Hi
@Sticks
I have a PM940M-VS-CNC . Took delivery in late 2017. I cannot be positive that it is exactly the same as the currently sold machines as it was purchased as a CNC and then later PM stopped selling it. But I tend to think that the fundamental castings etc are the same. There were several of these CNC machines sold back then. I too wanted a larger set of motions and a heavy machine, but this is not always the best way to go. I do not regret purchasing it, but there have been lots of problems. The machine is ok, but the China QC leaves a lot to be desired! Others have purchased 940Ms with a lot fewer problems than mine. Over the years I have attacked one problem after another and have made the machine better. But there are some problems which are pretty hard to fix. I still like the large motion, but I would look long and hard before getting another China made machine.
Doing it over, I would get the belt driven version without lots of bells and whistles and use it for a while before making changes. Getting the PM DRO is ok, but for considerably more money you can get a fancier one. The Delta VFD has also been fine.
The geared head is too heavy for the column and tends to fall if the gib is not over tightened. (fall: When the stepper power is turned off the stepper motor does not have enough magnetic torque, cogging, to hold the head position. So, the lead screw slowly unwinds as the head falls.) When the gibs are over tightened the backlash increases. The only solution I have figure out for this is to counter balance the head weight either with a cable and weights or via air springs. Attaching air springs is difficult and may limit the z-axis motion. The best I can tell, the head weight about 250-275# without excessive tooling. It tends to have about 0.0015" dynamic nod (tilts) when changing direction. There is considerable (several thousandths) backlash and this is not just because of the looseness of leadscrew bearings/nut. The ways are not totally parallel and if you set the gib to be correct at a region where the ways are narrow then when you hit the other regions the backlash increases. I found that if the gibs are set loose, so that friction at the ways is not causing excessive backlash, then the y-axis backlash can be made to be under 0.003". (When I have time I work on the y axis ways to make them more parallel, and hence have improved its backlash considerably. The 0.003" is probably due to the motor coupling and the shaft thrust bearings.) The Z axis is similar including the dynamic nod mentioned and the x-axis seems to be more like 0.005+. These certainly were not what I expected from the spec sheet. I am still working on improving these. I am sure you can get better performance than I have via an after market CNC conversion kit that uses backlash compensated lead screw bearings. But don't think you can get by with cheap lead screws. Some of these even have bad threads. Other HM folks have used and seem to like the kits from
https://www.arizonacnckits.com/pm-940-cnc-kit.html . PM even seems to reference them. If you look at these you can get an idea of what it will cost to do this mechanical conversion. I would get the heavy duty version with 32mm lead screws. You will still need electronics, motors, switches, computer, cables, etc. Get the good ClearPath motors with the in-line battery backup so that you do not loose the stepper position when the power is turned off. These motors are more expensive, but have the drivers built in. Also, they have built in encoders so that they actually measure the motion not just send out an open loop step pulse and ignore whether or not the motor actually turned the lead screw.
At the time I paid about $9K+ but this included the 4th CNC axis and rotary table/chuck. With a few accessories, Vise, collet set, spindle chucks, clamp set, etc... it came to ~ $10K. It has the hardened ways, cast iron base, geared head with a maximum speed of 3200 rpm. (I am not for sure why an earlier comment was made that the max speed is 2000 RPM, except that their are two gear settings and the label says 90-980 for setting 1 and 1000-2000 for setting 2. But these are not correct they need to be scaled up to a max of 3200 as this is what the VFD causes.) It has a 2HP (1.5KW), 3 phase motor and the VFD (Delta) converts 220V single phase. It came with an automatic lubricating system which is timer programmable for the ways and lead screws ( currently I have it set to pump way oil for 30 seconds once an hour, but I find I have a lot of idle time so the machine is over lubricated (dripping) at this setting. It also came with a very healthy coolant pump and recycle reservoir (coolant is largely caught by the stand channels and guided back down to the reservoir to be pumped again. That is most of the features. Because it was CNC it did not come with a DRO. I wish it had and have thought about add it. In theory you have the positions via the the CNC software, but with the stepper motors the positioning is open loop, no feed back. Steps can be missed. An external DRO would provide better info about where the tables are independent of the PC saying where it is suppose to be. It runs on Mach3 CNC software which is cheap, but works fine. I should probably purchase the Mach4 version as I think it is better documented. I will attach the quote spec sheet that PM sent me prior to my purchase so that you can see what was described at the time. Part of the reason I purchased it was it was already on a boat and about half way here from China, it was CNC ready and, mostly, I knew very little about mills at that point in time. When I purchased this I thought I had to have a PC with a parallel port to interface to the controller. So I actually went out and bought a used PC which had an old fashioned parallel port. This was not the case as the controller interface was USB 2.0. Anyway, I should replace the PC as it is pretty old and sometimes crashes! (This is not a PM940's problem.)
I will attach a spec sheet that I got around the time of purchase. It is not exactly what I received, but is very similar.
What ever you purchase, you should spend time carefully checking out the backlash. This is pretty easy with the CNC, but can also be done with a little more trouble and skill on a manual machine. Make sure the ways are parallel. If the backlash is not the same at all locations along the length of the lead screw travel then the ways are not parallel. Non-parallelism may be the hardest thing to fix on a machine as it is fundamental. Do this while the machine is new so that you can get the issue resolved. I suggest the following method:
1) Tighten the gib so that you can actually feel a springing at the manual crank. That is you can actually feel the handle spring back after you try to turn it. If you are good at this you do not need to measure anything else at this point. (The springing effect is due to the lead screw actually twisting. It is long so if there is resistance it tends to twist and then spring back. A smaller diameter lead screw will take less torque to twist. Obviously, the twist is also a function of the length of lead screw between the ball screw and the handle where the torque is being applied.)
Now crank through the length of the lead screw. You should be able to feel, and see via the crank dial where the springing increases or decreases. Find the location along the lead screw where the springing back is minimum. There maybe be a place where there is none. If so then use this spot to tighten the gib a bit more. Repeat until you find the location where the springing is the smallest. This will be where the ways and gib are the loosest fit, or said differently where the way spacing is the narrowest. If there is not springing then tighten the gibs a bit more until there is some. If the torque required to turn the handle, or the springing is excessive do not over do it. It maybe possible to actually damage the machine.
2) At the point where the ways are the narrowest. Slightly loosen the gib until the springing is minimal, but not so loose as there is none. Now measure the backlash in a conventional manner using a good dial gauge. Measure it when moving in both directions and record the results. Make a note of where this measurement was made along the lead screw. ( I suggest always measuring from one end of the lead screw motion. Or for example when doing the y-axis the table is against the column.) My method of measurement is to use a good digital dial gauge. I fasten the gauge to the spindle and crank the table until the dial gauge deflects several (> than any expected backlash value) thousandths. Note the value and note the distance measured via the crank dial. Reverse crank a distance greater than the backlash value. Subtract the two differences in the two sets of readings to get the backlash. When you reverse crank you will see when the dial gauge finally starts to move. You might be inclined to say that this is the backlash, but it is actually slightly worse than this as this deflection is commonly not linear. I searched for a while to find a good digital dial gauge and finally settled in on one made by Clockwise (Amazon). I checked its calibrated it and found that it is good to about 1 or 2 microns over short distances and to about 5 microns over the full 1 inch motion. (1 micron = 0.00004") This is very good for a full 1 inch of motion. (Clockwise Tools DITR-0105 Electronic Digital Dial Indicator ~$65) There are a lot of cheaper look a likes out there, but this is the one that is both robust and accurate. I did extensive calibration measurements on this one and then purchased two more for other measurements. It also uses the coin style Lithium battery and retains the data, at least for a while after the display goes blank. You can also get a special RS232 digital USB cable for it which will allow you, at the click of a push button switch, to record the read out into Excel or Word. (Clockwise Cable DTCR-01)
3) Without changing the gib, crank to a position where the springing is the worse and repeat the backlash measurement. If cranking to this lead screw position requires excessive torque then stop. It maybe possible to damage/crack the saddle if it is really bad. Measure the backlash here. Loosen the gib, while counting the turns on the gib screw, until the springing is similar to what it was at the point used in step 2. Measure the backlash and make sure it is similar to that of step 2. Keep track of the number of turns of the gib screw. By knowing the TPI of the gib screw you know how far you had to move the gib in or out. Take the gib out and measure its taper, the thickness per inch of length. (You can removed the y-axis gib without fear. If you want to remove the Z-axis gib you will need to clamp the head in place.) Hence, you have a measure of the difference in the way spread because you know how far you have to move the gib in and out to make achieve that same way spread. A perfect machine will have perfectly parallel ways.
4) If you really want to know more then measure the backlash as a function of the full lead screw positions.
On my 940M I have done this only on the y-axis so far. I found the ways were very non-parallel. At the narrow point I found the loose gib backlash to be about 0.002 to 0.003" which occurred near the column. As I moved the table out, without changing the gib position, I found the backlash increased to > 0.020 and was increasing very rapidly with lead screw position, .... I had not gone but about 2/3 of the way out on the y-axis. I stopped out of fear of damage. If I loosened the gib up a lot I found that the backlash could be the same as at the initial narrow way position. It was very bad when adjusted for the initial position, so I have been removing some of way material at the wider spacing of the ways. I am slowly getting there and it is now much improved, but it takes effort to remove material from a hardened steel surface. Not impossible but time consuming. I started out scrapping, but moved over to old fashioned sanding! I use the clockwise dial indicators to measure how much I have removed. Lots of measuring to avoid not taking too much! The problem with just running the gibs loose is that the table motion may not be square or may even twist during machining. For CNC operation if you have backlash then circles become ovals or some other shape. By the way, I continue to use the mill and have used it to built fixtures to do these very measurements! I think that scrapping procedures call for one to take the machine apart and make the measurements as one scrapes. I do not have the tools for this and I doubt that the results would be as good a measure as this backlash measurement procedure.
There is also, potentially, a fundamental design issue with the 940M and probably others, which may or may not be a problem. This maybe true on all of the machines I just do not have sufficient machine experience to know. That is that of squareness between the x and y axes. For the 940 the angles between the x and y directions is fixed when the dovetails are cut in the saddle that connects them and so it not adjustable. So if they are not cut correctly at the time of manufacture then the x and y angle will never be exactly 90 degrees. To check this you need a really good square. Perhaps this can also be done by machining a square with the mill. When you then measure two sides of square with a dial gauge it should look perfect, but this is false. However, if you flip it over and around and re-measure you may find differences and so will know that the machines x-y are not 90 degrees etc. I have not bothered trying this yet, so have not tested the process. Sometimes you do not want to know the answer! Enough said.
Good luck,
Dave L.
