So in trying to eliminate as much equipment and instrument error as possible for structural staking, I began doing some research on the accuracy of mini prisms. We have always used as short a rod as possible and mini prims for high precision layout but I had never paid much attention to the different levels of accuracy of the mini prisms.
After doing a little research I noticed that the prism accuracy is typically titled "centering accuracy" when looking at prism specs. The mini prism we typically use is rated at 2mm while some of the leica mini prisms I have found are rated at 1mm and even 0.3mm (large price difference obviously).
Can anyone tell me what they mean by "centering" accuracy? I had assumed that prism accuracy would be measured by its ability to consistently provide accurate distance measurements in conjunction with the edm.
I apologize for my ignorance, this is something I feel like I should know.
Some claim they can achieve centering accuracies of 1 or 2mm with a standard prism pole with an 8 minute bubble i.e. that the prism can be centered vertically over the point and would not be out of line in any direction by more than 2mm at the top of the pole.
I imagine the specs for the mini prisms are calculated by the accuracy of the bubble on the mini prism with 8 minute being the highest accuracy and 40 minute the lowest.
I see what your saying but I envision that the accuracy specs would have more to do with the prism itself rather than the pole or the bubble.
I say that because I find mini prisms with 8' bubbles, some with a centering accuracy of 2mm, others more like 1, 0.5 or 0.3mm. I really don't know if this is the case, just seems to be right from my research.
That said, I tend to have ideas as to what the accuracy is not calculated by, but really no idea as to what it is calculated by.
Any other ideas/opinions?
Thanks.
I've never done any survey work where I'd care about the difference between 0.3mm and 2.0mm. What kind of stuff are you doing where that would make a difference?
Actually, I spoke too soon before. I have had tight tolerances where I tried to achieve 0.005' which would be roughly 1.5mm, so I guess maybe it would make a difference. I think the actual tolerance for the steel was 0.02' so I was trying to stay way under that for the layout.
Are these nodal prisms you're looking at?
I would expect that the term "centering accuracy" with regard to a prism would refer to the accuracy with which the centerline of the glass is positioned over the center of the mount side-to-side. "Prism constant accuracy" would address the fore-and-aft positioning. But I've never seen either of these terms used.
I agree with Jim, that is what I would think it means.
The accuracy of my 2 mini prism bubbles is hardly useable over 1ft HI and neither is adjustable, they are glued in place.
The 25mm prism themselves are as accurate as most any of the 64mm generic, 2 1/2in Seco & 73mm Retro prisms in my box when used on rods or tripod-tribrach setups.
Some are sealed and others are not. If the glass is loose in the assembly, take apart, clean and like everything else, put together snug tight.
I tested everything by setting up a tripod, tribrach and puck with a removable brass center to swap all my prism & targets on for comparison.
There is little or no difference.
Dan Patterson, post: 353524, member: 1179 wrote: Actually, I spoke too soon before. I have had tight tolerances where I tried to achieve 0.005' which would be roughly 1.5mm, so I guess maybe it would make a difference. I think the actual tolerance for the steel was 0.02' so I was trying to stay way under that for the layout.
Are these nodal prisms you're looking at?
Dan, here is a link to one of the prisms I was referring to made by Leica, they claim a 1mm centering accuracy with this prism. The second link is another Leica prism that claims 0.3mm centering accuracy. The price difference is substantial. I will most likely drop the $225 for the 1mm prism, just can't see justification in the benefit for the cost difference for the 0.3mm prism. The 0.7mm difference equates to 0.002'.
http://www.allenprecision.com/survey/robotic-total-stations/accessories/prisms/mini-prism-set/
2mm is .0065'. For all of our structural staking projects, I like to try to keep our tolerances in the thousandths, anywhere under .010' is what we would consider acceptable (of course we know there is additional error there although it should be very minor). That said, we do layout for industrial manufacturing facilities where equipment center lines are expected to be "perfect" according to the manufacturer. We have done this enough to figure out that these types of people do not understand what they are saying when they say "perfect''. However, I do want to eliminate as much error as physically possible to eliminate any type of noticeable discrepancies with anyone checking my work.
Last summer we were laying out 8 points along a center line 840' long. We were told by the manufacturer that there was a "0" tolerance for these points and that they all had to be "perfectly" online. I then watched them set up in the middle with a 10" theodolite (who knows how long it had been since their instrument had been cleaned and serviced) site to one end and and turn a single 180 to check points in the other direction. I looked through his instrument and saw that he was missing one of my points by roughly 1/8"-3/16", he said "that's perfect". I tried not to laugh.
agrimensor06, post: 353531, member: 11020 wrote: one of the prisms I was referring to made by Leica, they claim a 1mm centering accuracy with this prism
The 1 mm refers to the accuracy of the Leica ATR to center on the prism. It's an instrument specification -- actually, a system specification -- rather than a prism specification.
I have 3 of the GMP101 prisms, and I use them for probably 95% of my total station work, almost all of which is robotic. They're kind of spendy, but well worth it in my view.
Jim, thanks for your help, that makes sense.
What type of short rods or pin poles do you typically use with your mini prism when performing high precision layout?
Also, any idea what makes the leica prism featured in the link below so much more expensive than others? Is it something other than the step up from 1mm centering accuracy to 0.3mm?
Thanks to everyone for all the replies and help!
agrimensor06, post: 353536, member: 11020 wrote: What type of short rods or pin poles do you typically use with your mini prism when performing high precision layout?
I use one of these for close-in on-the-ground work:
I don't do a lot of close-in precision work, but a few years back I had a project that required setting 4 tie-out points roughly 30 feet from each of 30+ corner monuments to be perpetuated, and restoring the corner positions from those points post-construction. I used the mini-tripod with a GMP101, took ample redundant measurements and did a network adjustment at each site. Residuals were consistently less than 0.015 foot, with most less than 0.01 foot.
The GMP101 built-in bubble is good, but its placement is not conducive to work down low -- it's too hard to see the bubble from directly above. That's where the bubble on the mini-pole is useful, and it can easily be checked against the GMP101 bubble.
agrimensor06, post: 353497, member: 11020 wrote: So in trying to eliminate as much equipment and instrument error as possible for structural staking, I began doing some research on the accuracy of mini prisms. We have always used as short a rod as possible and mini prims for high precision layout but I had never paid much attention to the different levels of accuracy of the mini prisms.
After doing a little research I noticed that the prism accuracy is typically titled "centering accuracy" when looking at prism specs. The mini prism we typically use is rated at 2mm while some of the leica mini prisms I have found are rated at 1mm and even 0.3mm (large price difference obviously).
Can anyone tell me what they mean by "centering" accuracy? I had assumed that prism accuracy would be measured by its ability to consistently provide accurate distance measurements in conjunction with the edm.
I apologize for my ignorance, this is something I feel like I should know.
hello agrimensor,
i think centring accuracy would likely refer to a tolerance for how well the EDM and ATR return from the prism glass would coincide with ground mark or spigot on which the prism was set, for a sample of manufactured prisms assemblies. for a leica gpr112, to name one, the prism glass has to be cut; the 90* retro-reflecting back end and it has to cut circularly on the outside to fit into the prism holder. the glass has to be set and glued in 3 dimensions in the prism holder. 2 threads are set/pressed? into the holder to screw in the pivot screws. the holder assembly sits in the bracket which has a sleeve cast/pressed into the bottom to fit on the spigot. a spring loaded button helps secure the prism on the spigot but also rocks the prism sideways when the button is released. all of these factors introduce a centring error which can work against the X,Y&Z position of the return signal coinciding with the mark over which the prism is set.
as part of my efforts to reduce the horizontal systematic errors from our equipment to negligible levels i recently tested all our gmp112 prisms in order to determine a range of X & Y offsets with respect to our 3 instruments. these prisms are the 2mm spec prisms if i recall correctly. our 8 prisms ranged from about -0.1mm to 1.0mm offset for distance and up to about 0.5mm offset sideways. interestingly the locking button induced a sideways rocking movement in the prism when locked to the spigot, and in the same direction for all prisms. to counter this i pulled the prisms apart and shimmed the pivots with some plastic card to get all of the prisms within 0.2mm when sitting on the spigot. so i made the low budget prisms the same sideways centring accuracy as the mega dollar prism. for distance offset there are enough prisms that you can select one that near as perfectly matches the instrument you are using.
as far as mini prisms go, at least with the leica ATR system, if you're using a reflector of too small a diameter then the size of the prism makes the aperture stop at the prism end, not the instrument, which shifts the position of the centre of the return signal. the paper i read suggests 20mm diameter or so and mentions cat's eye reflectors and reflective tape. so if you're using ATR or autolock, a well centred 'full size' prism on a short pole should be as accurate as you can get.
If one is talking about the optical properties of the prism alone, I thought that unless the prism is nodal, you can't use the prism to sight the point (at least not manually...I can't speak to robotic operation), but that you must use the target around the prism, the plane of which is located directly above the point (within the precision of the circular bubble of course). Is this wrong?
rfc, post: 353542, member: 8882 wrote: Is this wrong?
It depends on what you mean by "wrong." John Nolton has previously made reference to a 1978 study by J.M. Rueger as well as a 1981 study by L.A. Kiviola and William A. Oren that studied the effect of prism misalignment on angle and distance measurements. In a nutshell, for a single prism (as distinct from a cluster of prisms) you have to misalign it by almost 30å¡ in order to produce a 1 mm error in distance. The effect of angular errors is dependent on distance to the prism and the amount of misalignment, but a 2å¡ misalignment will produce an angular error of 0.6" at 100 m, while a 30å¡ misalignment will produce about 12" of error at 100 m.
I fancy one of the SECO nodal mini-prisms with the changeable plumb bob point. http://tinyurl.com/j963fwy
Preferably with the bipod http://tinyurl.com/hww5x8n
I think it was John H who was using this set-up for monitoring dams?
The centring error on the mini prisms will be a result of manufacturing tolerances. Couldn't see the link to the 0.3mm prism. I think the GMP101 is the top mini prism assembly they do.
I once had a couple of GMP101 that were supplied to me for wall mounting on leica spigots. I actually found a couple of mm discrepancy in the prism constant so I stopped using them and used normal Leica circular prisms instead. The method for checking that the prism constant is correct for a particular instrument/prism is well documented. I would suggest this should be done regularly when precise radial stakeout is being undertaken.
Can you elaborate on what you mean by checking the prism constant? Are you verifying that the manufacturer's published constant is correct? Or just that your instrument and data collector settings are correct?
both, it's a check on the whole setup - instrument and prism.
as per this thread - Prism constant question
I usually do AB, AD then CB, CD.
Ie 4 points in a line A to D.
Set the instrument up at first A and then C, measure to prisms at B and D.
AD-AB should equal CB+CD when you have the constant set correctly.
AD-AB should be the true distance as any EDM/prism constant cancels out.
AB, AC and BC is another way of doing the same thing.
Jim Frame, post: 353564, member: 10 wrote: It depends on what you mean by "wrong." John Nolton has previously made reference to a 1978 study by J.M. Rueger as well as a 1981 study by L.A. Kiviola and William A. Oren that studied the effect of prism misalignment on angle and distance measurements.....
Jim
I was at Purdue when this work was done, and remember Will Oren presenting his findings on this as well as other error propagation topics. That said, I am very grateful for the Purdue Education I earned.
We regularly buy Spherically Mounted Retroreflectors from PLX
Typical issues affecting error propagation in prisms include
1. Flatness of the prism material, typically glass
2. How well the proper intersecting angles were manufactured
3. How well the prism material is centered within the prism frame.
Each SMR comes with a calibration. The worst calibration I have seen for the prism being centered in the steel ball is 0.00007 on an inch.
This discussion has been pretty good at listing several of the errors that propagate in measurement. I find it interesting that some only look at the "measurement error" cited for their total station.
