#VLAB vs. Flyer
20 messages in this thread
Ed,
I got the system back today with 4.04 loaded. There appears to be no difference
but this may be "correct" for these reasons:
The motheboard swap seemed to be related to soem gvpscsi board level errors
that apparantly were preventing NT form getting a good Flyer test environment.
I had seen these errors intermittently when the system was hot, yet idle. It
apparently was worse in Topeka.
The "Crouton Failed to Load" and disk lock-up problems, I believe, are
scheduled for 4.05 fix, so I would not see any difference in that area at the
present time.
The failures associated with the digital effects (drive:NewTek/Effects/Wipes
drawer) were, as I recall attributed wo problems on my Toaster board which they
supposedly fixed. However, I see no difference there, and hav EMail to Chuck
to see if there is also a 4.05 software component that needs to be in place
before those no longer act whacky.
wmc – via Autopilot!
Actually the drive and Crouton load failure stuff cleans up if you turn off the
CPU caches. Our code will be fixed for this in 4.05, but in the meantime
killing those caches is the (performance penalized) workaround.
Your bars and tone projects used a particular effect that breaks up and flys
out like the color bars dividing up. When the project was played sync was lost
on the next clip. This was due to the Center Frequency circuitry fault, which
was repaired. So, play those projects, (with caches off), no loss of sync
after the effect executes.
I'm not sure what percentage of the effects are currently functional. The
constallation of which effects work and which don't changes with every
significant change in Flyer code. The whole set will be remade for Flyer 4.1.
So, if you've been running with caches on since you got the machine back, the
CLF and drive lockup stuff will look just like it always did. I did not change
the default in your user-startup. One of your clips was corrupt as well, you
mentioned this in private mail as well.
Up in second paragraph "play those projects, (with caches off to prevent CLF
and drive lockups)"; please apply the change. Caches off or on would have no
bearing on whether the effect causes sync loss or not. Sorry my first phrasing
was confusing.
If the effect does cause loss of sync, then you may need to set center
frequency again. cfadjust.txt is in the library.
Chuck.
After I got your Email I tried that and it did solve the CLF/lock-up thing. I
remember you told me that on the phone, but lastnight when I tried it I did it
after I had the Flyer loaded and it didn't quite do it. (Or maybe I hung on
that large damaged clip and mistook it for caches off not making a diff – don't
remember for sure). But that makes life ALOT easier.
It did force me to redo the timing on a CG project since the page load times
were impacted, though, so looking forward to the perma-fix.
the CBARS explode works fine (as it did when it left here – mor shipping damage
I guess). The ones I specifically had and stillhave trouble with are in the
Toaster:Effects/Wipes drawer. In particular the ones named Wipe B-T and Wipe
T-B. I believe these are among the ones you mentioned would have to be redone.
These two (or at least one) are real handy for comparing sources (like TBC
output to TSG output for color balance and alignment). Hope they get fixed in
4.05 (hint).
But, I'll try CF adjust tomorro and your other Email suggestions to get
restarted and see how it goes. I think the effects I am having problems with
are the standard "broken" ones since many other effects in this drawer do work.
So if I get the luma levels balanced again I'll be relatively happy as it will
mean pretty much all the problems I have remaining are on your fix-it list.
(tired of me and my machine yet? ;^}
wmc – via Autopilot!
The effects guy is not ecstatic about having the redo the whole pile, but
that's life! Yes, I think at this point, except for the corrupt clip, the rest
of what was going on is on our fix-it list.
I'm still just relieved how much the motherboard exchange cleaned up!
Chuck.
>>I'm still just relieved how much the motherboard exchange cleaned up!
It must have been another shipping effect, because if I leave my caches on I
see no differences. The SCSI board errors came so in frequently and seemed to
stop altogether when I swapped out the oldest HD that I thought all that stuff
was O.K.
Its amazing to me how much stuff was loosened and shifted around even when it
was all double boxed. Even on the main unit the right sided of the floppy and
drive adapted into the 5.25 bay were pushed in far enough to have about 1/8
inch daylight shinning into the box along that edge. I had to re-adjust their
position on the PS bridge when I put the MegAchip back in. I also got a fair
amount of movement on all the socketted chips when I pushe on them and Agnes
came out without a scuffle. It must have undergone a nice vertical drop to the
floor/ground at some point in transit.
Anyway, I ran all the diags on it and it checks out OK – but then so did the
old board before I sent it so, who know. Seems to be running OK and that I
guess is the best measure.
wmc – via Autopilot!
Wayne,
<<<Its amazing to me how much stuff was loosened and shifted around even when
it was all double boxed.>>>
Have you ever been there when one of the major carriers (UPS, FED-X etc.)
loads, or unloads a truck? You'd swear you were at a basketball practice!
Something should be done about it.
Regardless of how it's packed, if you toss a box that weighs 80 lbs ten
feet, from the back of a truck that is five feet off the ground, onto a hard
cement surface!!!!!, well you get the picture.
Good luck with your system. ;^) (I'm not being sarcastic)
Take care,
Mike,CyberGrafx.3D
Yeah, whent they delivered it here they guy cartwheeled it like a spare tire
through the door then tipped it sideways and let it flop down onto its bottom.
I was going to say something, but it was already in, plus it would porbably
have only garunteed harsher handling of my packages in the future. As brutish
as some of those people are, they have remarkable memories. :^}
wmc – via Autopilot!
Here, caches on or off changed what was happening with the Flyer, but made no
difference to the constant GVP SCSI.device errors, with lockups on the system
drives, often followed by system crashes. Even taking the Flyer and Toaster
out, and commenting out the Flyer init material, the gvp scsi kept erring out.
I covered other items in the previous private postings, as well.
Chuck.
I think something that must have been marginal pooped out in shipping. I only
had about 3 instances of the gvpsci errors in as many years, and usually once,
then reboot and I wouldn't see it again. The last time it happened to me I got
it 3 times in one session about 30-45 mins apart, but it was about 90 in the
studio so I turned it off, let it cool and for good measure changed out an 8
year old quantum drive that I figured was possibly contributing. That was
about 10 days before I shipped it and I didn't see the error at all after that.
(Air conditioner was installed about then too so I was never sure which did the
trick).
wmc – via Autopilot!
I'd have to agree from the sound of things, shipping must have finished off the
weak links on the way here, and it sounds like we are just lucky that shipping
didn't wipe anything out on the way back. Given that, I'd wonder about solder
joints on the CPU bus connector, problems in the CPU slot support circuitry as
far as the GVP errors.
Given the other items you listed in the history of the motherboard, I'm not
sure whether it would be worth the expense to try to revive as a backup, but it
may be possible.
Chuck.
If I do anything with it I may try to find a service shop that will give me
afew bucks in trade for a new/refurb one.
wmc – via Autopilot!
Trading for a refurb would be a good choice.
Chuck.
Chuck,
I did the re-autohue, recal etc. No real change in the luma shift and CG page
load flashing. But, in the process I found the Toaster is MAJOR hosed up.
With CBars input to input 1 direct froma camera, I autohued the Toaster (under
3.1/3.5 Toaster software both with and without the Flyer connected and the
FLyer init in user-startup). After autohue and starting the toaster I selected
input 1 on the program bus. Boy, were the colors off! Then I loaded CBAR
graphic into both DV buffers and did a half wipe between them using the "Wipe
Up" effect (normally proj 11, bank D, 34). BIG difference in the display.
I hooked the program out to the WF/VS and found that the CBars from the camera
when selected on the program bus, were 30 degrees off (that is, to line up the
color points in the boxes, I had to rotate the burst line 30 deg below its
correct horizontal position). The levels and relative distance between the
points was right on, however. With the CBar graphic in both DV buffers, the
levels on the VS were lower than I've seen them before (1/2 to 1/3 of the way
between the "X" and the lower edge of the box) and switching betwee DV buffers
shows about a 5-10 degree CCW rotation of the vector. DV1 appears to have the
correct vector angles, but DV2 is 5-10 degrees off.
Then if I look at the WF with the CBar in both DVs, the levels are about 3 IRE
low for DV1 and the tops of the bars in flat or low pass mode ar slightly
"bowed" – not nice and flat like I'm used to seeing from the Toaster. When I
flip to DV2, the bowing shifts to the left and the bars compress by about 5
IRE. That is the tops of the bars come down around 2.IRE and the bottoms of
the bars go up by 2.5 IRE. Looking at the wave form of the program out with
the camera CBars input (1) selected, the levels appear depressed from the "raw"
camera to WF/VS display by about 3-4 IRE.
Do we need to RMA it and get the to revisit the repairs or can these be
adjusted here by some "sneak" screen facilities? (I don't see any other
adjusments on the Toaster besides CF.)
SPeaking of CF, if I adjust it correctly, it seems to drift upward at about
1 V/hour (e.g. from -2.15 to -2.05 in the span of an hour). I don't know if
there is a lwer limit, but since the toaster doesn't loose sync, I expect it
stops prbably around 2.0.
NOTE: re-enabling Flyer init in the user-startup, killing the cache/burst of
cpu and then running autohue and flyer cal for the FLyer, then doing the same
WF/VS tests as above shows similar results. The camera CBar signal looks
perfect when taken direct to the WF/VS and nearly so when passed through the
Toaster 4000 in the other box and tested off its program out.
This leads me to believe that when they put the new Aircap on and the new
resistor they did not adjust for proper phasing and level matching or inputs to
outputs and between DV's.
BTW, do I still qualify for RMA on the Seagates eventhough the drives were part
of a Safe Harbor Flyer drive/tower package that was put together after NT
stopped certifying HD's?
If the Toaster needs an RMA (adjustments cannot be done by me), and I do
qualify for the Seagate RMA's what is the possibility of arranging RMA's so
that the repair periods run in parallel, rather than serial? (I do have the
original Seagate drive packaging, BTW).
wmc – via Autopilot!
Wayne, something just struck me while reading about the fouled-up Toaster. Have
you tried going into the special setup menu? You have all the necessary gear
(ext. bars and WF/VS). I should think you could clear up all the faults
(except the CF adjust, of course)…
Please excuse this if you've already tried it.
Trent/ReAnimators
Trent,
It's an original Toaster board. I don't think it has a set-up screen like the
Toaster 4000, does it?
wmc – via Autopilot!
I don't think the original boards had the setup feature. You could try hitting
the key combination and see what happens, but if I think back to when I had my
original board as well as the T4000, I seem to remember nothing happening…
Trent/ReAnimators
You are right – nothing happens.
wmc – via Autopilot!
The baby is definitely misbehaving. I would recommend RMA, from what you
describe; yes, we can do it when your drives are out for repair. My
understanding regarding the Seagates is that any Seagate 4G or 9G of correct
model and purchased for Flyer use is eligible for the program. I will
double-check on this and advise you.
Meantime, if you want to try it, rt-alt/rt-shift/F10 simultaneous gets an
adjustment screen in 3.x software, from Switcher. It's manual version of what
Autohue does.
CF is best done after the equipment is thoroughly warmed up, as is AH and the
rest. If you do this, and still see drift, there may be further problems in
the circuit.
Chuck.
Chuck,
I think the RA-RS-F10 only works on Toaster 4000 boards. At least this is true
of my boards (unless the 3.1 load on my Toaster original board set-up is
faulty).
wmc – via Autopilot!
OUch! I do believe you're right, that's only in the VT4000 switcher code.
Sorry!
Chuck.