[andrew2085]

I've had my mill for about a month and I've finally ordered everything I'll need to convert it to CNC. I'm using the cncfusion deluxe conversion kit, kelling 570oz/in steppers, KL-5056D drivers, and I have a very beefy 48V 57A power supply that I got for $30 on ebay.

I'm still waiting on all the parts, so I wired up a 240V receptacle near my machine for the PSU. It's "48V" but really ~51.4V. The KL-5056D drivers specify 50V max. I'm sure they'd be fine, but I'd be more comfortable if I could lower the voltage to around 48V. I took it all apart and managed to find a pot covered in silicone that let me lower the voltage, but only down to 50.5V. It's a 1K pot. I removed it and soldered on some wires that run out of the case and attach to a 5K pot with a dial. This lets me adjust the voltage from 45V to 53V, I just leave it at 48V.

I don't have any machining experience, but from messing around adjusting the gibs and using a dial indicator I can tell that the fit sucks. Even if I tighten them so hard that the table is hard to move, I can pull up on the table and see a few thou of movement and see oil sucking in and out of the ways. I broke down the mill and I could feel that the fit is very inconsistent along the length of the table. I ordered a cheap scraping kit and a 9x12x3 granite surface plate (biggest without painful shipping charge). I also got some Dykem blue hi-spot and 400, 600, and 1200 grit lapping compound that Brian linked in his build thread. I've never scraped before, so I might just try to get the gib to fit right and very lightly lap everything else to be a more consistent fit.

[SCzEngrgGroup]

Am I understanding this correctly? You open up a complex, server switching power supply, find a sealed potentiometer, and *assume* it's there to adjust the output voltage? Then you modify the circuit by adding another potentiometer, and take the adjustment even further? You have absolutely no idea WHY that pot is really there or what it *really* does. While it may slightly modify the output voltage, it is exceedingly unlikely that simply adjusting output voltage is it's primary function, as modern power supplies do not generally adjust that way - they don't need to because they typically have internal precision voltage references to compare their output voltage to. And, modern electronics will NEVER contain calibration adjustments unless they are critical to the safe, reliable operation of the device. You have now changed a factory adjustment, having no idea what the effect of that will really be, apart from the superficial change in output voltage. Don't be surprised if you find out down the road, in a most unpleasant way involving lots of blue smoke, that was not a very good idea.

Regards,
Ray L.

[bjones]

nice work andrew, do you have any pictures of which pot you replaced and where it is on power supply?

[andrew2085]

HimyKabibble: I did some research first and found that other people were replacing this pot in order to lower the output voltage. I gave it a little stress test afterwards at about 30A for a few minutes and all seemed fine.

bjones: I meant to take a picture to share but my phone died. I just found this page though with a bunch of pictures of it. This really would have helped me, it took a while to find that pot because it was covered. 48v 60amp strømforsyning

[hoss2006]

Thanks for that too Andrew, I have the same power supply.
Hoss

[andrew2085]

No problem.

After I get this all put together, I think I might have a use for this beast. Its rated at 6500W and is 130 rpm/volt. I suppose I could use the power supply I have or get another. There are pretty cheap high voltage RC ESCs rated for 50.4V. It would also be pretty easy to program a microcontroller to convert mach3's PWM to servo pulses to control speed. Some day

[andrew2085]

All of the electronics arrived yesterday and I've started getting everything put together. I was going to make a separate enclosure for the electronics, but I've found that they can all fit inside my PC. I've mounted the motor connectors on the back of t he PC and there will also be a 48V plug that my power supply will plug into. I'm mounting the breakout board and drivers on a sheet of aluminum that will in turn attach to the removable side panel of the pc.

[andrew2085]

Almost done. I had a few dumb mistakes slow me down, otherwise I think it would've been done this weekend.

Friday:
To install the Y screw I had to remove the ballnut. I let the plastic tube holding the balls in slip out part way dropping balls all over the place. I found most of them, and spent a couple hours trying to put it back together which I did (incorrectly). After I got the screw installed and was testing it with the motor, I heard about 10 balls fall under the base of the mill. I took the screw off again and brought it and the balls home with me (mill is in my father's shop). I managed to get the balls back (possibly incorrectly, too easy).

Saturday:
X and Y screw installed without incident. backlash was 0.004" on the Y and 0.002" on the X. I have a feeling the Y is higher because I'm missing some balls in the Y, and I'm not sure that I even put It back together right (Sunday's efforts further reinforce this).

Sunday:
The angle on my Z gib is way off, and I have a free replacement coming from Grizzly. For the hell of it I scraped the gib I have to fit. I had already got it much closer to fitting right by lapping on a surface plate and 1000 grit sandpaper while putting 75% of the pressure on the thick side of the gib. I used Dykem hi-spot on the mating surface of the column and slide and pushed the gib in then lightly tapped it with a small hammer and screwdriver. I've seen scraping videos on youtube where the spots are kinda hard to see, but this stuff spots awesome, very dark blue spots even with a very thin coat of the dye. I'm guessing I had to take maybe 15-20 passes to get it to fit nicely. It started getting a lot slower towards the end, and I felt it was good enough so I quit. by the time I had it fitting right the gib had to be inserted about 1/2" past the maximum depth the screw can push it in. I cut off about 1" from the thin side and ground a V onto one end, and a corresponding "female" V in the thick side of the gib. I used that little piece to put between the gib and the screw, it worked great. The slide is much smoother now and has no visible play. Before I did all this I could tighten the gib so tight that the slide couldn't move and I could lift up on the bottom side of the slide and there was at least 0.010" of play.

I got the Z screw installed and then it happened. I ran the ballnut part way off the screw, balls everywhere. I took the Z screw back off and brought it and the balls upstairs and got comfortable on the couch in front of a coffee table. These ballnuts are the type with 4 internal recirculators. They're ABBA 16mm x 5mm pitch. After a few hours of dropping balls and lots of swearing, I feel like an expert at putting these things back together. I had to actually take it off again and reload it a second time because I had too many balls in one of the circuits. I could tell something wasn't right because of the sound. No big deal though, by this time I had the thing back together in 5 minutes. You can watch all the videos you wan't about how to put them back together, but its something you really need to sit down and do to learn. There will be lots of swearing, but you will eventually get it right. I'm not even really going to try to explain how, but I get one circuit installed using the provided black plastic tube (with the hole plugged so balls can't fall through) then thread the screw into that first circuit, then Install each circuit one by one. A pinky finger and paper clip were essential. I definitely lost some balls. I think there are supposed to be 16 balls per circuit. I ended up putting 16 in the two circuits on the ends and 13 in the middle circuits. I think I put 17 in one of the circuits the first time which was causing the bad noises.

I'm sure I'm missing a lot of balls in the X screw, it went together way too easy and I didn't do it correctly. I think that's why there is 0.004" of backlash on that axis. Also when I was putting the Z ballscrew back together while I had the seals off, I could just let go of the ball nut and from its own weight it would spin down the entire length of the screw. I think that means I can fit bigger balls in there to lower the backlash even more. Each nut has +6 or +0 etc engraved on it, the one with +6 I measured at 0.1256". I guess that number means tenths over nominal. I'm not really sure how far over to go to get a better fit 0.0001", 0.0002"? I'm definitely going to try because I don't really wan't to run the mill with balls missing in some of the screws.

[andrew2085]

I ordered some new balls, 0.1253 and 0.1260. Now I can try over-sizing the Y by 0.0003" or 0.001", and the Z by 0.004" (stock Y balls are 0.1250" and Z are 0.1256"). Hopefully at least one combo should work.

[andrew2085]

The 0.1253 balls worked great for the Y axis with 0.1250 balls originally. Its not tight but snug, if you give the nut a spin it turns about 2 rotations and is smooth throughout the length of the screw. It brought the Y backlash from 0.004 to about 0.0015, and that's with love-joy couplings. I have some helical couplings on the way. The 0.1260 balls were too big for the Z axis with 0.1256 balls originally. I can get 2 of the races loaded and its too tight to get to the 3rd and 4th. the only other size between that and 0.1256 from toolsupply is 0.1257, which I guess will have to do. I also tried alternating 0.1256 and 0.1260 and it was still too tight, and its really hard to do with these internal recirculating ballnuts.

[arizonavideo]

Bal-tec has balls in more exact sizes.

Bal-tec, Manufacturer of Precision Balls, Ball Bars, and Kinematic Components

The do cost more.

One way I re packed ballnuts was to use heavy grease and roll the balls in it and grease the inside of the nut lightly. Then with a bent wire fill the return lines and then stick a row of balls in place and insert the plastic plug as you go.

I could fill the whole nut most of the time with no plug but they were not too deep.

[Fastest1]

You are replacing the Lovejoys with helicals? Why? My understanding is the lovejoy is a much stronger joint, tolerates misalignment better and doesnt wind up and or snap. I run oldhams. The helicals are fine in my Sherline but not sure I would use them in my G0704.

The 0.1253 balls worked great for the Y axis with 0.1250 balls originally. Its not tight but snug, if you give the nut a spin it turns about 2 rotations and is smooth throughout the length of the screw. It brought the Y backlash from 0.004 to about 0.0015, and that's with love-joy couplings. I have some helical couplings on the way. The 0.1260 balls were too big for the Z axis with 0.1256 balls originally. I can get 2 of the races loaded and its too tight to get to the 3rd and 4th. the only other size between that and 0.1256 from toolsupply is 0.1257, which I guess will have to do. I also tried alternating 0.1256 and 0.1260 and it was still too tight, and its really hard to do with these internal recirculating ballnuts.

[andrew2085]

Its not about strength etc, the helical coupling should have zero backlash, or maybe a tiny tiny amount due to flex. The lovejoy couplings have a lot of slop and even if they are tight with no slop the rubber deflects much more than aluminum. Alignment isn't really an issue, I can get the shafts to line up by connecting the coupling first with the motor mount screws loose. Should be well within the helical coupling's misalignment tolerances. Strength really isn't going to be an issue. 570 oz/in is only 3 ft/lbs, and that's with full stepping at stall. I'm running 1/8th stepping and 1/2 current at idle, so it will be considerably less.