Still chasing tail on FPU innaccuracy/logic - guesses appreciated

[Stinkhorn Press]

Inverted triloc FPU a la Alan.
Film output on 4880, all printed the same direction (horizontal).
Press RPM with central OC - accepts M&R style platens. Triloc jig effectively unmodified (welded in place rather than bolted). Triloc stop block unmodified (no it does't move).
At setting of zero OC, platens touch the screens, Triloc has smallish gap to the bottoms of the screen hangers.
Static screens, similar 20ish N tension.

We get consistently inconsistent results.
Example test run UB white and top white:
Looks "good enough" on FPU.
Prints out of register on press - we tried both screens on various heads, couple different people doing the locking in place - they all look very similarly the same out of register each time (top a little high and rotated slightly).
Wash screens and stick them back in the FPU, still "perfect."

What are we missing?

I know that DEPTH of where the three blocks hit the screen is hyper important to make sure is the same on a roller screen with shoulders BUT my "logic" says if the stop blocks are close enough to perpendicular to the screen/jig/fpu AND the square outer edges of the static frames are also close enough to perpendicular to the mesh side of the screen it shouldn't matter.
Besides adjusting for the depth of the stop, are we missing something???

[Stinkhorn Press]

more pics
the actual screens and print referenced above.

[Homer]

an issue we had, our stop blocks on the tri loc platen where hitting the corners of the screens, right where the weld is. I jb welded a piece of .040 aluminum to the platen stop block and it fixed it.

[Colin]

If Homers response doesn't fix it:

How is your screen tension between all screes again?  Below 25 newtons you can see dramatic shifts with a 2-3 newton difference.

Is the squeegee pressure different between your base white print head and the other colors?  This will exaggerate low tension screen issues.

Is pressure equal between both sides of the squeegee?  Again, low tension will magnify these issues.

[Stinkhorn Press]

an issue we had, our stop blocks on the tri loc platen where hitting the corners of the screens, right where the weld is. I jb welded a piece of .040 aluminum to the platen stop block and it fixed it.

i'm not following this. the stop blocks hit the outer edge of a static frame, couple of inches from the corner (only weld i see?), large flat-ish surface hitting large flat-ish surface(?)

How is your screen tension between all screes again?  Below 25 newtons you can see dramatic shifts with a 2-3 newton difference.

Is the squeegee pressure different between your base white print head and the other colors?  This will exaggerate low tension screen issues.

Is pressure equal between both sides of the squeegee?  Again, low tension will magnify these issues.

noted.
we print usually with 20-28PSI (RPMs might read a little low compared to say an M&R) so I am guessing that's not this issue, but I'm hesitant to throw out ANY suggestions at this point...
i can test that out with non-white inks that flow right through.

[ZooCity]

My golden pre-reg rules:

• Make sure it repeats.  Can you reload your exposed screen to the PRU and see that's it dead on?  If CTS, can you reload and reprint and have it dead on?  Can you load/unload/reload and print on press and it's dead on?
• Match the three contact points to the frame exactly on PRU and on press.  Keep off contact at loading in mind here.

Start there and see if the problem shows itself.  Good luck!

[IntegrityShirts]

Warped screen/frame perhaps? Did you see if they're both completely flat? You said you tried BOTH the screens in different heads or just the top screen?

I have one head that, when I hit the air locks, it always shifts the screen a little based on how the plungers come down and hit it and I think one of the hangers might be skewed a little.

[Stinkhorn Press]

Had a talk with 244, thanks man!
He was chock full of relevant advice - the three we seemed to have agreement my system was most lacking (talking it through, not seeing it) are:
depth. probably the largest issue. it DOES matter even for static screens. I think I'm going to raise the outer section of the fpu to hit the same points as opposed to adding shims under the platen stop blocks.
not spinning the press by the tri loc platen/jig. whoops.
checking that the screens are FLAT and that a flat screen sits flat on the screen hangers (no opposite corner movement) - else the air cylinders are going to bend one corner down if not physically move the screen.
Good stuff.

[alan802]

Use 3 regi marks if you're not already.  Yeah, 1 more regi mark 4" from the center in whatever direction you prefer helped us dial things in to a surprising degree.  Our results were consistent, and not that bad, but adding the 3rd mark made it really accurate.

[Doug S]

Does your RPM have pneumatic screen locks?  If so, try slowly flipping the switched for the locks and 1 at a time.  Just a thought.

[mk162]

Yes, the RPM has them.  We flip one side first and then the other...it helped immensely, we also put very little pressure on the screen so the tri-lock jig doesn't flex at all.

[Stinkhorn Press]

still chasing. getting closer. We mostly do 2-4 color butt registered solid prints.

My golden pre-reg rules:
Make sure it repeats.  Can you reload your exposed screen to the PRU and see that's it dead on? 

it repeats OFF press. Dead on OFF press.

Use 3 regi marks if you're not already.  Yeah, 1 more regi mark 4" from the center in whatever direction you prefer helped us dial things in to a surprising degree.  Our results were consistent, and not that bad, but adding the 3rd mark made it really accurate.

we have 4, but use three. top and bottom on centerline and one shooting off to the side in line with the top center mark.

Warped screen/frame perhaps? Did you see if they're both completely flat? You said you tried BOTH the screens in different heads or just the top screen?

we run statics. Almost NONE were flat. We're using our ShurLoc meshing jib to bend them closer to true EVERY round through.
we've tightened up our parallelism - within each head (making sure a FLAT screen has no movement and sits flat side/side and front back in the screen holders) and head to head.

I have one head that, when I hit the air locks, it always shifts the screen a little based on how the plungers come down and hit it and I think one of the hangers might be skewed a little.

we are starting to think this happens on MOST of our heads. See the video in link below. There is a switch for the left side air clamps and a switch for the right side air clamps. in the video I am hitting them individually and you can see the screen lurch when it happens.
Still not clear on exactly WHAT would make that happen - air cylinders at an angle to screen or bottom of the screen hanger c channel?
How to resolve that?
video link. straight down view:
https://photos.app.goo.gl/4DLgVpzPGX4l9dNd2

attached a pic of our screen hangers. RPM uses side hangers. micros are built into the Right arm.

[Homer]

are you holding onto the screen as you hit the air locks? We pull on the screen into the stop blocks and THEN hit the air locks. Our screens don't move at all when we lock them down

[ebscreen]

are you holding onto the screen as you hit the air locks? We pull on the screen into the stop blocks and THEN hit the air locks. Our screens don't move at all when we lock them down

This.

Evenly pull or push the screen into the blocks. Unfortunately you introduce human differences here so try and have one
person locking in all screens per job.

On the Roq's with the air clamps you can slow the clamp action down by slowing down the switch. IE slowly flip the switch
as opposed to flicking it. If this doesn't change anything on your machine, look or ask if there is a flow (preferable) or
pressure adjustment for the clamps. A slower, smoother clamping action will introduce less jolt than a fast one.

[Stinkhorn Press]

are you holding onto the screen as you hit the air locks? We pull on the screen into the stop blocks and THEN hit the air locks. Our screens don't move at all when we lock them down

in the video, no.
when setting up a screen with the jig, yes we hold. 
that is not a satisfactory answer to me though. a bandaid with flaws to boot.
1 the jig is a flimsy thing at the end of a long stick. doesn't take much to disrupt it. grip too strong and you move it before you latch. too weak and the latching moves it for you.
2 if you need to micro, releasing the left side may MOVE the screen before making any adjustments

On the Roq's with the air clamps you can slow the clamp action down by slowing down the switch. IE slowly flip the switch
as opposed to flicking it. If this doesn't change anything on your machine, look or ask if there is a flow (preferable) or
pressure adjustment for the clamps. A slower, smoother clamping action will introduce less jolt than a fast one.

that might help. no speed or pressure control built in on an RPM. binary on/off.

--------------
offline someone suggested this is just a reality of imperfect static screens.
newmans you can torque to spec and make it FLAT. a static is always a little off.