Ditron vs Sino...wich is better in your opinion

[rwm]

I have 5 micron scales, they are working fine.

Hey Pontiac- Do you have this on a lathe? Have you tried it in diameter mode and compared with DTI?

I'm thinking the 1um scales are probably just more accurately made. If I get 5 times better accuracy that would be an acceptable error. .002 becomes .0004.

 

[pontiac428]

No, I have mine on a mill, and the default raw display is decimal plus five digits.
It's not more accurate, but the 1mm has a finer resolution, so it can potentially be more precise.

The issue is on diametric feed, where Dia=infeed x 2, is your .0002 step resolution becomes .0004 step resolution. But I don't see it that way, read heads are reading two lines at once, coarse+abs and fine. The head can interpolate steps between steps using some math tricks, so I do not think you would actually see .0004 steps on your screen. My mill scales are all 1:1, my display is set for four zeros (.0001), and I get valid microstep DRO data without steps at .0002 intervals due to the interpolated result.

Wikipedia has a good article on linear encoders that explains how it works, but it's typical of multiplexing schemes and not that cutting edge. I don't buy any talk of rounding error here, it does not fit this particular implementation of technology. For all we know, the head could be reporting a 16-bit decimal and incrementing it, it's a black box until we see the source code. I don't think .0002 would be an acceptable limitation otherwise.

 

[rwm]

Can you translate that into what I should do?

 

[pontiac428]

Can you translate that into what I should do?

I think everyone here that has chosen 1um scales have been happy. I think several posted early in this thread.

I stayed with 5um scales so I wouldn't hit the speed limit on the mill, where it is possible to with jog. On the lathe, I don't think there are any movements you can make that would be fast enough trip up an ultrafine scale. So I have no detractors to that recommendation.

 

[rwm]

Are you saying that fast motion could make the read head miscount? I did not know that? Does it just count pulses or is there some additional roadmark?

 

[pontiac428]

Yes, and it is the Achilles heel of 1 micron scales! The rate that the pules come is faster than the head can read and refresh. More isn't better, usually it comes at a price.

I think you need it on your transverse, but maybe stick with a 5 micron scale on the long travel- you can envision why.

 

[pontiac428]

The A channel is the Sin channel, and B is the Cos channel. When the signals are read together, they take the form of a Lissajous figure. Because the sign of each forms 4 combinations (in this example),, there is 4x more data being read than is actually printed on the scale.

Because the signals are waves, midpoints between readings can be obtained through interpolation. My head does this to six decimal inches, even though the scale is no finer than .0002. Not only does it work, the error is miniscule. Xerox was doing this stuff by the end of WWII.

 

[rwm]

Thanks for the detail! Dumb question. What is the physical scale of those square waves? Is one wave one micron peak to peak?

 

[mksj]

Your head maybe one thing, the conversion to data by the microprocessor in the DRO is another thing. If there was an issue with the data stream speed they wouldn't be selling 1 micron scales. On 3 lathes I had the same issues as to resolution errors on the cross slide when using a 5 micron scale, the issues became a magnitude less when using a 1 micron scale. It is general practice if you are using diameter mode on a lathe to specify a 1 micron scale on the cross slide. The long axis, a 5 micron scale works just fine. The quality of the individual components is what matters, I have seen people that have issues with the accuracy of the Ditron DRO's whether this is the DRO head unit or the scales I do not know. I have been using Electronica and Easson head units and have no issues with either other than the noted problems of using their 5 micron scale on the cross slide. Once they were replace with a 1 micron scale, I could easily get repeat turning accuracy of better than 0.001". If you do not need that accuracy than use a 5 micron scale, as I mentioned different scales have different accuracies even though they all have the same resolution.

Target diameter was 0.1550" +/-0.0005"

 

[rwm]

I think everyone here that has chosen 1um scales have been happy. I think several posted early in this thread.

I stayed with 5um scales so I wouldn't hit the speed limit on the mill, where it is possible to with jog. On the lathe, I don't think there are any movements you can make that would be fast enough trip up an ultrafine scale. So I have no detractors to that recommendation.

Do glass scales work the same way? How about capacitance? Is there some scale that uses an absolute reference rather than counting?