GBS 8200/8220 CFW Project
Re: GBS 8200/8220 CFW Project
Hey Rama I just remembered I actually run ttl hv sync into my gbs via one of the 2 vga ports, have cut the hv traces and and running 480ish ohm resistors inline no worrys.
You told us it wasnt designed to take ttl from the start so I did that.
Forgot my own solution lol. Brain fart
Dont know if it hurts passthru as everything gets routed back thru my Matrix which will reclock and boosts sync.
You told us it wasnt designed to take ttl from the start so I did that.
Forgot my own solution lol. Brain fart
Dont know if it hurts passthru as everything gets routed back thru my Matrix which will reclock and boosts sync.
-
NoAffinity
- Posts: 1028
- Joined: Mon May 07, 2018 5:27 pm
- Location: Escondido, CA, USA
Re: GBS 8200/8220 CFW Project
Got the parts in for the clock gen mod. Works great! The screen tearing is definitely gone.
:edit: Here's something interesting. The GBS now syncs much more likely (immediately?) at sync presence, allowing for video capture much earlier than GBS w/ CFW with no external clock gen, and also more quickly than OSSC. Some interesting startup video from SF2: World Warrior below, if you step through the video frame by frame. Not mind blowing by any means, but mildly cool.
:edit: Here's something interesting. The GBS now syncs much more likely (immediately?) at sync presence, allowing for video capture much earlier than GBS w/ CFW with no external clock gen, and also more quickly than OSSC. Some interesting startup video from SF2: World Warrior below, if you step through the video frame by frame. Not mind blowing by any means, but mildly cool.
Spoiler
Re: GBS 8200/8220 CFW Project
maxtherabbit:
Practically, the resistor seems only important for CVBS / Luma sync.
All other sources may overload the GBS in the worst case, but so far these things seem to be super robust in that regard.
It's only with CVBS / Luma that the termination is important. It prevents sync dropouts from sudden brightness changes,
and even then, only a few consoles suffer from that.
So one could simply install the resistor if they know they'll be using CVBS / Luma most of the time, and otherwise leave it off.
Granted, I've worked with these boards for so long, I recognize a sync problem cause by look alone by now :p
tiochico27:
Great to hear that it works at least for some people!
It would be bad if every niche console had their own special requirement :p
NoAffinity:
Yeah, I can skip some no-sync-out time with the external clock generator.
There's likely even more to be had, as this is the first working code for it
I need to get a capture card to fine tune the clock adjust algorithm.
Note that so far I have no method to check whether the tear line is actually off screen when it locks.
It seems to always land off screen anyway, so for now, I'm good with that.
But if it happens, you'll notice the tear bar just hanging around in the same spot for ages ;p
Syntax:
470 Ohms inline is good for most things.
Practically, the resistor seems only important for CVBS / Luma sync.
All other sources may overload the GBS in the worst case, but so far these things seem to be super robust in that regard.
It's only with CVBS / Luma that the termination is important. It prevents sync dropouts from sudden brightness changes,
and even then, only a few consoles suffer from that.
So one could simply install the resistor if they know they'll be using CVBS / Luma most of the time, and otherwise leave it off.
Granted, I've worked with these boards for so long, I recognize a sync problem cause by look alone by now :p
tiochico27:
Great to hear that it works at least for some people!
It would be bad if every niche console had their own special requirement :p
NoAffinity:
Yeah, I can skip some no-sync-out time with the external clock generator.
There's likely even more to be had, as this is the first working code for it
I need to get a capture card to fine tune the clock adjust algorithm.
Note that so far I have no method to check whether the tear line is actually off screen when it locks.
It seems to always land off screen anyway, so for now, I'm good with that.
But if it happens, you'll notice the tear bar just hanging around in the same spot for ages ;p
Syntax:
470 Ohms inline is good for most things.
-
tiochico27
- Posts: 18
- Joined: Fri Dec 27, 2019 3:03 am
Re: GBS 8200/8220 CFW Project
rama, i've a question about the PS2 VGA mode over component cables. How does it work?
-
- Posts: 40
- Joined: Wed Nov 28, 2018 4:27 pm
Re: GBS 8200/8220 CFW Project
I have rechecked the wiring of my Dreamcast and it is a bit chaotic so I will connect everything from scratch as soon as I have a little time. Regarding VGA I am looking for ideas that can help avoid the alteration of the Synchronization signal. I was thinking that since it injected RGBS through a Scart adapter to DB9 and from there to the GBS board it would be possible to place the resistance in the DB9 port between the Sync pin and ground with the idea that consoles with RGBS by this means get the required attenuation in Sync and if something is connected directly to the VGA port it is not altered by the resistor. Would this be possible? It's the best idea I have to avoid removing or putting a resistor due to the only console that uses VGA.
-
maxtherabbit
- Posts: 1763
- Joined: Mon Mar 05, 2018 4:03 pm
Re: GBS 8200/8220 CFW Project
I know the 5725 is only rated for 3.3V inputs from the datasheet you sent me but I'd agree from personal experience it seems to take TTL sync no problem. It may be doing some long term damage but at $20 for a board, eh.rama wrote:maxtherabbit:
Practically, the resistor seems only important for CVBS / Luma sync.
All other sources may overload the GBS in the worst case, but so far these things seem to be super robust in that regard.
What's the factory sync termination resistor value on the GBS again?
-
maxtherabbit
- Posts: 1763
- Joined: Mon Mar 05, 2018 4:03 pm
Re: GBS 8200/8220 CFW Project
Spoiler
The footprints R4 and R5 are there for series resistance on the sync outputs. Right now I just have them bridged, but I'd be amenable to putting something there to make the 5725 happier.
Re: GBS 8200/8220 CFW Project
Most boards: 520 Ohmmaxtherabbit wrote: What's the factory sync termination resistor value on the GBS again?
HDMI version: 1.8 kOhm
some boards: 1 kOhm
So yeah :p
With a standard VGA 5V level (the maximum one would expect from any source), I see about 4Vpp on the input with default termination.maxtherabbit wrote: I know the 5725 is only rated for 3.3V inputs from the datasheet you sent me but I'd agree from personal experience it seems to take TTL sync no problem. It may be doing some long term damage but at $20 for a board, eh.
The signal does not look distorted at all on the scope, so I'm tempted to guess that the 5725 may even be 5V tolerant.
That, or the AC coupling before input helps it dissipate the extra energy better.
-
maxtherabbit
- Posts: 1763
- Joined: Mon Mar 05, 2018 4:03 pm
Re: GBS 8200/8220 CFW Project
I'll hook mine up and probe it. A 330ohm series resistor on my board might not be a bad idea in any case - it should be low enough of a value to keep well above the logic high cutoff of anything with all ranges of high-Z sync termination.rama wrote:Most boards: 520 Ohmmaxtherabbit wrote: What's the factory sync termination resistor value on the GBS again?
HDMI version: 1.8 kOhm
some boards: 1 kOhm
So yeah :p
With a standard VGA 5V level (the maximum one would expect from any source), I see about 4Vpp on the input with default termination.maxtherabbit wrote: I know the 5725 is only rated for 3.3V inputs from the datasheet you sent me but I'd agree from personal experience it seems to take TTL sync no problem. It may be doing some long term damage but at $20 for a board, eh.
The signal does not look distorted at all on the scope, so I'm tempted to guess that the 5725 may even be 5V tolerant.
That, or the AC coupling before input helps it dissipate the extra energy better.
-
- Posts: 13
- Joined: Fri Jan 10, 2020 12:14 pm
Re: GBS 8200/8220 CFW Project
Just updated the gbs control. VGA presets are now saving and loading on my dreamcast. Thanks Rama.
-
maxtherabbit
- Posts: 1763
- Joined: Mon Mar 05, 2018 4:03 pm
Re: GBS 8200/8220 CFW Project
I get about 3.14 Vpp sync amplitude at my unmodified GBS when I put the 330ohm inline on my VGA dongle, I think I'm going to keep it that waymaxtherabbit wrote:I'll hook mine up and probe it. A 330ohm series resistor on my board might not be a bad idea in any case - it should be low enough of a value to keep well above the logic high cutoff of anything with all ranges of high-Z sync termination.rama wrote:Most boards: 520 Ohmmaxtherabbit wrote: What's the factory sync termination resistor value on the GBS again?
HDMI version: 1.8 kOhm
some boards: 1 kOhm
So yeah :p
With a standard VGA 5V level (the maximum one would expect from any source), I see about 4Vpp on the input with default termination.maxtherabbit wrote: I know the 5725 is only rated for 3.3V inputs from the datasheet you sent me but I'd agree from personal experience it seems to take TTL sync no problem. It may be doing some long term damage but at $20 for a board, eh.
The signal does not look distorted at all on the scope, so I'm tempted to guess that the 5725 may even be 5V tolerant.
That, or the AC coupling before input helps it dissipate the extra energy better.
Re: GBS 8200/8220 CFW Project
cryptblood1986: Great, enjoy!
maxtherabbit:
Yep, that sounds ideal.
Your dongles are pretty neat!
Would be great if we could build them in small volume ;p
The Mega Drive 2 solution is perfect for beefing up that weak CSync.
Having the resistors and caps close to the console helps the old CXA1145/CXA1645 deliver a good picture.
Aftermarket RGB cables should all be built like that, no matter what connector they end in (SCART, DSUB15, whatever).
For PSX, there could be a sync stripper to extract + buffer (it also buffers) sync from Luma, or just pass-through Luma.
PSX is particularly hard to mod a DSUB15 onto the console case. I just tried, and it's kind of janky and barely fits :/
One of these dongles would be much nicer
maxtherabbit:
Yep, that sounds ideal.
Your dongles are pretty neat!
Would be great if we could build them in small volume ;p
The Mega Drive 2 solution is perfect for beefing up that weak CSync.
Having the resistors and caps close to the console helps the old CXA1145/CXA1645 deliver a good picture.
Aftermarket RGB cables should all be built like that, no matter what connector they end in (SCART, DSUB15, whatever).
For PSX, there could be a sync stripper to extract + buffer (it also buffers) sync from Luma, or just pass-through Luma.
PSX is particularly hard to mod a DSUB15 onto the console case. I just tried, and it's kind of janky and barely fits :/
One of these dongles would be much nicer
Last edited by rama on Thu Jan 30, 2020 9:43 pm, edited 1 time in total.
-
maxtherabbit
- Posts: 1763
- Joined: Mon Mar 05, 2018 4:03 pm
Re: GBS 8200/8220 CFW Project
That's the idearama wrote:cryptblood1986: Great, enjoy!
maxtherabbit:
Yep, that sounds ideal.
Your dongles are pretty neat!
Would be great if we could build them in small volume ;p
Re: GBS 8200/8220 CFW Project
Gotta try harder Rama :p
Seriously tho it's a pain, but I needed csync for my Extron and hate removing any other video modes. Not enough free pins on that av port
Spoiler
Seriously tho it's a pain, but I needed csync for my Extron and hate removing any other video modes. Not enough free pins on that av port
-
maxtherabbit
- Posts: 1763
- Joined: Mon Mar 05, 2018 4:03 pm
Re: GBS 8200/8220 CFW Project
this pleases the Victorious God of AnalogSyntax wrote:Gotta try harder Rama :p
Spoiler
Seriously tho it's a pain, but I needed csync for my Extron and hate removing any other video modes. Not enough free pins on that av port
Re: GBS 8200/8220 CFW Project
Gotta work hard if I want to get into gamer heaven
I'll open it up soon and post a pic of mounting and location
I'll open it up soon and post a pic of mounting and location
Re: GBS 8200/8220 CFW Project
Best I was willing to do for this mod :p
Separate H+V taken from the GPU directly and maybe some day I'll even add a buffer!
The shield / metal shroud is in the way in all good locations :/
Separate H+V taken from the GPU directly and maybe some day I'll even add a buffer!
Spoiler
-
maxtherabbit
- Posts: 1763
- Joined: Mon Mar 05, 2018 4:03 pm
Re: GBS 8200/8220 CFW Project
I'll look into a PSX design, but it would require a different PCB with a footprint for a sync stripper. Could be useful for PAL nintendos and saturns too though.rama wrote: For PSX, there could be a sync stripper to extract + buffer (it also buffers) sync from Luma, or just pass-through Luma.
PSX is particularly hard to mod a DSUB15 onto the console case. I just tried, and it's kind of janky and barely fits :/
One of these dongles would be much nicer
Re: GBS 8200/8220 CFW Project
I buffered my csync but saw little point in a RGBHV 15k signal.
What would you use rgvhv 15k for??
What would you use rgvhv 15k for??
-
maxtherabbit
- Posts: 1763
- Joined: Mon Mar 05, 2018 4:03 pm
Re: GBS 8200/8220 CFW Project
pretty much only dreamcastSyntax wrote:I buffered my csync but saw little point in a RGBHV 15k signal.
What would you use rgvhv 15k for??
Re: GBS 8200/8220 CFW Project
Is it possible to have more than 5 custom presets (with the possibility to rename them) ? it gets kinda confusing with more than 10 consoles.
Re: GBS 8200/8220 CFW Project
Yeah but Rama did rgbhv 15k for a ps1...
Re: GBS 8200/8220 CFW Project
Iraito:
Renaming support is quite a bit of extra work, and very specific to the preset system.
It'd be great to have some help from people more versed in HTML and JS for these things.
It's not something I want to work on at this time.
Syntax:
The RGBHV PSX is just a development aid that provides a convenient legacy PAL/NTSC timing source.
It's entirely pointless otherwise, especially since separate HV Sync doesn't have a good code path in gbscontrol.
(The issue here is that separate sync doesn't seem to register in the 5725 Input Formatter unit correctly. It may be a chip bug or oversight.)
Renaming support is quite a bit of extra work, and very specific to the preset system.
It'd be great to have some help from people more versed in HTML and JS for these things.
It's not something I want to work on at this time.
Syntax:
The RGBHV PSX is just a development aid that provides a convenient legacy PAL/NTSC timing source.
It's entirely pointless otherwise, especially since separate HV Sync doesn't have a good code path in gbscontrol.
(The issue here is that separate sync doesn't seem to register in the 5725 Input Formatter unit correctly. It may be a chip bug or oversight.)
-
maxtherabbit
- Posts: 1763
- Joined: Mon Mar 05, 2018 4:03 pm
Re: GBS 8200/8220 CFW Project
I mean if the separate syncs are available might as well grab em. I would have buffered thoughSyntax wrote:Yeah but Rama did rgbhv 15k for a ps1...
Re: GBS 8200/8220 CFW Project
Rama have you also modded the ps1 to have both crystal's like mine?
-
- Posts: 40
- Joined: Wed Nov 28, 2018 4:27 pm
Re: GBS 8200/8220 CFW Project
Shiver_169 wrote:I have rechecked the wiring of my Dreamcast and it is a bit chaotic so I will connect everything from scratch as soon as I have a little time. Regarding VGA I am looking for ideas that can help avoid the alteration of the Synchronization signal. I was thinking that since it injected RGBS through a Scart adapter to DB9 and from there to the GBS board it would be possible to place the resistance in the DB9 port between the Sync pin and ground with the idea that consoles with RGBS by this means get the required attenuation in Sync and if something is connected directly to the VGA port it is not altered by the resistor. Would this be possible? It's the best idea I have to avoid removing or putting a resistor due to the only console that uses VGA.
Any response for this?
-
maxtherabbit
- Posts: 1763
- Joined: Mon Mar 05, 2018 4:03 pm
Re: GBS 8200/8220 CFW Project
I for one do not understand what you are saying. DB-9? Are we doing TTL RGB here?Shiver_169 wrote:Shiver_169 wrote:I have rechecked the wiring of my Dreamcast and it is a bit chaotic so I will connect everything from scratch as soon as I have a little time. Regarding VGA I am looking for ideas that can help avoid the alteration of the Synchronization signal. I was thinking that since it injected RGBS through a Scart adapter to DB9 and from there to the GBS board it would be possible to place the resistance in the DB9 port between the Sync pin and ground with the idea that consoles with RGBS by this means get the required attenuation in Sync and if something is connected directly to the VGA port it is not altered by the resistor. Would this be possible? It's the best idea I have to avoid removing or putting a resistor due to the only console that uses VGA.
Any response for this?
Re: GBS 8200/8220 CFW Project
Yeah the wall of text lost me too.
Re: GBS 8200/8220 CFW Project
Syntax:
Not that console, but I often experiment with that mod.
I usually salvage a 59.69MHz crystal from an NTSC-J and add it to a PAL console, leaving the stock 53.2MHz as well.
Just remember that the GPU clock in pins description ("N" and "P) is swapped in the service manual ;p
The Si5351 can easily do dual oscillator as well. Once my extra test boards arrive, I think I'll try and add it to PsNee
Shiver_169:
Regarding the DC, someone should go and find out whether CSync is available in VGA mode.
It makes VGA support in gbscontrol so much easier to have CSync, due to the Input Formatter bug.
Not that console, but I often experiment with that mod.
I usually salvage a 59.69MHz crystal from an NTSC-J and add it to a PAL console, leaving the stock 53.2MHz as well.
Just remember that the GPU clock in pins description ("N" and "P) is swapped in the service manual ;p
The Si5351 can easily do dual oscillator as well. Once my extra test boards arrive, I think I'll try and add it to PsNee
Shiver_169:
Regarding the DC, someone should go and find out whether CSync is available in VGA mode.
It makes VGA support in gbscontrol so much easier to have CSync, due to the Input Formatter bug.
Re: GBS 8200/8220 CFW Project
Renaming isn't that important really, but more custom presets would be amazing.