Accuracy issues

[itsme]

Hi there,

My mill has been converted for quite a while now, but I am having trouble with accuracy. Today I was machining a small model engine part in an area of about 3 x 3cm. I set the zero point to the centre of a pre-drilled hole and then machined the part. The part included some holes that were meant to line up with holes on the engine's crankcase. After the part had been machined, the wall thickness of one side of the part was thicker than the other side, when they were meant to be the same. When I put the part onto the engine, all 4 of the mounting holes were offset to the right and down a bit. This is a bit irritating, as the 2mm holes are out by more than 0.5mm. I don't think it is inaccuracy in the leadscrew; 0.5mm over 3cm is ridiculous -I know my leadscrews are more accurate than that! I think it may somehow be accumulating an error. The reason why I say this, is because the features on the part that were machined first are close to the correct dimensions/positions, while the features that were machined later on are further out. You can almost tell what order everything was machined in just by looking at how inaccurate it's position is . It is offsetting everything in the same direction.

All the code was hand-written, and I am confident that it has nothing to do with this, as the problem has occured on other parts too. I was using mostly G02 commands in the code. My machine is a Sieg X1 using the original leadscrews and TurboCNC for control.

The leadscrews are meant to have a pitch of 2mm and that is what I have set it up to be on the software. If for whatever reason my leadscrews are not quite 2mm pitch, but the software thinks they are, can this cause an accumulative error?

Does anyone have any idea what might be going on here? I need to sort this problem out, as having these slightly offset parts is worse than not having the part at all. It also gets irritating when you have to remake a part 3 times, just to find the third one is no better than the first...

Thanks
Warren

[lakeside]

Start With A Dail Indicator With One Inch Travel Clamp A Block To Your Table Put .500 Load On Indicator Set A Zero On Controller Move The Axis .300 See If Controll Is Reading .300 Next Move The Axis Back To Zero If The Controller Is Back At Zero Then No Backlass In Manual Move Do The Same Thing With Program Moves See What Happen Then Do Th Same With Other Axis If You Don't Have An Indicator You Can Buy One For Short Money

[cncuser1]

chances are that your motors are skipping steps. I suggest re-running the program as a dry run. No material, no cutter, spindle off.

Carefully observe the servo motor, look for odities in the revolutions in the motor output shaft.

It is likely that the motor is being asled to work harder than it is capable of.

[itsme]

Hi there,

I will see if I can get hold of a dial indicator and try that out.

I don't think it is the motors skipping steps. The motors are 250ozin steppers which are already on the large side for this machine. I was machining with a 4mm endmill at a feedrate of 200mm/min in aluminium - if this is too taxing for it, then there is something badly wrong. I can't get the motors to skip steps by trying to stop the table either - it just ploughs on...

Regards
Warren

[Zumba]

Originally Posted by itsme I will see if I can get hold of a dial indicator and try that out.

I take it from the fact that you don't own a dial indicator, that you never setup your machine correctly. Don't worry, it's easy, but you'll need some measuring tools. I'd highly recommend getting a test indicator (Brown & Sharpe BesTest or clone) and also another dial indicator on a magnetic stand.

Use the test indicator to tram your head. Place the test indicator in the spindle using a good collet, angle it at about 45 degrees, and check to make sure that it gives you the same reading in each direction - North, East, South, West. I can almost guarantee you that i'll be off by quite a bit, at least in the East/West direction.... several thousandths.

As for the dial indicator on the magnetic base, use that to do what lakeside suggested. You can also use it to measure spindle runout and quill accuracy.

[lakeside]

Brown and sharp stuff is very good and so is intrerapaded but cost take a look at spi and also craftsman you are not tiring to measure .0001 so why spend the money also e-bay is a go place to look

[phil burman]

Hi Warren,

Some thoughts:

1) Yes of course if the leadscrew is not quite 2 mm pitch then this would lead to a scaling error if you input 2mm pitch.

2) Did you input the screw pitch into the control software in metric or imperial. Your inaccuracy of 0.5mm in 3cm is the same as the inaccuracy in 1mm equals 0.040". Of course 1mm actually equals 0.0393701"


3) You could use the machine to scribe a series of lines (using a pencil on paper) and check the error every inch over say 10" with a ruler. This will give you an idea of scaling errors.

4) If you can't find the error in the hardware or your inputs to the software then you should talk with people on the Turbocnc forum (I assume there is one).

Regards
Phil

Originally Posted by itsme Hi there, My mill has been converted for quite a while now, but I am having trouble with accuracy. Today I was machining a small model engine part in an area of about 3 x 3cm. I set the zero point to the centre of a pre-drilled hole and then machined the part. The part included some holes that were meant to line up with holes on the engine's crankcase. After the part had been machined, the wall thickness of one side of the part was thicker than the other side, when they were meant to be the same. When I put the part onto the engine, all 4 of the mounting holes were offset to the right and down a bit. This is a bit irritating, as the 2mm holes are out by more than 0.5mm. I don't think it is inaccuracy in the leadscrew; 0.5mm over 3cm is ridiculous -I know my leadscrews are more accurate than that! I think it may somehow be accumulating an error. The reason why I say this, is because the features on the part that were machined first are close to the correct dimensions/positions, while the features that were machined later on are further out. You can almost tell what order everything was machined in just by looking at how inaccurate it's position is . It is offsetting everything in the same direction. All the code was hand-written, and I am confident that it has nothing to do with this, as the problem has occured on other parts too. I was using mostly G02 commands in the code. My machine is a Sieg X1 using the original leadscrews and TurboCNC for control. The leadscrews are meant to have a pitch of 2mm and that is what I have set it up to be on the software. If for whatever reason my leadscrews are not quite 2mm pitch, but the software thinks they are, can this cause an accumulative error? Does anyone have any idea what might be going on here? I need to sort this problem out, as having these slightly offset parts is worse than not having the part at all. It also gets irritating when you have to remake a part 3 times, just to find the third one is no better than the first... Thanks Warren

[itsme]

Hi there,

I think it probably is time to get my machine setup and running properly. It gets extremely frustrating when parts don't turn out the way they should, because of a poor setup.

I do have a dial indicator, but it only reads to 10mm, which I don't think is enough to setup the machine correctly. I had a look at dial test indicators too. There are a few details that I don't quite understand (since I've never used one of these before). They seem to have small dovetails that are used to mount them. How do you go about mounting the indicator into the machine spindle? Would I need to get hold of a matching dovetail holder that would go into a collet on my machine? Also, is the small 'needle' head the part that can be angled to, say 45 degrees?

Thanks
Warren

[lakeside]

try this site for some attachment as far as set up keep the needle of the indicator at a slight angle to the surface that you are indicating?
http://www1.mscdirect.com/CGI/N2SRHI?PMSECT=1321

[MIKE JEFFERS]

Warren
get yerself a plunger type dti (arc chronos) they have more travel
maybe your lead screws are wearing
mike

[itsme]

Hi Mike,

I certainly hope my leadscrews aren't already worn out, as I only replaced the x-axis leadscrew about 1 week ago, when I added an extra length table to my machine. Just in case you are wondering, I did have similar accuracy issues before I replaced the leadscrew.

I'm working on the DTI.

Regards
Warren

[phil burman]

Hi Warren,

You have people spending their time trying to help you. It would have been useful if you had tried to help them help you by putting forward this extremely pertinent information at the start of your thread. Unless both leadscrews have the same systematic fault then you should be looking at the software.

You work it out.

Regards
Phil

Originally Posted by itsme Hi Mike, I certainly hope my leadscrews aren't already worn out, as I only replaced the x-axis leadscrew about 1 week ago, when I added an extra length table to my machine. Just in case you are wondering, I did have similar accuracy issues before I replaced the leadscrew. I'm working on the DTI. Regards Warren