24" Widescreen CRT (FW900) From Ebay arrived,Comments.

Wanted to pick some brains here about an interesting challenge.

A large part of the WinDAS calibration involves adjusting parameters so that the image dimensions match specified values (in mm).

So for example, it may ask you to increase the horizontal size until the image is 474 mm wide.

I've found doing these measurements extremely tedious and time consuming, because of the parallax error issue. You need to ensure that your line of sight is perfectly centered the point being measured. You can easily see what I mean by doing the following experiment:

touch the screen with your finger, and try to line up the center of your fingernail with a point on the image (choose an edge of a displayed line). Now move your head around. Notice how the finger no longer lines up when you change your line of sight?

The solution I've been using for now is to hold a straight object, like a chopstick, against the screen, so that the end of the chopstick is flush against the screen, and centered on the point I'm measuring. I then use the chopstick as a line of sight guide and read off the measuring tape accordingly.

This is extremely tedious, however, and requires a lot of dexterity and flexibility!

I had another idea that might work. You'd have a clear plastic material with a fine grid structure printed on it, with millimeter markings - kinda like a "two dimensional" ruler. You would somehow attach this grid onto the screen so that it doesn't move. It doesn't have to be flush against the screen, but it would be better if it was.

You'd then be able to take measurements without having to hold a tape measure against the screen (which is actually the most tricky part of the whole process). You'd be free to move your body around so that your line of sight is matched up to whatever point you're interested in.

And, this would have an advantage for geometry adjustments. You could easily assess whether the lines are parallel to the bezels by using the grid as a reference.

Any ideas on where one could find such a plastic grid (or any other ideas on how to deal with this challenge)?
 
Kermie said:
I read from an eBay member that the value could be read by using WinDAS, realizing this I took a look at past documents and found that I had documented the number of hours the monitor had been running.

You cannot believe the number of modified G2 dats that I have.
Click to expand...

wait, so you were able to find the part of the dat file that referenced the number of hours? Can you share that?!
 
Hi guys.. I'm using custom resolution Utility to try and get specific Display modes on the FW900 to work

The FW900 manual specify horizontal pixel clock in usec... the CRU utility uses pixels for the horizontal setting..

How do I convert usec to pixels...
 
xln47862 said:
Hi guys.. I'm using custom resolution Utility to try and get specific Display modes on the FW900 to work

The FW900 manual specify horizontal pixel clock in usec... the CRU utility uses pixels for the horizontal setting..

How do I convert usec to pixels...
Click to expand...

Please see my post on this very question.
 
You haven't provided any relevant details that would help us answer your question. For example, what resolution are you using, and can you point us to the settings in the manual that you're trying to use.

edit: ok good :)
 
Last edited:
spacediver said:
You haven't provided any relevant details that would help us answer your question. For example, what resolution are you using, and can you point us to the settings in the manual that you're trying to use.

edit: ok good :)
Click to expand...

yea mode 5 is the 1920x1200x85

I got the monitor to spill it..

But it's weird that I can notice flicker at 85hz.. but none whatsoever @ 96 hz..

I mean.. seriously.. 11hz makes the difference? my brain breaks at 11hz?

Damn you brain, you suck...
 
Maybe you're extremely sensitive, or there is something else going on. I'd be curious to see if the issue persists if you move the monitor to a different location and plug it into a different circuit.

Also, maybe bring in a second observer and see if they perceive the flicker.
 
xln47862 said:
yea mode 5 is the 1920x1200x85


I mean.. seriously.. 11hz makes the difference? my brain breaks at 11hz?
Click to expand...

that's not the right way to think about it. It could be that your threshold for flicker sensitivity is at, for example, 88 hertz. Any thing at or below that is perceptible to you and anything above it is not.
 
spacediver said:
that's not the right way to think about it. It could be that your threshold for flicker sensitivity is at, for example, 88 hertz. Any thing at or below that is perceptible to you and anything above it is not.
Click to expand...

ah.. quite correct... I did not test anything between 85 and 96... :D

New question.. I'm full of them..If I want to create a mode. 1680 x 1050 @ 100hz. How does one determine the optimal FW900 Timings..

The mode 5, 1920x1200x85 and mode 3 1920x1080x60 and mode 4 1600x1024 seem pretty close to the CRT GTF equations.

But the mode 8/9 1920x1080x72 are very different. so it would seem optimal timings do not adhere to GTF in specific Vertical refresh rates.

If I had more data points I could do least square and get a line, but there's not enough to tell me what the engineers were using to determine the timings..

Thoughts?:eek:
 
Last edited:
I was able to successfully sync using Nvidia's advanced timings - the built in GTF formula calculated these parameters:

wvzupu.png
 
spacediver said:
not that I'm aware of. I was replying to xln's request (see the post directly before the one where i posted the image of the timings)
Click to expand...

Is there a particular "reason" for 1 FP 3 SW.. vs 3 FP 6 SW

And the Monitor doesn't seem to care about sync polarity.. but why - + in this case, while most other modes in the screen are - -

I tried all these combinations.. they all "work"... but I'm just curious as to why mode sony determined 3/3 for mode 8/9 (72hz) and used 1/3 for 60hz and 85hz..
 
I've always been curious about the polarity issue. As far as I understand it, it's just a convention about how the display interprets the voltage signal during the sync pulse. So long as the monitor is able to interpret the signal correctly, it shouldn't make a damn bit of difference in the resulting signal, although I could well be wrong.

As for the FP and SW, I don't have a good answer, but you might gain insight here
 
I think vertical sync width refers to the number of horizontal scanlines that the beam undergoes while it is resetting from the end of the last horizontal scanline back to the top left of the screen.

In other words, I don't think it necessarily goes back in a diagonal path, but rather zig zags,

You don't see any retrace lines because this occurs during the vertical blanking interval.

The Vertical front porch is the number of (horizontal) scan lines that the beam undergoes after the last active horizontal scanline, and before the vertical blanking interval.

So think of it this way:

the beam has a specified horizontal scan frequency: the number of times a second the beam can move from left to right and back to left again.

The total number of times it does this during one refresh cycle is the back porch (the lines it does before the active lines), the active lines, the front porch, and those that are done during the vertical blanking interval.

My guess is that if you tried to change the vertical sync width, while keeping everything else constant, this would mess up the timing, as the beam would have to operate at a different frequency during the vertical blanking interval in order to maintain the proper refresh rate.
 
Last edited:
spacediver said:
I think vertical sync width refers to the number of horizontal scanlines that the beam undergoes while it is resetting from the end of the last horizontal scanline back to the top left of the screen.

In other words, I don't think it necessarily goes back in a diagonal path, but rather zig zags,

You don't see any retrace lines because this occurs during the vertical blanking interval.

The Vertical front porch is the number of (horizontal) scan lines that the beam undergoes after the last active horizontal scanline, and before the vertical blanking interval.

So think of it this way:

the beam has a specified horizontal scan frequency: the number of times a second the beam can move from left to right and back to left again.

The total number of times it does this during one refresh cycle is the back porch (the lines it does before the active lines, the active lines, the front porch, and those that are done during the vertical blanking interval.

My guess is that if you tried to change the vertical sync width, while keeping everything else constant, this would mess up the timing, as the beam would have to operate at a different frequency during the vertical blanking interval in order to maintain the proper refresh rate.
Click to expand...


Very informative.. I think I will use 1/3 even though cru spits out 3/6 on default. Maybe this is because the fw900 is faster than assumed
 
Can you take a screenshot of the values that CRU spits out? And do you get a successful signal with those settings?

I'm guessing that the horizontal frequency will be higher than the one posted in my screenshot.
 
Last edited:
spacediver said:
Can you take a screenshot of the values that CRU spits out? And do you get a successful signal with those settings?

I'm guessing that the horizontal frequency will be higher than the one posted in my screenshot.
Click to expand...

Everything is identical to your nvidia output with the exception of

Front Porch: 3

Sync Width: 6

It also sets Back Porch to 312-pixel Horizontal, 56-lines Vertical.. Which I do not see in the nvidia menu.
 
yep, the horizontal back porch is 312 in my case. You can calculate it easily starting with the total horizontal pixels, and subtracting horizontal active pixels, horizontal sync width and horizontal front porch, and seeing what remains.

So total horizontal is 2304

2304 - (1680+128+184) = 312.

And for vertical:

back porch = total - (active+sync width + front).

In my settings, that would be 1112 - (1050+1+3) = 58.

In your CRU settings, that would be 1112- (1050+3+6) = 53.

You said it calculated 56 vertical lines of back porch. Are you sure the total pixels was the same as mine (1112)?
 
spacediver said:
yep, the horizontal back porch is 312 in my case. You can calculate it easily starting with the total horizontal pixels, and subtracting horizontal active pixels, horizontal sync width and horizontal front porch, and seeing what remains.

So total horizontal is 2304

2304 - (1680+128+184) = 312.

And for vertical:

back porch = total - (active+sync width + front).

In my settings, that would be 1112 - (1050+1+3) = 58.

In your CRU settings, that would be 1112- (1050+3+6) = 53.

You said it calculated 56 vertical lines of back porch. Are you sure the total pixels was the same as mine (1112)?
Click to expand...

ah.. ur right.. the CRU total pixel is 1115
 
One of these days I'm gonna have to pull mine out of storage... It still works, but it has the "pinging " issue. (*ping* sounds with a faint distortion line appearing on screen for a split second).

The last time I sent it to the repair shop they obviously failed to replicate this issue, so this time I won't take it for repairs until something fails big time. Can probably get a few good months out of it before that happens, as it is now.
 
spacediver said:
wait, so you were able to find the part of the dat file that referenced the number of hours? Can you share that?!
Click to expand...

I just talked to Unkle Vito; he told me there was no way to determine the number of hours a monitor has been running. Someone was yanking my chain on eBay, I apologize for the inconvience. I do know from personal experience it has been running for 10 months since the time I got it for 8-10 hours per day therefore it should have plenty of life in it.
 
Kermie said:
I just talked to Unkle Vito; he told me there was no way to determine the number of hours a monitor has been running. Someone was yanking my chain on eBay, I apologize for the inconvience. I do know from personal experience it has been running for 10 months since the time I got it for 8-10 hours per day therefore it should have plenty of life in it.
Click to expand...

Let me clarify what I stated to Taylor...

There is no EEPROM parameter(s) that tracks and/or counts the hours the CRT has been used. Simply, this does not exist.

However, there is a way to estimate the life of the tube and the hour-range that the tube has been used by measuring its emission and by the level of luminance adjustments in WinDAS/WinCATs on the first pass and second pass. I am not going to go into the specifics (I told the readers that I will no longer be posting on this forum, but since Taylor mentioned my name, I have to further clarify what I stated to him), but this procedure is been very effectively used in estimating the usage and the life left on the tubes of these Sony monitors.

Hope this helps...

Sincerely,

Unkle Vito!
 
LAGRUNAUER said:
Hope this helps...

Sincerely,

Unkle Vito!
Click to expand...

Hey Unkle Vito, Welcome back! I wanted to invite you to be a mod on the subreddit dedicated to the FW900. It isn't very populated right now but if / when it does start taking help requests and showoff posts it will be a WHOLE LOT more organized than this 6,000 page post.

http://www.reddit.com/r/fw900 Reddit Accounts are real easy to make so just message me on there and I will set you up if you would like.
 
glad to see you post again lagrunauer.. thx for clearing that issue out, i was curious about how much hours my fw900 had aswell.. tbh, i'm glad its a hoax.
 
had an appalling scratch on my FW900, thanks to the AG posts I now have a shiny new screen with improved clarity!

thx to all those contributors.
 
So this isn't exactly related to the FW900 but I hope you guys can help me out. I'm thinking about upgrading to a 290 or 290X once prices come down a little bit but it looks like the reference ones don't come with a RAMDAC so no VGA. Can AIB cards add a RAMDAC?

Am I stuck with going with an active DP-VGA adapter? I'm running at 1920x1200@94Hz so I'm guessing I need a dual-link one to get the full refresh rate. Anyone have any recommendations on an adapter?

Edit: do active DP-VGA adapter cause input lag?
 
ManyCowsMoo said:
So this isn't exactly related to the FW900 but I hope you guys can help me out. I'm thinking about upgrading to a 290 or 290X once prices come down a little bit but it looks like the reference ones don't come with a RAMDAC so no VGA. Can AIB cards add a RAMDAC?

Am I stuck with going with an active DP-VGA adapter? I'm running at 1920x1200@94Hz so I'm guessing I need a dual-link one to get the full refresh rate. Anyone have any recommendations on an adapter?

Edit: do active DP-VGA adapter cause input lag?
Click to expand...

i asked amd on twitter and they said no, digital only :/
and i am 99% sure that there are no high enough clocked active adapters; you have to go nvidia in the future i think
and yes, they cause lag, usually quiet a bit because they are designed for television use where input lag is not important

https://twitter.com/McD3rp/status/398616365105704960
 
That sucks about the R9. I hope Nvidia doesn't follow suit, but I wouldn't be surprised if they did. Looks like the GTX-780ti may be the fastest analog video card. :(

Also, in other news - I'm a hypocrite. :D Remember that Quantum Data SG I said not to go for because the seller didn't respond? Well, eventually he responded and his answer checked out. I went ahead and took the gamble and bought the generator. So far so good! It works just like it should. I got the Quantum Data 801GG signal generator, which is the higher-resolution version of the 801 series. Maximum dot clock is 360Mhz, which should be fine for the high-res GDM displays.

For those of you who are in the know - a signal generator is needed to perform spot-on geometry and convergence adjustments. I know that WinDAS can natively control one of the other 801 generators (801GF), so I'm crossing my fingers that it can control the 801GG too (in essence, the 801GG is a high-res version of the 801GF). Either way, it won't matter - because I can still make a bunch of custom WinDAS timings for my monitors and still nail it. Native WinDAS control, however, would make it a lot easier. :)
 
Last edited:
very cool, you are a brave pioneer! :)

For my own purposes, I see absolutely no reason (yet) to go with a signal generator, but I would love to have one for calibrating units at the lab, as it makes the whole workflow much more portable and consistent.

Let us know how WinDAS interprets the modes - (i.e. does it give the correct mode number when you feed in the correct signal timings). I mean, if you get WinDAS to interface with the generator automatically, you won't have to do this, but would be interesting to see if WinDAS still interprets the mode as a different number if you do it manually.
 
jbltecnicspro said:
That sucks about the R9. I hope Nvidia doesn't follow suit, but I wouldn't be surprised if they did. Looks like the GTX-780ti may be the fastest analog video card. :(

Also, in other news - I'm a hypocrite. :D Remember that Quantum Data SG I said not to go for because the seller didn't respond? Well, eventually he responded and his answer checked out. I went ahead and took the gamble and bought the generator. So far so good! It works just like it should. I got the Quantum Data 801GG signal generator, which is the higher-resolution version of the 801 series. Maximum dot clock is 360Mhz, which should be fine for the high-res GDM displays.

For those of you who are in the know - a signal generator is needed to perform spot-on geometry and convergence adjustments. I know that WinDAS can natively control one of the other 801 generators (801GF), so I'm crossing my fingers that it can control the 801GG too (in essence, the 801GG is a high-res version of the 801GF). Either way, it won't matter - because I can still make a bunch of custom WinDAS timings for my monitors and still nail it. Native WinDAS control, however, would make it a lot easier. :)
Click to expand...

Yes...please Nvidia -- please continue to support your CRT and high end video card enthusiasts. We're still here...
 
You must log in or register to reply here.
Back
Top