I have a question about the operation of the "slope compensation" section.
What is the expected behaviour of this feature? If someone is used to using an
RHK controller, which has a slope compensation control, how does the GXSM
feature compare to this? It does seem like something that I would like to
have, but I am not totally sure if that is a correct assumption. Certainly,
during some scans the line scan appears highly sloped.
I guess it might not really matter in the long run as this slope could be
removed in post-processing, but I would like know what the behaviour of the
slope compensation should be.
Hey, yes it does a additional slope compensation, or better I should say it
supports the feedback, as it adds the slope adjustment to Z in addition.
However, the way it works, you will actually NEVER see the "virtually" removed
slope in your scan as this (TOPO/Z) always represents the actual Z value --
including the slope any time.
How it works and how you CAN make it visible:
a) the slope correction is ALWAYS in an integral way and done in small dz
corrections to Z relative to any dx or dy (scan) motions. So NEVER a "jump"
b) A trick to make it "visible": (better make sure your tip is far enough away
1st for this)
manually center Z some how (at least it should not be fully at top or bottom so it can change) -- use small current/setpoint, play with CI until Z near center and freeze by CI=0, also put CP=0 for now.
start a (repeating) scan to play with, open the "Red Profile" to view the scan line, TURN OFF the "LR" (Line Regression!).
then enable the slope compensation and play with the slopes for X and Y
... now you should see the the slope/plane induced by doing this!
let me know if any questions!
Thanks for the tips. I can reproduce that behavior, but I guess I am looking
for a more complete idea of what the slope compensation is trying to do, or if
it is even something I should care about.
Okay, reflecting on what you have written above I think I get that I will
never see the correction, because even if the slope compensation is turned on,
the z piezo will still be moving with the added slope.
Take for instance the two images that I have linked below.
These are two scan lines (of different size) but same area. They are scans
along the positive x direction. Now I would think that this would require a
correction of either:
(550Anstrom/500mn)=+0.110 or -0.110
So, operationally, would I add this before the scan begins, or during the
scan? Is this something that would be suggested to do, even. I mean, the scan
line can look fine when I look at it with the LR clicked, but it does seem
like I am requiring the feedback to work extra hard.
I guess I could check how well the slope correction has been applied by look
at the "direct" image and not the "quick" image?
I think you have some experience with the RHK controller. On that one, there
is a x slope and y slope correction knob. So, when I see a scan line that
looks highly sloped, I think of correcting that slope to flatten it. I cannot
seem to replicate that behavior with the slope compensation as is.
Perhaps I am using it incorrectly, or maybe I am not getting it's true
Any suggestions would be great.
So I would think that the slope
This may be confusing, but this is what it does:
a) you will never see the "slope" actually removed from your data or Z and
always will see the original Z.
b) if you set and enable the slope compensation this is what happens:
You will lessen the "load" of the digital PI feedback regulator as the slope
compensation for X and Y will provide and add the proper amount to Z while
scanning -- this works seamless together with the feedback computation and can
not simply be separated.
Usually the feedback is fine with a gentle slope and you will not gain much if
However, to visualize what happens do this "self test trick" (better make sure
your tip is not neat the surface or in Z range for this):
"manually" center Z approximately and freeze it there via setting CI to zero.
(may be also CP)
If you now start a scan you will see a plain constant Z -- put the view mode
But now, enable the X (and play later with Y) slope compensation, say just put
a 0.1 in there (10% of X goes to Z) and enable it (you can do this while
scanning!) and watch what happens....
One more note:
Yes I know the RHK controller behavior -- there the analog signals X and Y are
added/mixed to Z and compensate on analog level and thus the "Z" (before
adding) and the feedback will not see the slope any longer.
As here with the DSP MK2-A810 case everything is digitally computed the DAC
has to output the signal with slope -- unless you introduce a similar analog
compensation (mixing Xmx + Ymy + Z after the A810 in analog).
But as long as the Z-range is not exceeded (excessive slope) I do not see any
need for this.
Yes, that is helpful. It took me a bit to realize that seeing the z signal
means that the slope will be there if it corrected or not.
I think I have a follow-up to this thread (regarding forward and reverse
scans), but I will wait to see what the results of the corrections are and see
if I will answer it myself.