Eliminating Prism Errors - SURVEYING TRAINING 

Prisms in this case, are actually retro reflectors. As such, they have a certain depth to them in relation to their size. The prism constant is an offset to the center point of the reflector, in relation to the center of the prism pole.

Look up something called "nodal prisms", and non-nodal prisms. 0mm offset is what Leica tried to do, when all their true prism constants are -34.4mm . But their software will do the math with the assumption that 0mm = -34.4mm. In effect, it's like saying "Water freezes at -34.4 C, but 0 is a nicer number, so lets work off that instead". Except they have other prisms that aren't their new 0, and have other numbers on them. For example, as mentioned in this video, he likes 17.5 as being more accurate, when inactuality that's just a simplification. 17.5mm for leica, its true prism constant is -16.9mm. The main reason why different prisms exist that have different prism constants is largely because of choice, and cost. It costs more money to make a prism constant bang on exactly 0, and secondarily, even if you did a 0 prism offset wouldn't be ideal. When you have a prism offset (true)-typical of Lieica Mini's GMP101 or GMP111-0 (or maybe it's GMP111-1) of -16.9mm, because they are nodal, AND they have low centering error, if your prism isn't aligned perfectly with your total station, like perfectly square targeted, and bang on, a 0mm offset prism will have far more errors based on that angle deflection. A way to illustrate that point, is the put your total station in layout mode, keep it in tracking if you want too. When your;e staking out a point, keep your crosshairs fixed (if using a robotic, keep autolock off), and just rotate the prism and watch where it's telling you you need to go. On a nodal prism (-16.9mm, +17.5mm-Leica) -mind you there are other nodal prisms that do have different prism constants- you'll notice that if you rotate the prism either left or right 10-20 degrees, and are still able to get signal from the EDM, that it will tell you to go away, or left or right more or less than a few mm than when you're bang on lined up perfectly square with the total station. That means a few things: 1. You should not be doing robotic total station accurate work with a non-nodal prism. I'm not going to debate this with anyone. I've tried it, its too unreliable. And that prism will not have a 0 mm constant. 2. You can absolutely use a non-nodal prism if you're doing things the old fashioned way for high precision layout (IE: You have someone looking through the instrument, and aren't relying on autolock servos, that can tell the guy "hey, straighten out your prism". 3. You don't need to spend as much money on backsight targets with non-nodal prisms, but you really should make sure that you square them up with your total station, even if they're far away to get the most accuracy. You will still need to do the same thing with smaller prisms too. You do also have to consider that EDM's will get a bit wonky the farther away you're shooting, that's why I don't necessarily agree with the idea that smaller prisms are always more accurate, I'll argue that they are for layout,