Hello,
I thought I'd start a thread about the Mach3 conversion of my 1986 Abene VHF-3 vertical knee-mill. I've mentioned this project in various threads here on CNCZone and I thought that it's about time I'd start a thread on the subject. The project has been ongoing since December '07 and is still ongoing so I'll jump a little between the past and current, I hope that's OK.
Anyway, I'll start from the beginning. I found the machine on an online advertising site, here's what the photo in the ad looked like:
What I "fell for" was the fact that it already had been converted to CNC by one of its former owners so ballscrews and servos already existed but it lacked the control - a perfect scenario for a retrofit in other words. After some discussions with the seller I made him an offer and he accepted. This was a bit of a gamble since I hadn't actually seen the machine in person but he felt honest so I went all in....
After quite a struggle with the freight company it finally arrived at it's current destination where it sat during the x-mas holidays '08:
Here's a random shot of the "back" of the Y-axis ways and ballscrew after some inital cleanup.
The machine has a 40 taper DIN2080 spindle, 4.5kW (6HP) with a 12-speed gearbox, dovetail slides on X and Z and boxways on Y. The ballscrews are 32mm*5mm rotating in dual tapered roller bearings. The servo-motors that came with the machine are some REALLY solid Indramat PMDC-servos, you can see one of them here, next to a little Maxon DC servo with a 3600line Scancon encoder that I initially planned to use:
The X- and Y axis motors are both of the same type, (MDC10.30D) rated at 19A continous, 150A peak, 0.47Nm/A while the Z-axis motor (MDC10.40C) is a little bit bigger, 24A continous, 200A peak and has an integrated break. I did get the original Indramat 3TRM2 3-axis servo amplifier with the machine and the original plan was to get some Pixie P100 boards from Skyko but I found out that they where no longer availble. Besides, the transformer for the original SCR servo amps was single phase 400V (well, two phase really) and I want to be able to run this machine from a 400V/16A outlet.
While looking at, thinking about and searching for suitable a drive I started the design of the powersupply. My motors are rated 2000rpm and to get that speed under full continous load I need ~110V at the motor terminals. Some (quite little actually) of the voltage is dropped across the drives internal resistance and a bit more due to the less than 100% modulation of the drive so I settled for 130VDC. Since not yet knowing which drive I would end up using I designed an energydump circuit that dumps regenerative power in order to prevent overvoltage:
This is a double sided PCB with 105um thick copper on each side, it mounts right on top of the two capacitors. In the background of the second photo above you can see the variac and three transformers I used for testing purposes. You can also see the Y-axis motormount. You might also notice that I replaces the original pulleys with new AT5 profile pulleys. The previous pulleys were mounted with keys and setscrews but I went for those "collet-type-thingies" that clamps to the shaft. (I can't find the correct english word for them at the moment)
Quite early I found that there was something wrong with the Z-axis on the machine. The knee is driven by a ballscrew and a 2:1 bevel-gear. However turning the shaft sometimes did and sometimes didn't move the knee - something was slipping. In order to remove the screw-jack for the knee I had to remove the table and then hoist the knee up as high as possible with an engine hoist:
I removed the screwjack and found the problem. The bevelgear was mounted to the ballscrew with a setscrew that was partially recessed into the screw shaft. The problem was that it was recessed too deep making it more like a small tube or pipe than a solid piece of steel.
I ended up drilling, reaming and mounting a solid conical pin right thru the gear shaft and screw instead, I hope that will hold up better that the setscrew aproach.
OK, back to the issue with servo-drives.... I'll simply make a very long story fairly short here and say that disregarded Gecko from the get-go due to low voltage, I thought about Rutex but didn't want to go there due to various reasons. I looked and the Viper200, I tried the Mammut from CNCDrive and the VSD-A from Granite Devices. Although I LOVED the VSD-A it simply couldn't supply enough current. When I'm writing this I'm using the HP-UHU and although the motors are still on the bench I beleive that this might actually work.
The motors, as stated earlier, didn't have encoders on them so I had to figure out a way to fit those. The back of the motor had the Indramat logo embossed in it so I took them off, chucked them up in the lathe and turned them "clean":
I then fabricated three shafts that fits in the internal thread at the back of the motors as well as three mounting adapters that would let me mount various encoders. The little bushing is an mounting aid to help align the adapter coaxially with the shaft when mounting it to the back of the motor:
The shaft where mounted and secured with LOCTITE and checked for run-out:
Due to the limited bandwidth of the UHU-chip the nice 3600lines Scancon encoders I had couldn't be used. I ordered up a set of USDigital E7P's and ended up spending another 4+ weeks chasing a noice problem before finally scrapping that piece of s**t and ordered a set of Renco R35i's. With those mounted everything just worked. Here's an overview and closeup of the back of the motor with the R35i mounted:
The last photo in this post shows one of the aluminum housings for the encoders that I machined. I made these while the plan was still to use the Scancon encoders so they are a lot longer than needed for the R35i's but it doesn't really matter as there's plenty of room. The groove was originally meant for an o-ring but it turned out that particular dimension, although listed in the catalog i looked in, was a "special" so I made some seals from rubber sheet instead.
OK, that's it for this first post, more later.
/Henrik.