Still chasing tail on FPU innaccuracy/logic - guesses appreciated

[ebscreen]

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.

We don't use statics but I think it's the reality of the process regardless of frame type.
Hell, with rollers you are clamping on a round surface. Without lateral forces holding it, stuff is gonna
move until it finds home.

Your movement seems pretty significant though. I'd look at a few things, namely if the frame holders
are parallel to eachother, and the condition of the clamping bars, make sure nothing is binding and the screw
are bottomed out.

[Stinkhorn Press]

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.

We don't use statics but I think it's the reality of the process regardless of frame type.
Hell, with rollers you are clamping on a round surface. Without lateral forces holding it, stuff is gonna
move until it finds home.

Your movement seems pretty significant though. I'd look at a few things, namely if the frame holders
are parallel to eachother, and the condition of the clamping bars, make sure nothing is binding and the screw
are bottomed out.

grass is always greener! ha.
truly groking what 'in parallel" means and requires takes some time. Hangers are pretty true AND checked pretty much weekly at this point just to see if/what shifts over time.
no clamping bars on RPM, a pair of air cylinders with a thick stem and a screw on "foot" - so four individual points per screen.
what screws bottomed out, the ones bolting the cylinders to the screen hangers?

[ebscreen]

If'n it was me, I'd find the flattest thing I could that would fit in the head, clamp it on one side, then
adjust the other to match. Making sure you're parallel to the pallet of course.

The feet are likely the biggest issue though. By screws I meant the ones holding the bar (or in this
case, foot) to the air cylinder piston. If they are protruding you're going to be knocking stuff all over.

A bar retrofit should be pretty easy and would provide a more distributed clamping force across a larger
area. Literally a piece of aluminum stock with two recessed holes drilled through. Rough up or groove the bottom.

[Stinkhorn Press]

John Diehl stopped by. Thanks John.

Had a good, obvious point - but one that I hadn't figured out. He sells CTS machines. Part of the sales pitch is improved registration accuracy, which led to his point:
can you load the same screen into each head on your press and have it print in the same location?

we're printing vertical and horizontal lines on pellons from head #2. then using the SAME screen, printing a 2nd color from every other head.
and then we're doing it again to see if we get the same results.
and then maybe doing it using only ONE platen to get that variable out (requires many more TriLoc-ings to make that happen).

we're printing .25 point lines.
HOW CLOSE is GOOD ENOUGH for a test like this??

[Colin]

IMO - Dead F*cking Nuts

Thats why you buy a registration system right?

This is of course removing the differences in mesh tension/squeegee hardness/pressures/print stroke speed/ etc from the equation.

I would love to see if any one has measured what those variables add up to - minus screen tension - for registration.

My 14 color M&R has almost never been dead nuts (or close enough for my picky sensibilities.  Sometimes its WAY the f*ck off) but my 12 color is always very close if not dead on..... 

M&R Iimage here....

Realistically.... I would want all the lines edges to at least be touching.  No more than a .25 point off.

[Stinkhorn Press]

UPDATE.

90% there.

when we started this process, we averaged 12 minutes per screen to setup color (total setup time from empty press to approved test print divided by # screens). we're slightly over 5 minutes per screen at this point. YYYYYYUGE improvement.

what worked:
adding in everyone's advice 30%
getting the press in decent parallel 70%

special THANKS goes out to:
alan at SRI - for proof of concept!
John Deihl @ Douthitt - for pointing out the obvious way to CHECK for press accuracy (see last post)
Rick Fuqua - for his time explaining how he parallels his presses (we run his press)

you can, get reasonable accuracy from film + fpu + triloc + press in parallel.
give us another month and I'll put our system up against your CTS for accuracy. it's that good.

pictures: example of our test print to check parallel.
print the ONE screen on ONE head on all platens with black ink.
move the same screen to each other head and print red ink on top the black. (one head per platen).
not perfect. not "dead nuts" on yet. heads not perfect. we still tape film (not that hard guys, just a time investment).

the MOST any head is off at this point is 1/32" most much less than that. WOO!

[Stinkhorn Press]

re-reading the thread - shout out to ZOO for properly diagnosing from the get-go as well - he noted we needed to print in the same location on each head each time in his reply, but it took us talking to John to comprehend a good way to actually TEST that!

[mk162]

Nice!  We gave up on leveling our old RPM.  The new version of that press is INSANE.  I will be doing a full write-up on it soon.  If you like the RPM, you should check the Vector out.  It's got some nice upgrades.