KR155E wrote:
A good suggestions, thanks! I have reduced the number of articles to 5 now. 🙂
Excellent job, as usual. Thanks for the speedy response!
Maybe I can build something for me that generates the needed img tag with js code and somehow gets a screenshots in the size of the video.
I’m sure you can pull it off 😉
For the screen-shots, since it doesn’t need to be done all that often, I would just do it manually with “Print Screen” and IrfanView. Its “auto-crop” function (Strg+Shift+Y) would make it easy.
BTW, if you add the “/clippaste” option to the command-line of the shortcut you use to start IrfanView, it starts up with the clipboard image already loaded. All of this assumes you’re running some flavor of Windows, of course…
thevirtualboy wrote:
Do you think its possible to just use a glue gun to reinforce the glue so I wouldn’t have to risk melting anything?
The glue is between the cable and the display board, so you can’t really reinforce it with more glue. Unless you warm the glue and press the cable into tighter contact, you won’t fix the problem, even temporarily.
Once you have it working, you can put packing tape over it and wrap it tightly around to try to hold it on better, but it might make it harder to permanently fix later.
If you decide to do it, take Dor-Si’s advice; he seems to know what he’s talking about.
Dreammary wrote:
I used my screwdriver and took apart my spare VB because it displays nothing. Then I got to thinking, and realized that hotglue guns dispense the glue around 200 degrees, so I squirted hot glue on each ribbon and now it displays a perfect image.
I’m glad it worked for you, but you have to realize that it’s no more permanent a fix than the oven method, yet it’s next to impossible to re-do it when (not “if”) it fails again, since adding more glue on top of what’s there won’t heat the original adhesive at all. It may be possible (though probably very messy and possibly painful) to do the oven method, but the glue will almost certainly hinder the heat getting to the joint, making that method even less effective than it already is. On top of that, you’ve made it very difficult to apply the “actual” (solder) repair method. Your only hope is that the glue doesn’t actually stick to the cable (which it tends not to do on such smooth surfaces), and that peeling it off doesn’t permanently damage the cable or board.
All that being said, you may be on to something. If a similar substance can be found to evenly conduct the right amount of heat to the joint without sticking to it, you may have found a “less dangerous” alternative to the oven method.
Other than possibly insufficient input capacitance, that circuit looks fine to my very inexpert eyes ;-). I assume the RC timer is there to simulate the “power good” signal as described in this post/thread:
Do you have the means to test the reset signal? It’s possible the mirrors will still start even if this signal isn’t working, but the CPU probably requires it.
I’m still hoping it’s running and just not showing any outward signs (like bad display cables combined with a faulty audio amp or volume control).
virtualnoob wrote:
made custom regulator (capacitor,9k resistor and regulator)
now the mirrors turn on
but soon after powering on the voltages and current slowly drops near zero. im losing my hope on this : s
Don’t give up! We’ll find out what’s wrong, eventually.
What are you powering the regulator with? Make sure it can supply enough current and that it meets your custom regulator’s dropout voltage requirement (without exceeding it’s maximum voltage, of course).
Hmm… It’s odd that the LED on the regulator is working when the regulator itself isn’t. Of course, that may be how it goes; I’ve never had one go out, personally.
But, if you have it powered up through the controller port, and it’s still not booting, I guess it’s time to start looking over the PCBs for damage. There could be a short, a broken solder joint, or a damaged trace. A regulator suddenly going out could be caused by a power rail being shorted to ground. Electrolytic capacitors are another possible culprit—they are getting up there in years, you know.
Have you checked all the internal cables and connectors? Maybe one is loose or making intermittent connection. Without being there to look at the “patient” first-hand, I can’t think of anything else to check.
Keep us apprised of what you find out. You may help save other VBs in the future!
I recently had issues with one not accepting the cart if it was pushed in too firmly. Try pushing the cart in with different amounts of force, or pushing it in and then pulling very slightly on it, before applying power.
If the voltage regulator lights and the mirrors vibrate, it has to be a cartridge or cart port issue (or dead display cables; see below).
BTW, what carts did you try? Some games take a few button pushes and some waiting before they make any sounds. If the displays are completely gone (as in: the cable no longer makes any contact), you may just not know it’s running.
WoLfMaN wrote:
Looks like i am experiencing the same now. Happened today when I installed my newly repaired screens. Shows up on the left eye, starting from the far right reaching to the middle (fading off after 2/5 of the screen lenghth).
I always kept it clean and treated it well. It cant be a scratch…what else could that be, except for dust?
I’ve noticed that it’s very hard to keep dust off of the mirrors. This can cause a darker, horizontal streak that doesn’t seem to span the whole width of the image. And you have to be careful cleaning it, because they smudge and scratch very easily, being “first surface” mirrors.
Also, you have to take a bit of care installing the display PCBs. I’ve had the foil on the non-soldered end delaminate and cause shorts. You can usually straighten them up and carefully reinsert them into the connector, though.
I don’t think double-posting is going to get this noticed any better…
If you really want this to happen, put up a bounty. At the very least, you could offer your VB as a “lab rat” to be experimented on.
If you have more brains than money, to coin a phrase, then just (re)read this thread and check the wiki, the VB patents, and anything else you can find on the subject, roll up your sleeves, heat up your iron, and get down to business! Hey, you might even learn something useful along the way 😉
KR155E wrote:
This thing looks awesome. I have donated a last minute $333 last night, now can’t wait for the Oculus to arrive in December – if they can adhere to their schedule now that they fundraised 10 times their initial goal. A friend of mine will be receiving his Leap Motion prototype dev kit in December as well, which ia a new motion tracking system. Maybe we’ll try to combine both systems.
Anyway, I hope Ryphecha got an Oculus Rift as well and will add a special display mode to Mednafen. 😉
Actually, with that kind of cash, they should be able to get these things on shelves in Best Buy within the deadline!
I’m not sure using that Leap thing without a screen would be that much fun… But, it’s worth a try!
As for Mednafen support: it looks like the OR just uses standard side-by-side mode. You might need to tweak the scale and positioning, but I think Med. already has options that should at least get you close.
wolfman wrote:
So back to field 1 – it can only be done within the VB?
What parts do we need and where will we have to put them?
Unfortunately I don´t have electronic skills, but I am quite good at soldering….if I had a spare unit and sort of a manual, I´d start immediately 😀
AFAIK, all you need is a new crystal with which to replace the stock one. Of course, this will result in the “squished screen” problem, but I don’t know any specifics about how to go about solving that problem. If you have an oscilloscope, you could check the wires leading from the mainboard to the servo controller and try to find the mirror drive signal. I’m thinking a PLL could be used to derive a clock of the old frequency from the one produced in the OC’d VB. The good news is that, I think the VIP (graphics chip) actually synchronizes itself to the mirrors, rather than the other way around. My theory is that the squishing is caused by the VB trying to move the mirrors faster than they are physically able. Like any pendulum, they have a specific period. There is also the inductance of the drive coils to deal with.
I have a VB to use, and the desire to experiment with stuff like this, but I have neither the time nor the special equipment needed 🙁
If you send me a few hundred dollars for an o-scope and parts, I’ll post a “how-to” video for you in a couple of weeks 😉
Okay, here’s what I could find with a quick search. It’s been edited for relevancy. All times are CST (GMT-6).
Sun, May 22, 2005
[02:58:45a] How far can you overclock one without trouble?
[02:59:00a] I wouldn’t really know.
[02:59:06a] DP?
[02:59:11a] The bottleneck is probably the RAM.
[02:59:11a] You tried this yet?
[02:59:34a] I had one running at 33MHz, still ran, but had some glitches, probably RAM or cart ROM speed problem
[02:59:50a] 40MHz wouldn’t boot, not sure if it was the CPU or the VPU though (or both :P)
[03:00:10a] Just replaced the Xtal?
[03:00:23a] I wanna try just overclocking the CPU, leaving the VPU at regular speed… but I haven’t gotten around to it yet
[03:00:25a] yeah
[03:00:56a] CPU overclock would be great, because most games are timed to the VPU.
[03:01:06a] exactly…
[03:01:35a] I’d probably also need to do something about the mirrors…
[03:01:46a] It might actually be timed properly, actually. Just not using interrupts.
[03:02:12a] ugh, I used the same word twice in one sentence.
[03:03:22a] I think I’d need to make the mirrors move faster, or maybe with the VPU slow, it’d still stay sync’d… I dunno
Mon, May 23, 2005
[09:09:06a] at 40MHz, the CPU wasn’t hot, but it wouldn’t boot, and at 33, it was fast, but glitchy (but not hot)
As far as parts wearing out faster, I don’t think you need to worry with such modest speed gains. It likely draws more current at a higher speed, but I can’t say if it exceeds the VB’s voltage regulator’s capability.
If you try it, be sure to post your experiences here. Like you said, it’s quite an interesting topic.
It’s been done. DogP did it sometime in 2005 (maybe earlier). It should be possible to make it fully functional, but during experimentation it was found that the mirror oscillations were driven by the main clock, too. This caused the screen to “squish” horizontally. This might be remedied by either building a separate clock circuit for the display, or deriving the original clock from the new clock with a divider.
I’ve got IRC logs of a discussion about it, if anyone is interested…
Do you have a problem with a misaligned display, or are you just concerned that they may become misaligned if you ever repair the display cables?
If it’s the former, you’ll need to describe the actual problem you’re having in more detail. It would help if you could take a few photos of the display hardware.
If it’s the latter, you don’t really need to worry. The only things that need to be removed to do the fix are the display PCBs. They’re held on with two screws and there is only one way they can go on. There are registration pins to prevent misalignment, and the rest of the optical system is unaffected by the removal/reinstallation, as long as you don’t take them out with a hammer or crowbar 😉
The galvanometers (mirrors) and the focus/IPD system is actually very robust. As long as no attempt is made to force it into a position it wasn’t designed for, it should need no manual realignment. Also, having made a fairly thorough study of the insides, I’m of the opinion that any drop or impact hard enough to affect the optical system would most likely damage other components, as well. Such a mistreated system would probably best be used for spare-parts.
I hope that helps. If not, try to clarify your question/problem.
It’s possible there are some lost gems there, but I doubt it…
In fact, there’s probably more info in the wiki now than on his site, although much of what is there was discovered by him during the development of his VB emulator, Reality Boy.
VBrulez wrote:
That looks awesome. I can’t wait to play Doom 3 on one of those. This may be a dumb question but is this going to be it’s own system, or will you need to plug it into a system to play it?
The impression I got is that it’s just an LCD monitor and motion tracker (most likely in the form of a USB or, eventually, Bluetooth HID). Software support will likely be limited to PCs at first, but the hardware is probably compatible with the big three game consoles (and a few older ones).
Yeti_dude wrote:
When I try to use gccVB through cygwin I encounter some funky issues. I’ve looked through different makefiles for the exact commands to issue to v810-gcc but I always get the same string of errors. Most of the errors I get are all related to macros in the video.h and world.h files. Here’s a snippet of some of the compiler errors.
../libgccvb/video.h:31: warning: type defaults to ‘in’ in declaration of ‘VIP_REGS’
../libgccvb/video.h:31: conflicting types for ‘VIP_REGS’
../libgccvb/vip.h:37: previous declaration of ‘VIP_REGS’
../libgccvb/video.h:31: ‘b’ undeclared here (not in a function)
../libgccvb/video.h:31: warning data definition has no type of storage class
../libgccvb/video.h:31: stray ‘/’ in program
…
tron.c:39: warning: implicit declaration of function ‘vbSetWorld’
…
These errors look like they might just be caused by a mismatch between the libgccvb version you have and the compiler you’re using. Hard to say for sure without seeing the code.
I use this line from the barebones makefile when trying to compile through cygwin and :
v810-gcc -Wall -nodefaultlibs -mv810 -xc -o $(TARGET).o $(TARGET).c
I’d like to know what all these options mean and why they are included. Any help would be greatly appreciated!
“v810-gcc –help” will explain most of these, and the manpage can handle the rest, but here’s a quick rundown:
“-Wall” means to enable all warnings.
“-nodefaultlibs” obviously prevents gcc from linking in the default libraries one would use on a desktop system.
“-m” specifies the “machine” type. In this case, it’s “v810”. I think it might be optional in the tools with the “v810-” prefix, but I’ve never tested it.
“-xc” just means to treat the input file as one containing “C” code, rather than trying to figure out what it contains from the file extension.
“-o ” specifies the output filename. Any other files specified are input files.
Far be it from me to question your methods—especially given that you’ve made working ROMs—but if you’re going to go to the trouble of using WINE to run a win32 compiler, why not at least use the VBDE, which includes a more modern GCC and the latest crt0, linker script, and libgccvb headers?
Have you tried compiling GCC 4.x with the VB patches applied? I think it’s been done under OSX, so it should work fine under Ubuntu. Here’s a relevant post that should get you on the right track, at least:
I’m almost certain I used that very tarball (the non-OSX 4.4.2 one, that is) to build the Win32 version of GCC included in the VBDE (under MinGW). The only thing that might be newer is the crt0 and/or linker script. The one I have was last updated by dasi, but I don’t know if it’s the same one included in the VBDE package.
I hope that helps you streamline your VB dev. environment.
