Ye old *axis following error...

Ok, so I've been getting x or z axis following errors. 
I can't see any errors from bios expeditions 
Shielded cable used.
Using a pair of 8i20 cards for the servos. 
Modified fanuc servo motors with a,b,z encoders - I do the thump on startup.
it seems to do errors at any speed fast/slow. I can't pin down what the cause is. Tried a multitude of servo tuning settings,  still does the same. If it was a bios expedition, is there a way to track this, as running axis from a terminal dosent show any errors. 

Can you provide a halscope plot of the velocity and following error?
 

I'll organise one, the error that occurs is usually during a rapid movement not possible to catch and map usually as it occurs randomly.

You can trigger halscope on a following error to capture the actual error

So, I was checking halscope and modded the "P" value higher - which gave less error. Then i found that It was running off more frequently. So, ive had too high a "P" value. The reason for this really is when I've looked up pid tuning of 8i20 cards you've reccomended setting as high a D value as is stable, but there's no range stated, - is 3 a high value? Is 500 a high value? I don't seem to get good position holding with a high d value as P generally ends up around 185, which you have said before is pretty much saturated.
I'm running a high thread rate 225000 off the top of my head.

I would add as much D term as possible until you reach instability
Have you added FF2?  (the FF2 value depends on the system inertia)
 

Ok, so what do you mean by instability?
D tends to make it be slow to reach position once over 25, with d at 25 and p at 285 it holds position very well on movement. The problem is while on rapid movement it tends to stick and give a following error, seems to be on rapid moves, - there's no error bit that sets if the 8i20 goes overvoltage is there? - straws, I have no braking resistors fitted and wondering whether this may be an issue. The bus voltage doesn't seem to track well with halscope.

By instability I mean you will start to seen oscillations if D is too high
Once D is as high as possible, you can do the same with the P term
(raise it until you get oscillations and then back off about 30%)

What is the issue with plotting the bus voltage?

Can you provide a plot of the position error and PID output?
 

After much stuffing around. I believe I found the problem.
The machine I have went thru a flood - it was 4 metres underwater- about 4 and a bit washing machines in American speak.
There's a rust mark in each ballscrew where it was parked. It plays up rapiding in these areas

Boy I've got a lot better at tuning my 8i20 cards though!