 485 so I opted to get a longer serial cable. This decision was based on all the videos on youtube show a very very short 2 wire connection from the converter to the drive. I am hoping to be able to change the m3, m4 and m5 scripts when I get it to communicate. I have never used modbus before so this is a steep learning curve for me. I opted for the WJ200 for the dynamic braking capability. I have not stumbled across a good reading source on modbus yet or Mach3 setmodoutput command explanation yet.

[danilom]

it should be quite the opposite, rs485 can be run up to 300 meters and is immune to noise due to the signal being differential.

My adapter is directly transforming the signal in the db9 connector and outputting A B signals out.
You can see it here
http://www.mechmate.com/forums/showthread.php?t=3411

[pblackburn]

Well the modbus can talk now and don't ask me why using a longer serial cable worked but it did.

I did find another post with someone else using a wj200 and they had to set up using a brain. Another headache to try and resolve and understand. Modbus is a bit above my level of understanding yet.

Now onto the next problem. I had this before with using just a router but it was very random. External Estop Requested on Mach3 would show up on occasion. With the spindle it is more frequent. I assume it is noise but unsure on which cable. I think it may be one of the limit switch cables. All other cables are shielded.

[zumergido]

on youtube are some samples of brains use. and the mach3 page have some videos too.
http://www.machsupport.com/videos.php
its a logical program.
i have made a modbus controller for the mach3 to me looks better than lpt port that i use only for the motors drivers.

[KenC]

Pete, the E-stop issue is definitely VFD noise. The 1st thing I'll do is to inspect if your star ground is implemented correctly.
2nd step is to buy plenty of snap on ferrite; snap on all the VFD related cables... you may need to snap more than one in each cable.
Or, change all the VFD cable to shielded cables...
Lastly, use RFI EMI line filter on your supply line to the VFD.

[smreish]

...also change the debounce time of your system to something a little higher.
It's in Mach under Config, General Logic Configuration, (upper right corner of page)

[pblackburn]

I have ordered snap on ferrite cores plus zero phase reactors for the incoming and motor leads.

I saw about the debounce in other forums. It sounded from the posts that it was mostly for startup noise problems. I will check it out now.

[smreish]

I had one machine #28, that required an isolation transformer between the VFD and Mains to completely eliminate my noise issues.

Chasing your issue can be troublesome. Keep at it.

[pblackburn]

While I am waiting for the zero phase reactors to arrive Monday. First and foremost, Thank you to all you Veterans. Getting hold of Hitachi a second time is almost impossible it seems. From other posts I have read, the WJ200 only likes hex. The book says it can accept words but lack of information is the item here. I am trying hard to grasp this but I feel wanting. Is there anything anywhere about the settings of modbus with mach with a better explanation of how and why it works? Any links and advise would be appreciated. I have googled myself sick already.

[KenC]

here you go,
http://www.homanndesigns.com/pdfs/Us...with_Mach3.pdf
#ps, Brain included.

[pblackburn]

Thanks Ken. I read that before and it did not stick but this time it made more sense.

[pblackburn]

I went through all the grounding and for just in case I even grounded the common of the 24VDC. No help. Tried to increase the debounce and it seems to help a little. I even pulled the miniature relays that work with the alarm for water flow and work with my timer to allow the dust collector to stay on for 20 minutes after the spindle stops running. Still External Estop Requested.

I did find a brain for the WJ200 posted on MachSupport. After looking at it and modifying it to work with my machine, I finally can send the correct frequency to the VFD. Will post the brain and screenshots later.

I was looking a line reactors and came across an interesting article on practicalmachinist.com. The short of it was one guy made a filter with nothing but iron pipe. Crude but it worked then later in the article another post said to use a transformer core and wrap your incoming leads around it four times. Well I had a bad xfmr so I tried it. Amazingly it did help so this give me hope that the Zero Phase reactors on the supply and the motor leads will reduces my noise significantly. With just the core on the supply I have found that the lowest I can run is 14000 rpm. Anything lower will cause the annoying External Estop Requested.

The debounce is my concern. I did not increase it much but I have read it can mess with you limit switches. Is this true?

[smreish]

how much is much? Is your debounce setting like 1000 or 10000?

If you OEM switches are wired and working then you will be able to test this. (that's the programed 1000 and 1001 codes to resolve your hard buttons to Mach 3 interface)

What you will find is the higher the debounce the longer button push required on the green - start or yellow - feedhold buttons. If your debounce is to high...it might take a 2 second push-and-hold to get that command to be resolved. It's a happy medium, but in reality, it should be set as low as possible and keep the fault at bay.

Also, if you have proximity switches, this is a very usual place to pick up interference too. I found that my z proximity switch was the noisy antennae. Once I isolated that circuit, the e-stop issue for me went away on machine #28.

[pblackburn]

I will have to try and break down where the problem is. Noise can be hard to follow. I will need to get the zero phase reactors installed first. I had to change the carrier freq to the motor. Motor running smooth now but this adds noise. I feel like I am in a never ending cycle. My debounce originally was 0. It is now 100. If memory serves me correctly, what I read was about prox switches not working and the machine hitting the hard stops because the debounce was high. I think it was on cnczone not that it matters.

[pblackburn]

Debounce at 200 I can run 8000rpm