1600SW/MLA with Octane MXE/V10?? -
mapesdhs - 05-25-2020
Does anyone have a suitable vfo file please to allow an Octane V10 to drive a 1600SW via MLA? Or with MXE?
None of the links I found on the irixnet archive work (or the attachment links are missing), nor those elsewhere such as on the gainos site, and of course the lnks from old neko threads on the Wayback machine don't work either.
I get signal not detected when using the supplied 1600x1024_60 in xsetmon.
Ian.
RE: 1600SW/MLA with Octane MXE? Or V10?? -
Jacques - 05-25-2020
Ian,
I have a bunch of vfo files, none of them have 1600x1024 though. I shall e-mail the whole lot to you.
I believe the 1600x1024_60 is for the F220, not 1600SW.
J
RE: 1600SW/MLA with Octane MXE? Or V10?? -
mapesdhs - 05-25-2020
(05-25-2020, 06:22 PM)JacquesT Wrote: I have a bunch of vfo files, none of them have 1600x1024 though. I shall e-mail the whole lot to you.
Thanks!! I'll put them on my site somewhere for all. Checking what refs I could find for 1600x1024 though, it looks like recondas was one of the main people who worked on this, also 3dchris, re this page:
http://archive.irixnet.org/apocrypha/nekonomicon/forum/3/16725672/1.html
Ian.
RE: 1600SW/MLA with Octane V10?? -
callahan - 05-31-2020
All -- I wanted to record for the record my complete failure to get this working in a MGRAS MXE for Ian. I compiled 1600x1024 vfo/sdb files for both Vpro and MGRAS (2RSS) boards using my Octane running 6.5.30. I wasn't able to test the Vpro files (according to Ian the correct resolution was displayed on his Dell LCD monitor, but the MLA would not recognize the display at all).
So, I turned my attention to getting a MXE Octane to display 1600x1024 @ 60. I thought this should be easy, given that I've had no trouble getting 1920x1080 and 1920x1200 working. However, each resolution I tried actually displayed 1280x1024 and, most oddly, appeared to have the leftmost ~320 pixels cut off from the screen.
My attempts revolved around the below .vfs built off of the modeline in Recondas' old nekochan post, copied below. I tried several variants with increased horizontal porches (up to 2x) and reduced framerate (down to 50 FPS), with no changes. I conducted all tests on my Dell U2412 monitor.
Modeline from nekochan:
Code:
Mode lines for the Silicon Graphics flat panel display:
These mode lines are required for use with the T2R4 (Rev 4) and the Silicon Graphics Flat Panel display.
Modeline "1600x1024d32" 103.125 1600 1600 1656 1664 1024 1024 1029 1030 HSkew 7 +Hsync +Vsync
Modeline "1600x1024d16" 103.125 1600 1600 1656 1664 1024 1024 1029 1030 HSkew 5 +Hsync +Vsync
Modeline "1600x1024d08" 103.125 1600 1600 1656 1664 1024 1024 1029 1030 HSkew 1 +Hsync +Vsync
Modeline "800x512d32" 54.375 800 800 840 848 512 512 514 515 HSkew 7 DoubleScan +Hsync +Vsync
Modeline "800x512d16" 54.375 800 800 840 848 512 512 514 515 HSkew 5 DoubleScan +Hsync +Vsync
Modeline "800x512d08" 54.375 800 800 840 848 512 512 514 515 HSkew 1 DoubleScan +Hsync +Vsync
These lines are required for use with the SGI Multilink Adapter and the SiliconGraphics Flat Panel display.
Modeline "1600x1024g" 108.0 1600 1616 1656 1704 1024 1027 1030 1056 -Hsync -Vsync
Option "OverridePolarity" "1"
My core .vfs file:
Code:
/*
1600x1024 Format for a SGI 1600SW monitor using a MLA adapter, v3
*/
General {
FormatName = "1600x1024_60_MLA1";
FieldsPerFrame = 1;
FramesPerSecond = 60.00;
TotalLinesPerFrame = 1056;
TotalPixelsPerLine = 1704;
ActiveLinesPerFrame = 1024;
ActivePixelsPerLine = 1600;
}
Active Line
{
HorizontalBackPorch = 0.445 usec;
HorizontalSync = 0.370 usec;
HorizontalFrontPorch = 0.148 usec;
}
Field
{
Vertical Sync =
{
{
Length = 1.0H;
Low = 0.0 usec;
}
repeat 2
{
Length = 1.0H;
}
}
Vertical Back Porch =
{
{
Length = 1.0H;
High = HorizontalSync;
}
repeat 25
{
Length = 1.0H;
Low = 0.0 usec;
High = HorizontalSync;
}
}
Active =
{
repeat 1024
{
Length = 1.0H;
Low = 0.0 usec;
High = HorizontalSync;
}
}
Vertical Front Porch =
{
repeat 3
{
Length = 1.0H;
Low = 0.0 usec;
High = HorizontalSync;
}
}
}
#if 0
postprocess
{
dump edge;
}
#endif
Which compiles fine and gives the following ascii output from vfc (using the -a=ascii option), clearly showing a 1600x1024 viewable area.
Code:
1600x1024_60_MLA1_v3_2.sdb:
Total lines per frame: 1056
Total pixels per line: 1704
Active lines per frame: 1024
Active pixels per line: 1600
Frames per second: 60
Fields per frame: 1
Swaps per frame: 1
Pixel clock: 107.965 MHz, period = 9.26222 nsec
Hardware pixel rounding: every 2 pixels
Line analysis:
Length: 1704 Pixels, 1 Lines, 15.7828 usec; (line 0)
Frequency: 63.36 KHz, period = 15.7828 usec
Horizontal Sync: 40 Pixels, 370.489 nsec; (line 29)
Horizontal Back Porch: 48 Pixels, 444.587 nsec; (line 29)
Horizontal Active: 1600 Pixels, 14.8196 usec; (line 29)
Horizontal Front Porch: 16 Pixels, 148.196 nsec; (line 29)
Field Information:
Field Duration: 1.79942e+06 Pixels, 1056 Lines, 16.6667 msec; (line 0)
Vertical Sync: 5112 Pixels, 3 Lines, 47.3485 usec; (line 0)
Vertical Sync Pulse: 5152 Pixels, 3.02347 Lines, 47.719 usec; (line 0)
Vertical Back Porch: 44304 Pixels, 26 Lines, 410.354 usec; (line 3)
Vertical Active: 1.7449e+06 Pixels, 1024 Lines, 16.1616 msec; (line 29)
Vertical Front Porch: 5112 Pixels, 3 Lines, 47.3485 usec; (line 1053)
The display manager also shows this as a 1600x1024 display, but the monitor only outputs 1280x1024.
On a whim (not really, also based on the post on the other forums
here), I also picked up one of the
famous dip-switch cables from ebay (I had always used a Sun 13W3 cable with all non-ground pins removed) and decided to give it another try. No difference in results, but I was very surprised to see that my Dell U2412 worked just fine with all switches on, the SOG setting, and H/V setting. The picture did appear slight fuzzy with the H/V sync settings compared to either all switches on or the SOG setting. Hmmm...
Edit: It gets weirder! Just after posting I decided to try my BenQ XL2730Z from my main PC out on the Octane. Primarily because it should be able to handle the 1600x1024@72 resolution that was pre-installed with Irix. The results were...very odd.
- First, the monitor handled the SOG signals without problem. Which is nice.
- 1280x1024 displayed well.
- 1600x1024 displayed poorly (at both 60 and 72 FPS). The monitor said it was displaying 1280x1024, but it clearly took up more lateral width than the 1280x1024 resolutions (about 20%, as expected) . For the 60 FPS variant (which also has reduced blanking), it appeared that the entire screen displayed, but there were substantial visual artifacts and tearing around the left side. Definitely not "working". I suspect this is a problem with the MGRAS rendering.
- 1920x1200, which displays perfectly on my Dell U2412, showed up as 1600x1200 and was compressed to match (i.e. it only took up ~65% of the lateral screen real estate in the 1:1 display mode, as expected for a 2560x1440 monitor). But otherwise, the entire screen was visible.
Moral of the story here is that when you find a monitor/resolution that you're happy with, probably best to stick with it! There is a lot of black magic going on that I don't fully understand.
RE: 1600SW/MLA with Octane V10?? -
mapesdhs - 05-31-2020
Thanks for the update!! I should add for the record that by using a different 13W3/VGA adapter I was at least able to get 1280x1024 working just fine with the MXE/MLA setup, so that's what the end user will be sticking with for now.
The other thing I tried was a combination of a cheap VGA/HDMI converter and HDMI/DVI cable, namely these items from Amazon (with of course the MLA set to digital mode):
https://www.amazon.co.uk/gp/product/B00S646VV4/
https://www.amazon.co.uk/gp/product/B00YOSA85Q/
Much to my surprise this actually worked for 1280x1024, though the image was a bit too washed out (I didn't expect the converter to cope with a SOG signal at all). Even more surprising though was that it did almost work for 1600x1024 aswell, in that the picture was the correct width & height (OSD showed it was definitely detecting 1600x1024), but alas the brightness/contrast was all wrong. Note I've not yet tried the equivalent tests with Octane/VPro.
Thus, I think a higher quality conversion from SOG analogue to digital would likely work ok, it would just be a lot more expensive, certainly too much for me to test.
Ian.