Re: Motherboard databus on a theatrical moving light working(but not quite)

Mark,

Thanks so much for your comments on this problem.

I have switched and swapped out not only the sensors, motors and drive
circuitry on the questionable board, but also have swapped in known
good parts onto this board and the problem doesn't/hasn't moved at
all. It's still all about this one color 2 motor not spinning during
its init procedure to figure out its position.

I will definately try the .1uf caps on the logic lines as I suspect
there is data glitch somewhere that the processor is reading in and
hence, thinking that it's aligned when it really isn't and thus,
doesn't even fire the motor move the wheel for positioning.

I've just not ever come upon a problem like this where the system is
99.9999% working, except for some discrete procedure.

That's what is really odd about this is that the motor drive works
just fine after the init finishes and I can move the wheel and it even
sounds and seems to keep up at the same speed as the other wheels.
The hall sensor on this wheel works correctly when I manually move the
wheel and look at the logic levels coming into the buffer. And since
there is another sensor sharing this same logic line, it always seems
to find its position so it would seem that logic line post the buffer
is being read by the processor correctly.

I will let you know what happens after I try out the caps and see if
that elicites a change or not.

Thanks again for the comments and help on this really puzzling issue.

See ya!


On Fri, 18 Dec 2009 19:44:17 -0800 (PST), Mark <makolber@xxxxxxxxx>
wrote:

On Dec 18, 12:30 pm, "JosephKK"<quiettechb...@xxxxxxxxx> wrote:
On Mon, 14 Dec 2009 19:24:06 +0100, Dan Charette <d...@xxxxxxxxxxxxxxxxxxxx> wrote:
sorry it's taken me a bit to get a reply posting going, but I did some
more experiments with the lighting fixture motherboard and thought I'd
present them for further eveidence.

The tests available in the factory menu are sorta of a continous loop
of the individual parts in the initialization procedure.  They do help
to isolate troublesome areas, like in this case, and save time since
you don't have to always power the whoel fixture up just to test one
small part.

I'll describe what's happening when I run the color 2/Rgobo test.

Under normal conditions, the light would have powered up and went
through its initialization, and the color wheels and gobos would be
positioned correctly.  When engaging the test procedure for color 2,
the LED display on the side panel simply says "ON."  Now, what this
means is that the shared data line for these two sensors is indicating
a true state, or low in this case.  If I manually move both wheels,
just a little so each of the two hall sensors are out of range, the
display says "Off"  Then, the processor will send a command to one of
the motors and rotate it until the wheel comes back into alignment.

Both of the wheels, when moved out of hall sensor range, will indicate
on the display "On" when I manually move them past the hall sensor.
It's just that the color two wheel, when "Off" just won't trigger the
processor to move the wheel, spinning it around to get it in position.

I've done some probing on the data bus ahead of the motor drive data
buffer, on the main data bus connecting the processor and the flash
RAM.  And, I can actually see activity when I rotate the Rgobo wheel
off position, then the data lines activate and the motor rotates the
rgobo wheel until it's in position.  I try this with color 2, and I
cannot see a change on the data lines.  So, this tells me that the
processor isn't giving a data write to the motor data buffer at all.

I have swapped out the flash RAM IC121, and loaded it with new
programming and no change, the problem is still there.  I've also
swapped out IC122, 74HCT574 motor driver data buffer, IC123 - 74HC541
Hall sensor data buffer, and IC126 - 74HCT138 Motor driver chip select
decoder.  

The only chip left that has anything to do with this function is the
microprocessor.  I just cannot imagine that this would be the problem.

As I stated in earlier posts, I am absolutely able to control this
color wheel motor under normal operation.  It moves, and steps as it
should.  It's just that upon initialization, it doesn't move and it's
not in the right spot for normal operation later.

My only real grasping at straws thought now is that somehow, the code
that the processor is reading for the init procedure on this one
motor, is somehow corrupted when the processor reads it or writes it.
As I said, I've swapped out the flash memory, so the chances of there
being a bad byte or two of data, in the same place would be
astronomical.

Which flash chip did you use for the master?  From the flakey board or
the proven good board?



Can a processor read something from RAM and just not write out
correctly what it read?  Without seeing the source code, I can't
imagine there being some strange sequence for spinning the motor that
is different from the normal operation procedure.  So, I would think
that in initialization, the procedure just calls a motor drive
function until the hall sensor is aligned.  So, why is the processor
just not spinning the wheel only in initialization?  That is the big
$10,000 question.

Any more ideas to try?

Thanks for thinking on this... it sure helps to bounce these ideas off
someone.  The manufacturer BTW is non-responsive.  I've sent a few
e-mails so far and not one response.

Have a great evening!  :)

Dan

On Sat, 12 Dec 2009 13:14:12 GMT, Geo <hw9j-s...@xxxxxxxxxxxxxxx>
wrote:

On Fri, 11 Dec 2009 12:56:05 GMT, Geo <hw9j-s...@xxxxxxxxxxxxxxx> wrote:

On Fri, 11 Dec 2009 06:40:18 +0100, Dan Charette <d...@xxxxxxxxxxxxxxxxxxxx>
wrote:

The question really is why doesn't that first wheel even start
spinning?  I'm just wondering if there is some data lagging on the bus
on the line that the processor is checking for that hall sensor's
feedback of correct position.  And even before the wheel starts
spinning, the processor reads a logic low, giving it a "Wheel 1 is in
position" indication.  That's why the procedure always exits once
wheel two finds its position.  At least, that's my theory anyway.

Ok -with the working motherboard - can you position (by hand) the problem wheel
to the sensor position then do the init. Does the good motherboard then behave
the same as the bad?

Not sure if this would give any info (from the manual):-
"Factory Test (SPEC/FTST): This menu provides an effects test (ETST), a movement
test (MTST), and a sensor test (STST) used for quality control. The sensor test
includes programs for testing sensors on the color and gobo wheels (COL1, COL2,
Rgob, and Fgob)."

Dan Charette {dan_at_thesonicfrogFUZZ-dot-com}
Remove the "FUZZ" and replace the underscores and
such from my e-mail address to contact me.

"I may not always be right, but I'm never wrong."



to OP..

you seem to have a non obvious problem like a marginal logic level or
switch bounce or something that doesn't show up obviously...

since you have 2 wheels and 2 sets of parts, if you can't puzzle this
out, I suggest swapping parts around good to bad till the problem
changes and "follows" a bad or marginal part...

If swapping a part suddenly fixes it completely, then you have a
combination of marginal things..

I'd also be tempted to add 0.1uF caps across all the logic signals to
eliminate glitches. There may be a 1us glitch pulse that the SW sees
that you don't

Mark




Mark

Dan Charette {dan_at_thesonicfrogFUZZ-dot-com}
Remove the "FUZZ" and replace the underscores and
such from my e-mail address to contact me.

"I may not always be right, but I'm never wrong."
.

Relevant Pages

  • Re: Motherboard databus on a theatrical moving light working(but not quite)
    ... I had replaced the flash memory on the flakey board with a known ... Both of the wheels, when moved out of hall sensor range, will indicate ... It's just that the color two wheel, when "Off" just won't trigger the ... I've done some probing on the data bus ahead of the motor drive data ...
    (sci.electronics.design)
  • Re: Motherboard databus on a theatrical moving light working(but not quite)
    ... Both of the wheels, when moved out of hall sensor range, will indicate ... It's just that the color two wheel, when "Off" just won't trigger the ... spinning it around to get it in position. ... I've done some probing on the data bus ahead of the motor drive data ...
    (sci.electronics.design)
  • And it continues to get more strange and puzzling.
    ... Both of the wheels, when moved out of hall sensor range, will indicate ... It's just that the color two wheel, when "Off" just won't trigger the ... spinning it around to get it in position. ... I've done some probing on the data bus ahead of the motor drive data ...
    (sci.electronics.design)
  • Re: Motherboard databus on a theatrical moving light working(but not quite)
    ... Both of the wheels, when moved out of hall sensor range, will indicate ... It's just that the color two wheel, when "Off" just won't trigger the ... spinning it around to get it in position. ... I've done some probing on the data bus ahead of the motor drive data ...
    (sci.electronics.design)
  • Re: Motherboard databus on a theatrical moving light working(but not quite)
    ... of the individual parts in the initialization procedure. ... It's just that the color two wheel, when "Off" just won't trigger the ... processor isn't giving a data write to the motor data buffer at all. ... Hall sensor data buffer, and IC126 - 74HCT138 Motor driver chip select ...
    (sci.electronics.design)