If it isn't obvious, the process in which I used to machine the dogs is the simplest and most prone to accuracy problems with the end product. I turned the shank of the dog to the MFT hole diameter and did not machine the head (no skim cut) which relies on the consistency of the diameter of the material AND the accuracy of the chuck on the lathe.
For this exercise I checked all six and found their OD's to be within a thou. An up-close of one of them.
In thread, Sapper indicated that the shoulders of the part looked like they weren't even. Here's a single dog from 4 different angles. The shoulders look (visually) to be identical.
If they were machined correctly, there shouldn't be a visible difference from any angle.
Not to check for concentricity. I chucked up one of them and set up the dial indicator. Spun the chuck while the indicator rides on the unmachined surface marking the high and low spot.
The deviation between measurements is ~.003". This indicates that the material (extrusion) is out of round.
To verify the part is in the chuck straight (as possible) I move over 1/4-inch and re-measure. If the part is straight in the chuck, the measurements will be the same. They are.
Without removing the dog from the chuck I relocate the dial indicator to the machined surface and re-test. If the spots on the machined surface have similar high and low spots then the part is slightly offset in the chuck.
The difference between high and low now is only .0005". That means the machined surface is not only in the chuck straight, but machined true. This also means that the extrusion isn't extremely accurate (as far as a machinist thinks). As a wood worker, it's darn good. I ran similar tests with half of the others with nothing interesting to report.
If my dogs are straight, I can assume them a constant and do some hole comparisons on the MFT.
I spent about an hour moving around the MFT trying many different locations checking to see if I could get a 3-point test to fail. Of all of the tests, so only found one on the surface that stood out as suspect. The pencil is pointing at it in the pic.
I moved the dogs around it and he led it from all sides.
The results are as you would expect if the hole wasn't in the correct location. I switched dogs out and rotated them as well to ensure the .003 run out of the dogs wasn't messing the test up. No matter what I did, I cannot get the square to seat on three points if one of the holes is this one. I found no others that were obviously off anywhere on the table.
I'd probably recommend checking the holes and dogs to ensure the ones you want to use are straight the first time you use them. They may not be.