Another Hurco KMB-1 converstion (this time in Sunnyvale, CA)

Strange... This forum site seems to be very 'buggy' with my Mac browsers! :( Anyhow...

I thought I'd start a build log regarding my Hurco KMB-1. I've had it for about 10 years. It started as a B control and I upgraded it to a BX. I'm an electrical engineer and have chip level schematics so maintaining it was never a problem (fairly simple design actually). However, it has never been the most stable machine. There are *so* many connectors in there that it would go through phases of shutting down. Looking at the signals with a scope showed enough noise that I knew the problem but didn't want to fix it. Plus, I have always wanted to dabble with a 4th and 5th axis.... So, I finally bit the bullet and yanked the control a week and a half ago. It was a form of catharsis, actually!

I am planning on going down the following route:

• keep computer in the main cabinet, not the control pendant (which will just have the monitor and switches)
• keep all servos and encoders (may lower CPR of encoders, if required)
• keep brake system controlled by computer
• keep mist/flood option
• 17" touch screen for input (if I don't like it, I can always swap it out)
• E-Stop chain that is a *real* chain... so, it will cause an open of a relay controlling servo power, not be 'interpreted' by the computer
• Linux CNC as a control
• Mesa 7i77 interface card
• 3 Granite Devices VSD-E boards running in analog mode (so, full feedback to the control software, not step/direction)
• 2 Electrocraft DC-25 amps for 4th and 5th axis
• wiring for 220V single-phase
• dumping original transformers except for the 60V servo transformer

Although I work on hardware/software all the time, machine tool design is still a little different. Part of my motivation here is to get ideas for a retrofit of a Deckel FP3NC that I also own. That is a totally different beast though. It has a logic chain that would make a person cringe as it shifts speeds, etc. More on that later though.

I shall have to post pictures of my machine, although my garage is a TOTAL mess and I barely have any room to work. I'm more than a little embarrassed by the though of "showing my underwear" there, so to speak. But, I'm hoping people will have more interest in the machine retrofit than my #*$% so they'll look the other way and not rib me. Plus, the cool thing is that all the junk is actually not junk at all. Most of the boxes hold far too cool machining tools or other stuff. :-)

Alan
Sunnyvale, CA

Looking forward to it Alan

Tim

Here are some photos of when I first got the machine some 8 or 10 years ago:
Attachment 233560
Attachment 233562
memories.... she was so young then!
Attachment 233564

And moved into its home in the garage (very tight):
Attachment 233566

The guts of the old machine (main cabinet, not VFD):
Attachment 233568

Where I am today:

This is where I'm putting the main guts of the control, which will be Linux CNC. Looking at the cable channels as delimiters, the upper left will have fans along the left side and then 3 Granite servo amps and two Electro-Craft amps. The servo filter caps are along the top. Upper right has the high power terminal strip, associated relays, and the interface relay board (computer controlled), driven by the Mesa 7i77 board that will sit below this. Lower right houses the main CPU board and the power supply (mounted and covered in tape) below that. Lower right will have the SSD for the computer, the lower power terminal strip and the encoder/tach/limit terminal strip:
Attachment 233570

My machine had a fan in the door of the cabinet added but no real exit for the air (duh!). I've added this in the lower right side, as shown below:
Attachment 233572

The control cabinet is coming along well... If you look carefully, the standoffs are already mounted for the CPU and the 7i77 Mesa interface board. I'm presently working on setting up the fan bank for the amps. I'm also drilling some holes for the connectors for the 4th and 5th axis (BTW, one set of pins will be devoted to energizing a relay for each of those axis. In this fashion, the amps will remain unpowered when no axis is connected).

Long ago I removed the Parajust VFD and replaced it with a GE unit. I'm rewiring all of that, hence the reason the front is off, but the general idea is shown below. The relay will form the main control relay for the control power (namely energizing the servo amps, etc). The coil will be latched in via a chain of switches in series, any one of which will trip it out and immediately stop the movement of the machine (but not loss of the encoder position, etc) by virtue of the fact that the servos will lose their drive power. One of the switches in the series will be a relay under control of the computer itself so that it too can issue an E-Stop if it senses a fault requiring such action. Pretty standard stuff, but there seem to be several people that believe software control is sufficient for E-Stop... I'm passionately not one of them! I believe in a fail-safe drop out of the main power that is completely independent of the computer (although the computer can participate because, again, it has its own switch in the chain):
Attachment 233574

For those of you with sharp eyes.... There are only 2 fuses in the main disconnect switch. Furthermore, the white leg, which has no fuse in it, is not taped off to a different color. That is because after I acquired the machine, I changed it over to single phase.

Although I have a Phase Perfect converter, I saw no benefit in running this machine with 3 phase when it would work fine as a single phase device. Ironically, even the original Parajust drive was running with only single phase. So I removed the isolation transformer from the circuit (which is now being physically removed too during the retrofit) and rewired accordingly. I had never needed the traditional neutral leg, however, so I 'parked' it on the disconnected. I'll be changing that now because I will be using 110V in the new system.

Yes, the original machine had a 220 to 110V transformer, but using it seems kind of silly when I have a neutral available. In fact, the way I've wired the power distribution in my garage, I have three types.... 220V single phase that has the neutral available too, 220V three phase (again, with neutral available... so a 5 conductor twist lock), and 380V three phase (driven with a 5KW transformer, no neutral here because I can't get the 110V easily). The 380V is there because I have an Emco Maier EmcoTurn 120 that utilizes this power. I also have a Deckel CNC mill that will eventually replace the Hurco which is also 380V.

Alan

Hi,
Is there a way I could obtain a copy of the KMB1 chip- level schematics? I am working on restoring a KMB1 with a ATC.
I am an electrical engineer with a lot of experience getting old equipment running.

BTW, I agree with your E-Stop philosophy. I have seen a few control CPU's hang on execution, but with the servos still engaged and racing to their stop limits....

Hi,
Please send me a private message with your email address. The machines are all based on the same basic board set, although I am intrigued.... I didn't realize they made a KMB1 with an ATC. I thought that the ATC was something limited to the MB1 series, which is a bigger beast.

Best regards,
Alan

There was a KMB1 with a toolchanger, a KMB1X with a toolchanger and also a KM3P with one. The KMB1 was a bizarre thing with a nutty exchange arm mechanism that had a GAG cylinder that jammed a pin in the arm rotate mech to stop it in the right places in the magazine and a fourth servo to rotate the chain carousel.

Good work on the machine so far. I was wondering why you are going with the granite devices VSD-e instead of their latest and greatest Argon servo controller. I know cost may be a factor, but is there another reason? I am considering upgrading my KMb1-M and thought that the cost difference was low enough that the extra features may be worth it.

I need to update the pictures as I've gotten way further now. In fact, I'm presently working on the servo amp installations. Which is a good segue into....

I bought the VSD-E drives because they were economically priced and appeared to have a good bang for the buck. The Argons were not out at the time and, even if they were, would probably have been outside my price range. On paper they look very nice, but they are getting into a range where there is more competition from others and, more importantly, are getting well beyond the budget of the average home builder who needs to buy 3 or more drives.

On paper the VSD-E drives look like they will be very nice. In practice, they have been a nightmare so far to setup. I'm really not sure what is going on so I'm not ready to blame the drive at this point. It could well be the FTDI USB configuration cable, but I can't get a reliable sync to the drive. This makes tuning very unreliable, to the point that sometimes the tuning software sees the encoders and sometimes it does not. <sigh> Now, *if* this is all just some strange cable problem and the drives are not at fault, then the drives will, I think, be stellar additions. If not, I'll have to buy something else. But what does worry me is that there do not seem to be tons of people using the VSD-E drives and that makes support spotty. There are two folks on the Granite site trying to help me now, but so far we have not made a great deal of progress and what I really need is to get help from Granite itself, I suspect. However, they have not responded to the posts or to a direct support request. This may be related to some vacation, I really don't know... I certainly hope that is the case, though, because if it isn't then I have some expensive boat anchors and I'll have to find something else that will work. I am not sure what those would be but I'd also be frustrated because of the cost and the fact that I'd have to change my system setup. Personally, I'm hoping GD will come through with help. :)

Did you keep the servo drives originally in the machine? I did an EMC2 (still called this at the time)/LinuxCNC conversion on a KM-3 and kept the Servomate drives and electrocraft motors. It worked really well. Maybe you can continue progess with the old drives until you sort out the VSD-E problem.

I also bought some VSD-E drives a few years back, but sold the machine I thought I was going to use them on. Never unpackaged much less wiring anything up. Maybe its time to bench test them. Are you working with just a single drive and motor on the bench for troubleshooting? Did you buy the mount or backplane, IIRC?

I went with an "all in" type of approach, and pulled the WestAmp/ServoDynamics drives out. I still have them so I *could* put them back in, I guess, but I didn't lay the system out for that. I'd have to see if they fit. Plus, my intent was to move to a new system that was digitally tuned. I'd really like to move forward with that.

I have chronicled my woes with the drives in the following post:

http://www.cnczone.com/forums/granit...160-drive.html

But, in summary, I have three drives installed. I have a motor on the bench so I don't crash the machine, but I've also tentatively tried with the axis as well. Each VSD has a different issue. I tried a hodge podge of variations early on in an attempt to establish some type of patter. I got absolutely nowhere except to become very frustrated. Having failed to find something that would give me an entry into something working, I've moved back to a single drive, single axis and am trying to dig into that. I just purchased a new FTDI cable and the problem persists with that so I can now effectively eliminate the USB cable and computer. That is bad. I was hoping it was one of those because they are easier to deal with... <sigh>

123CNC: Would you like to sell your drives? If so, please send me a message. Thanks. I tried sending one to you but I was unable to do so (you are blocked)

I figured I'd try one last time.... Would you like to sell your drives 123CNC?

BTW, I didn't keep the original drives primarily because of real estate. I wanted to have room for 5 amplifiers in the chassis so that I could add a trunion table.

I think that I've licked the VSD-E issue. Noise was a huge part of it, which required adding some caps. That seems to have alleviated a LOT of the issues getting a stable USB connection to the drives, which means I can now work on tuning them. One of the drives, however, was bad. It was unable to see the encoder pulses at all. I repaired this, although I have to admit that I hate working on SMT designs without an SMT rework station. It is a pain!

Anyhow, again, I *may* be interested in the drives.

Quote:

---

Originally Posted by MrMetric

I need to update the pictures as I've gotten way further now. In fact, I'm presently working on the servo amp installations. Which is a good segue into....

I bought the VSD-E drives because they were economically priced and appeared to have a good bang for the buck. The Argons were not out at the time and, even if they were, would probably have been outside my price range. On paper they look very nice, but they are getting into a range where there is more competition from others and, more importantly, are getting well beyond the budget of the average home builder who needs to buy 3 or more drives.

On paper the VSD-E drives look like they will be very nice. In practice, they have been a nightmare so far to setup. I'm really not sure what is going on so I'm not ready to blame the drive at this point. It could well be the FTDI USB configuration cable, but I can't get a reliable sync to the drive. This makes tuning very unreliable, to the point that sometimes the tuning software sees the encoders and sometimes it does not. <sigh> Now, *if* this is all just some strange cable problem and the drives are not at fault, then the drives will, I think, be stellar additions. If not, I'll have to buy something else. But what does worry me is that there do not seem to be tons of people using the VSD-E drives and that makes support spotty. There are two folks on the Granite site trying to help me now, but so far we have not made a great deal of progress and what I really need is to get help from Granite itself, I suspect. However, they have not responded to the posts or to a direct support request. This may be related to some vacation, I really don't know... I certainly hope that is the case, though, because if it isn't then I have some expensive boat anchors and I'll have to find something else that will work. I am not sure what those would be but I'd also be frustrated because of the cost and the fact that I'd have to change my system setup. Personally, I'm hoping GD will come through with help. :)

---

Any updates on your build I am in the middle of retro fitting my machine and would like to see the progress you have made on yours.

Yes, as a matter of fact... I got *very* slowed down due to issues with my servo amplifiers. However, I think that those are resolved now. So, theoretically, I have the major components installed and the next thing is to tune the amps. I've done two of them but still have to do the third... I'm also wiring the machine for 2 extra amps so that i can do a trunion table in the future. I'll see if I can post some pictures this weekend. It is progressing well, I just kind of lost steam with the frustration on the amps.

Yep I understand. I have started on my build and cant figure out which way to go there are a lot of options out there and I keep going back and forth on making a decision on how to proceed. I have installed a TECO7200 5hp VFD got it running but am having some issues with it trying to sort it out.

Thanks for the post and will be looking forward to seeing some pics.

Here is the cabinet. The fundamental approach in the layout was to have four basic section that have cable channels running between them. Everything in the following photos is opened up and very messy (some temporary wires are connected, etc). Remember, this is work in progress. The idea here, however, is as follows. The upper right has the relays and the high voltage section. The upper left is also high voltage (see a pattern here... high voltage is at the top) but dedicated to the servo amps, etc. On the far left you'll see a bank of fans that cool the servo amps. From top to bottom you have X, Y, Z, A, B. All are mounted on their sides. I'll be cutting fins on the VSD-E amps eventually. The bottom two amps (only one installed but not hooked up) is an AMC amp. When the bottom two amps are not running (because the trunnion table is missing), the fan will likely not be on there:
Attachment 255356

The relay section on the upper right side. The terminal strip is the high voltage. From the top down you have hot, switched hot, 120VAC system relays (not controlled by the computer), fuses for the 12V and 24V supplies, and grounds. On the right you have the relays that are controlled by the computer. As is the theme, the high voltage outputs are on top. :
Attachment 255358

Not the best photo, this is below the terminal strip and description of the last photo.... Separating this strip/relays and the "stuff" below is a white cable tray. Below the tray is the Mesa 7i77 interface board, the main control point between the CPU and the rest of the system. Tucked in to the lower left of the Mesa board is a 12VDC to 24VDC power converter because I needed a small number of power there:
Attachment 255364

The receptacle is switched and feeds the CPU power supply. Under this there is an isolation board for the limit switches and, finally, the computer SSD. The terminals here are all low voltage: 5VDC, 12VDC and 24VDC. There are also terminals dedicated to the encoders and limit switches.
Attachment 255360

The computer (lower left), with the power supply underneath:
Attachment 255362

Not shown:
On the back of the cabinet, I have a Hobbs hour meter, ethernet connector, and I will be mounting an external WiFi antenna. On the front face I have two military connectors for the trunnion table. In the rear cabinet I removed the isolation transformer but kept the servo transformer. I did remove about 5 coils, as I recall, to lower the voltage a bit though. I also have the full wave rectifier for the servo power. In the opposite cabinet I have the VFD and a relay (redundant, actually, that feeds the VFD power).

The only change to the spindle area is that I ran an ethernet cable up there that i can someday use as the feedback for a spindle encoder. That way I can set the system up to do rigid tapping if I so desire.

Clearly, as stated, this is all work in progress. Once everything gets buttoned up, the wires going here and there will be gone and it will look much cleaner. However, I think you get the idea. :) I promise to take better photos sometime later.

The pics look fine didn't you use the granite devices for your drives ? Why did you pull out the original drives and replace them ?
Are you using Linux to run your machine ?

The progress you are making looks great I hope to get some hardware ordered this week for mine and maybe start putting things back together next week.

Still udecided which way to go if I keep my drives I think I am leaning towards the Vital Sytems ? if I purchase new drives I think I will go with the Viper 200F drives and
PMDX hardware ?

Have to decide soon so I can get things under way.

Thanks again for your post.

Terry

My XYZ are all Granite Devices VSD-E drives (two 80V, one 160V). The trunnion table drives are AMC... I replaced the original drives for a few reasons. First, I needed the space. Second, I thought it would be nice to be able to digitally tune the servos amplifiers; in retrospect, that was not all that big a deal. Third, I wanted to move between EMC and Mach should I decide I like one versus the other. Mach (regrettably) only supports step/direction so the original analog drives wouldn't work well there. However, the Granite can work in either mode. Finally, I have very good hearing and the origin all SDC/Westamp drives used to grate on my nerves because the squeal. I thought I'd try a different drive that had a higher chopper frequency. From the tuning I've done thus far, it seems like this was a good move. Mind you, however, there is a lot of general noise still from fans and such, but not so much of the high frequency squealing.

Yes, as mentioned above, my primary interest is in LinuxCNC, so that is what I am anticipating as my platform. If I move to Mach, I'll need to change to a different interface board.

I talked, briefly, with the Viper developer/owner. His drives look nice and he seems to provide good support for them. I don't think you'll go wrong there. For me, I kind of liked the versatility of the Granite drives in that they could run in many different modes. The Viper owner didn't feel as though there was a market in pursuing that type of flexibility. In truth, that is probably true. The market will define what you spend your time on developing. However, in my case, I actually liked the ability of switching from analog to step/direction.... Plus, I had the drives (although I had a good number of problems getting them to work correctly).

Having had to repair my drives, I spent some time looking at the design. When you get right down to it, there were not a lot of surprises. Like most things, the drives have migrated to a world where most of what you are getting is in the software. A microcontroller has some interface connections (encoder input, configurable I/O, LED outputs, analog input/output, serial, etc) and control over a HV section (the MOSFETs). The Freescale processor works very well for this. So, the hardware isn't exactly terribly complicated anymore. While there is definitely analog in there, a heck of a lot of it has been moved to the digital world. Again, no giant surprise there because this is true in pretty much anything now. The point being, though, that "support" for features is generally more an issue with whether or not a developer/manufacturer wants to invest the time to code the functions, and less about whether or not the hardware can support it (or it may take very few additional components). Market needs ultimately drive whether or not the effort is made... But I suspect that on many drives right now, the hardware is all pretty good and a decision on what to use is more a function of features/support from the manufacturer.

I'm not really familiar with the PMDX hardware. I looked briefly at it, but I decided to go with the Mesa. Again, though, I think this had more to do with what I wanted to drive my system (LinuxCNC) and an ability to take whatever hardware I chose and adapt it fairly easily should I decide to move to Mach.

Alan