NOTE: I’ve been using this mill heavily for several months now after this initial review, and my opinions have changed some. This mill is still a good mill, but I would NOT RECOMMEND it for production use. Without ballscrews, limit switches, and an R8 or MT3 spindle, production is *extremely* frustrating. It’s a great mill to explore CNC on, but I will be getting a new mill based on production needs.
To see my updated options, look for the red comments
I recently bought KDNTool’s CNC conversion of the XJ-95 mill.
The conversion uses the original screws and brass backlash nuts with a DeskCNC / Gecko based control box.
The manual machine can be purchased from Emachinetool and Penn Tool Co
The most important part… pictures:
The mills performance:
I’ve been using rapids of 60 ipm. This may be a bit too high. I haven’t run repeatability tests, but so far so good. I’m now using rapids of about 25 ipm. Part of this is due to problems with DeskCNC.
I’ve got the physical backlash to about .0015″ on the X axis, and .002″ on the Y axis. This is removable with the software.
The Z has no backlash adjustment, but the head is heavy with no counterbalancing weight or gas spring.
The NEMA 34 motor seems to move it around just fine though.
The travels can vary a bit depending upon if you are willing to go past the edge of the dovetail, if you have trimmed part of the table end plates, and how you have the bellows set up. Without disengaging any part of the dovetail, and my bellows on I’m getting a bit over 9″ X and a bit under 6″ Y. On Z I’m seeing about 8″. Of course, without limit or home switches you have to keep an eye on things.
So far, I’ve been able to hold tolerances of +/- .002″, but I feel I should be able to get that down to +/- .001″ as I learn more about cutting conditions.
What I had to do to the mill:
Adjust backlash. The bearing play in both axes needed adjustment which required removal of the steppers.
This is relatively easy to do when you get the mill. But it needs adjustment way too often if you are using the mill for 8 hours a week.
To adjust the bearings I locked the screw in place with the backlash nuts locked down and then twisted the nylon locking nut on the end of the screw with a socket wrench.
This has caused me some more headaches.
Then to see if I had it adjusted properly, I loosened the backlash nut completely before checking the bearing backlash.
After that, adjusting the nuts is straightforward, if a bit of a pain to get to the Y nut.
(Flip the machine on it’s back).
Install the DeskCNC software and get used to it.
Add way covers.
I don’t want chips in my ways, especially as I am using sticky way oil (Mobile #2) on the ways.
I used flat bellows from McMaster, but I plan to re-do them with upside down U shaped bellows.
I will be using McMaster square bellows and cutting them into two U shapes.
The U shaped bellows were a big improvement. I would also add a large sheet metal tray between the vise and the mill to keep the coolant from draining over the saddle.
I did NOT have to clean the mill. KDNTool already did that.
Things I especially like:
It’s pretty easy to add way covers.
It has an E-stop that works. It kills the spindle and steppers. You do lose position in DeskCNC.
It uses Geckos and a 48 volt power system so it could handle a much larger machine.
It has adjustable backlash nuts. That are hard to adjust and require too frequent adjustment.
The DeskCNC software is pretty easy to use, and does everything in one package.
It has some major issues though…. see below…
The mill itself is better than the X2 mini-mill. (Anyone want to buy my X2 with power feed?)
Thinks I don’t like:
(All of these are very minor issues.) If you are learning CNC, maybe. But if you want to get work DONE, some of these are major issues.
It didn’t come with way covers.
The Y backlash adjustment screws are hard to get at.
Adjusting the bearing pre-loading requires removal of the steppers.
There are no handles for manual movement. I think you should always have the double shaft motors if even just for adjusting the backlash and bearing pre-load.
It’s hard to oil the X axis ways with my oil bottle.
The DeskCNC software has a few quirks I haven’t figured out yet. “Quirks” is a nice word. I hate this software now. It’s “okay” for tool-path generation, but sucks as a controller. You can’t adjust the feed speed without adjusting the rapid speed the same amount. So, if you have the rapids set for just below stalling speed, and push up the override to get a faster speed, the steppers will stall next time you do a G0 (rapid) move. Also, there is no indication of what workspace (G53, G54, G55) you are in, which has caused me further problems. I still haven’t figured out how to use the arrow keys to jog. I will be moving to MACH 2 or 3 on my next mill.
Is uses regular screws instead of ball screws. So far though, I’m not missing them. I think the mid-band resonance damping of the Geckos makes up for the need to drive the screws faster.
Without ballscrews, the backlash changes over time, requiring frequent, annoying adjustment. The ballscrews would be much faster to boot.
It has an MT2 spindle.
If you need to do jobs that require multiple tools, you need at least an MT3 or R8 in order to not lose Z position when changing tools. You also are limited to 1/2″ tools with an MT2. I can’t believe the Wabeco mills use an MT2.
Why the machine itself is better than an X2:
It has both a screw feed for the Z column and a quill for the head.
The gibs are tapered and are longer than on the X2.
The Y travel is much larger (5″ to 6″).
The Y travel would be a HUGE advantage if the machine had the ballscrews, limit switches, R8 spindle, etc.
The table is much larger, making it easier to add a vise like the heavy duty 3″ vise from Little Machine Shop. If you have a mini mill, GET THIS VISE! It is SO much better than the screwless vise I was using.
The motor is more powerful and goes to 3000 RPM.
It’s MUCH easier to break apart and re-assemble for transport. One person can handle it easily. I can dis-assemble or re-assemble it in 10 minutes.
The head can turn for horizontal work.
Where the X2 is better than the XJ-95:
The X2 can be bought with an R8 spindle, where the XJ-95 has a MT2 spindle.
The X2 comes with way covers.
You can get a belt kit for the spindle.
You can get replacement parts from Little Machineshop.
There are several ball screw conversions readily available.
I think these are fairly minor issues.
I was wrong. These are huge issues.
Make sure to check out the comparison page that KDNTool wrote.
Where do I go from here?
I need to add the better way covers. Square bellows from McMaster cut into two U shaped parts is the way to go.
I need to spend some time checking the tram of the column and head. If the column is out of square I’ll need to use shims, but it should be pretty straightforward. Out of the box facing cuts look good.
I plan on adding stream coolant or flood coolant. Probably flood coolant to move chips out of the way. UPDATE: Check out the coolant flood table that I made for this mill.
After the mill breaks in a bit I will be re-adjusting the backlash nuts and bearings.
I would really like to add homing switches. Edge finding the vise everyday is a pain.
I want to slit my wrists when I have an “oops”. I am toast because if I pull the tool to use an edge finder, I will lose my Z height on the tool and need to re-indicate that as well.
Make parts and make money!
Conclusion:
In total, the XJ-95 is a better mill than the X2, and KDNTool has done a great job with this conversion.
I would like to say that dealing with Ken of KDNTool had been great. He has been a big help and answered all my questions.
Check out his website: www.KDNTool.com.
If this machine had ballscrews, limit switches, an R8 spindle, and used MACH, I’d have this mill cutting hundreds of parts a week. If you need to cut lots of parts, look somewhere else. If you want to make a few parts in your spare time, and learn CNC, this is a good mill to start with if the price is right. I feel that a X2 based machine with ballscrews and limit switches would be a much better option even with the reduced Y travel.