Problem Proving Flatness with laser tracker vs dial gauge

Hi everyone im new here and would love some help i have a pretty dumb question but wanted some others input on the issue. We had a project that involved checking the flatness of a slew ring about 4 metres in diameter. Our customer had been told to install hydraulic jacks equally spaced and hold it under equal pressure this was because they did not want to remove 220 tonnes of equipment that sat ontop of the chassis. We inspected it on 8 different occasions after welding and other processes. We also made them clean the surace up to almost a mirror finish. Other things that were making me not confident with our accuracy was the fact we were on brick pavers. On the end of an airport runway (vibration) and had heavy equipment cranes, trucks and forklifts all around us which caused vibration intermittently. So now you now how bad the controls were... oh wait and it was outside in the australian heat... think thats most of it.
So we were happy with its flatness when it was first inspected prior to welding then we saw twist/pulling where we expected after gauging and welding. This required the need for milling it back into spec which is 0.2mm flatness over the entire 4 meter diam. So a team was contracted to use a large climax portable mill and machine it flat... this is where it gets hard to explain. We went in and showed the machinists our numbers and where it was warped. Then we left and returned to check it after its first pass. The item was 0.4mm flatness so we said sorry you have to remachine it however they then did not trust our laser tracker because they couldnt see the 0.4mm error on theyre dial gauge (fitted to the mill and run around the surface) i explained that the dial gague wasnt the right way to check it as it is not checking 2 virtual planes in a minimum/maximum or lowest - and highest +. Suffice to say the machinist didnt agree so we checked it 12 times in different setups to prove it was 0.4 eventually we said just trust us remachine it and lets see if it gets better... it did 0.14mm flatness. So how do i explain why they cant see 0.4 on the dial gague and am i right in saying the dial gauge is not helping of mounted to a machine ( which i discovered had 0.1 of play in the tool height) on a side note the machinists are well known to be the best where i live and did great considering it was theyre first time using a climax mill and furst time machining a large slew. I cant fault theyre work i just want to be able to give them feedback so nextime they understand.

I'm not quite understanding why an indicator couldn't check this part. In this particular instance, using a portable mill to check its own work is suspect, but I'm not understanding your more general statement. If you lock the Z-axis, you'll be measuring in Z from the mill's XY plane down to the surface. You should end up getting basically the same result as you get with the laser tracker.

Yes, there are some slight differences depending on how you post-process the tracker data, but the indicator data will, excluding the errors of the mill itself, be more conservative. Because you know the SMR locations in 3D space, you can find the minimum distance between two planes that will contain all of your points. Whereas with an indicator, you're actually finding the maximum and minimum Z distance between the bearing face and the mill's XY plane. If these aren't perfectly parallel, your flatness reading will be high.

Laser trackers are great, but in field conditions, you're approaching their limits. This is especially true when measuring a flat horizontal surface. Many trackers are more accurate in distance than they are on angles, and measuring a flat horizontal surface, the measurement you really care about is coming from the elevation encoder. There are laser devices built for measuring flatness of surfaces and flanges, so if you do a lot of this type of work you might look into them. They're not as versatile as a tracker, but they're more accurate for what they're designed for.

What brand/model of tracker are you using?

You can explain by saying that if the circle that the machine swings is not entirely flat, then mounting an indicator on THE SAME arm that holds the cutting head will only MIRROR the warp it's cutting (if it is indeed running a gentle "wave" around the 360 of the slew). If you are in car driving up the mountain, and you are leaned over with your fingers touching the pavement, does that mean the road is not rising? It's about using the indicator off of a dead flat reference plane.