View Full Version : IH accuracy poll

02-09-2007, 11:23 PM
I bought my IH mill last spring and hadn't really used it for anything that needed to be accurate until a couple of days ago. I was dismayed to find that I had a lot more runout in the x-axis than I had bargained for. I have about .005 in 8 inches of travel. So I fired off an email to Bob Warfield ,whom I am sure most of you are familiar with, and asked what the stats were on his machine. Well Bob informed me that his machine only has about .001 from end to end. I sent an email to Industrial Hobbies to ask what the specs are supposed to be for this machine but haven't recieved an answer yet. So I wanted to take a poll of the IH owners here and see what the specs on your individual machines are. I would also like to know how many of you lapped your machines. :confused:


02-10-2007, 12:47 AM
Welcome to the Zone!

I lapped my way. Personally, I can't recommend it to anyone. Kinda seemed to be a complete waste of time.

Concerning errors... a thou or two is right in line over most of the travel of the table (not at the extremes of the X though). I have .0025 out in both the X and Y over 6" of the head travel but that could be improved.

The quill was pretty bad on my machine so that is never unlocked anymore.

02-10-2007, 10:26 AM
Accuracy is only going to be as good as your setting up, and what little backlash left is going to be cancelled with software comp. Be sure to use a low comp speed setting to not have a snap in the comp. mine now will interpolate a bore and hold it round and true to "less than" .oo1"

02-10-2007, 10:28 AM
As far as what happens if I run an indicator from the head to the table, table reading as it were, I get less than 0.001" deviation from 1 end of the table to the other in the X. The Y is a little different. Starting at center & going both ways gives me 0.001" in each direction which tells me the table is a little bowed front to back. The column error is less than 0.001" over 6" head travel in the X & 0.002" in the Y. Also adjusted the head in the X & shimmed it in the Y which got me down to 0.0005" error there. The table isn`t real flat. I can take a precision square & lay it on the table in the Y axis & see light coming under it in certain areas. I would like to have the table taken down about 0.005" or so to even things out but I haven`t looked into what the cost would be to do that yet. Not sure it would be cheap.

02-10-2007, 10:49 AM
Hello all, I did nothing to the mill and have about the same as Mr. Warfield. I did, however, make sure backlash in Mach3 is on :) I first had set it to 70% of full speed, when they 'switch' to compensate it over-shoots about .005. Yesterday morning I turned it down to 40% (which slows down your over-all runtime) and what a difference. I know, I know, steppers ;p

02-10-2007, 11:05 AM
I slowed mine down to 7% speed in backlash comp, before with 12% i could hear it, now it is smooth, quiet,and no line remaining in surface ! read "no measureable inaccuracy to speak of"

02-10-2007, 04:17 PM
I measured my X and Y axis table runout and cannot measure hardly any deviation over 12" with a .001 indicator.

But I just got a .0001 indicator so I measured the same with it. I am showing up to .0013 variation in the table surface over 6", but when I put a B&S parallel on the table, I get .0003 runout in both X and Y directions, over 6".

Looks like they had a good day at the factory when they made mine, at least on this part of it.

02-12-2007, 01:15 AM
I can see from some of the responses that I didn't articulate what I meant very well. The run out that I am refering to is a measure from the head to the table as the table is traversed in the x axis. My actual backlash on the lead screw isn't that bad, only .003 after a little adjustment. Easily compensated for. I discover the out of square condition while squaring blocks. After machining a four inch block a check on the surface plate revealed that the block sides were out of parallel by .006. So I started checking. I thought that I would check the table first just so I could eliminate that. Well it seems that the majority of the problem lies in the table and ways. I haven't tried to put the table on the surface plate, just a little hard to handle. But I put the vice on the surface plate and it only has a .0001 deviation from end to end and that could be coming from the surface plate. I think that my next step will be to remove the table and check for contamination and bruises. I am actually hoping that I find some. A couple of bruises would be easy to correct with a scraper, but if I have a bad out of square problem then that is a case for remachining the table. I have access to the equipment to do it, but haven't tried anything that large that had to be more accurate that +-.005 which is about where I am at now. I think I could do much better than that but I haven't ever tried to get that accurate before so am a little apprehensive. I'm also not sure if it is the table top or the ways that have the problem, so until I get it on the surface plate for inspection I can only speculate. Keep the info coming I appreciate it.