Thread: DRO Accuracy?

I've finally got time to play around with the mill and noticed for every turn of the handwheel (resulting in 2.5mm table movement) the DRO reader shows slightly less. It adds up to quite a bit as the table moves and I'm wondering if I installed these correctly.
I have it *very* parallel to the table movement so I don't think it's cosine error. Both x and y show the same thing.

One thing in installation I'm unsure of is the reader setup. This is my DRO, should I leave the red thing attached and screwed on? Should the gap between reader and scale be ever so slightly more than its thickness?
Online guides for Easson and similar glass scales recommend 0.8-1.5mm gap which is roughly what I have it set at now. I have also found ones for 3mm gap though.

post image link

I adjusted the linear compensation but that only fixed up part of the travel error, it seems the error varies in different parts of the table.

Any advice appreciated.

Wear in the screw? Cross check the DRO with a long travel dial indicator.

The mill is pretty new it barely has any hours on it. HM-46 rong fu 45 clone from hare and forbes.

The red piece is just to keep the reader stable during transport etc. It should be removed once the scale is set up. I agree with checking the table movement with an indicator. You could try with a caliper to get reading over a longer distance.

Dean

Does turning the hand wheel one turn actually move the table 2.5mm? I would be inclined to believe the scales rather than the dial markings, funny things are known to happen with lead screw pitches and dial markings at the lower end of the machine tool market, I can tell you about my Hafpos AL340 lathe for instance.
Can you do 10 or 20 turns and measure actual movement with a vernier?

The red thing is used to set the appropriate gap when setting up the read head. Screw the head down and then remove the red plastic, the gap will then be correct. The resolution of the DRO is quite small, so you need to maintain very high rigidity on the scale and read head. You should be able to apply pressure to both the read head and the scale and not see a massive variation in the DRO reading (there will be some variation however as they are very sensitive).

Ignore what the hand wheels say and actually measure a known standard. Modern DROs are generally very accurate. If you don't have long gauge blocks you could use something like a micrometer standard or similar. Set your linear compensation back to zero and measure the actual travel from your standard at different parts of the table.

Well shoot thanks for the advice, I've done one set of readings on a 1-2-3 block (long side) and got:

76.20 theoretical length (3" = 25.4x3 = 76.2)
76.19mm measured with mic (I am new to using micrometers though)
76.21mm measured with DRO
76.30mm measured with handwheel of mill

And with my cheap thrown about chinese digital vernier with a cracked screen I get 76.2ishmm.

Looks like the DRO is pretty accurate over that particular portion of the travel. I might buy a slightly more expensive vernier this week and calibrate it using that.

if you're particularly concerned about accuracy I have a set of full set of cej gauge blocks here. you'll more than likely find the DRO is accurate but the leadscrew pitch is both slightly off and not consistent over the entire length.

I personally use an indicator to measure travel when ever possible, usually on the lathe when I have to hit a critical diameter on a small run of parts. on the mill the travel needed makes it a little inconvenient unless you have an indicator with 5"+ of travel but you'd rather not know the price.

21392809_10155708173389993_1575527100_o.jpg

Originally Posted by j3dprints

Well shoot thanks for the advice, I've done one set of readings on a 1-2-3 block (long side) and got:

76.20 theoretical length (3" = 25.4x3 = 76.2)
76.19mm measured with mic (I am new to using micrometers though)
76.21mm measured with DRO
76.30mm measured with handwheel of mill

And with my cheap thrown about chinese digital vernier with a cracked screen I get 76.2ishmm.

Looks like the DRO is pretty accurate over that particular portion of the travel. I might buy a slightly more expensive vernier this week and calibrate it using that.

The DRO will be generally more accurate than the lead screw. I've found the Easson scales to be pretty good generally speaking.

Did you know that the DRO has a calibration factor? Which you could check to see what it is, but, If you don't have the means to check it, I'd leave it at the factory setting.

Here's the calibration check we did on the HM52 with Easson scales, ended up with +86 ppm correction factor.

Yep it seems pretty accurate but the numbers I am getting suggested it is out 0.56mm per 1m. I input that into the linear correction factor and it's now reading closer.

Interestingly/worryingly if you approach the same point from different directions - that is, the same reading on a dial gauge - the DRO reads different amounts. I tightened the axis locks and this minimised this behaviour - but they had to be pretty tight. Also, tightening the locks alone moves the DRO value. I guess this has something to do with the gibs, and them not being perfect ?

Originally Posted by j3dprints

Yep it seems pretty accurate but the numbers I am getting suggested it is out 0.56mm per 1m. I input that into the linear correction factor and it's now reading closer.

Interestingly/worryingly if you approach the same point from different directions - that is, the same reading on a dial gauge - the DRO reads different amounts. I tightened the axis locks and this minimised this behaviour - but they had to be pretty tight. Also, tightening the locks alone moves the DRO value. I guess this has something to do with the gibs, and them not being perfect ?

you're reaching the point where no amount of fiddling with your mill in its current state will truly get the accuracy you're looking for.
Its time to break out the surface plate and embrace the metrology life.

In my experience if the scales have been installed parallel to the axes and a horizontal, the DRO will be correct and you won't need compensation. The way they actually work means that they will normally either work to their specifications, will be unreliable, or won't work at all. It would be unlikely they would be inaccurate yet consistently so, especially in two axes.

0.5 mm/m isn't much in the grand scheme of things, and with all due respect from what you've said I'd suggest you don't have the capacity with the equipment you've got to conclude that accuracy from the DRO.

I would check the methodology of how you're measuring things and work from a known standard. When I tracked down an issue with my DRO I used a 100 mm gauge block, which is the longest I have. The issue turned out to be a mount that was flexing slightly (I wasn't getting repeatable results when approaching from either side). At the moment I feel you're making all of your instruments read the same, which will make everything very precise, but sadly also very inaccurate.

Originally Posted by bob ward

Does turning the hand wheel one turn actually move the table 2.5mm? I would be inclined to believe the scales rather than the dial markings, funny things are known to happen with lead screw pitches and dial markings at the lower end of the machine tool market, I can tell you about my Hafpos AL340 lathe for instance.
Can you do 10 or 20 turns and measure actual movement with a vernier?

I agree the DRO is probably correct . Correct me if I'm wrong but a hand wheels segments are only going on thread pitch with any associated backlash involved whereas DRO scale is measuring the movement independent of thread pitch and backlash .

Yep, you are correct. The x-axis also measured correctly with a 1-2-3 block. The block was 76.2mm and the DRO was pretty much dead on.
I did adjust for backlash of course previously when testing things.

Originally Posted by Retromilling

I agree the DRO is probably correct . Correct me if I'm wrong but a hand wheels segments are only going on thread pitch with any associated backlash involved whereas DRO scale is measuring the movement independent of thread pitch and backlash .

Unfortunately not entirely correct. "Backlash" won't affect the handwheel accuracy as you're only measuring with the axis going in the same direction. Backlash only becomes a factor when you reverse direction, but even then won't affect accuracy. So you could have a machine with an extremely worn or loose laedscrew nut, but it could still be quite accurate. On the other hand an inaccurate leadscrew, either through wear or poor manufacture, will affect accuracy to the extent of the wear.