Forums

I made a little program for the upcoming competition and I encountered a strange bug I don’t really know what to make of.

When I test it on real hardware via Master Everdrive X7, the TV type detection seems to fail. The program runs in 50 Hz mode even though the system runs at 60 Hz, so everything is going too fast. On the other hand, the same program works absolutely fine in Emulicious and also on the Mega SG hardware.

I can’t rule out that it’s my Master System, as that was originally a PAL machine that’s now permanently stuck in 60 Hz mode after a botched attempt at installing a switch. However, I used the exact same detection routine in Flight of Pigarus and that have always the correct results.

Any thoughts or ideas?

Hi,

I also have an everdrive X7 with a modded Master System 2 PALto switch from 50 to 60hz.

Did you use the Master System region detection to perform your test ?[/url]

If you have a code example, I might be able to take a look and test it on my own.

KingQuack wrote

Did you use the Master System region detection to perform your test ?

No, in this case I use this code. I’ve also used the exact same code in the past, both in Flight of Pigarus and in my WIP Mermaid game, and it has never failed me in either. Hence my question if maybe the Everdrive writes something to System RAM. The byte that stores the TV type information is located at $C000, right at the beginning of RAM. In the other two games, it’s positioned somewhere else.

I wouldn't expect it to disrupt your RAM, but... did you try moving that byte in some other place?

I perform a test myself, with the everdrive X7 on a SMS 2, and I can confirm that the TVTypeDection code is working well.

I can switch from 50 to 60hz and the code detects the change.

You can also try to use sverx's SMS Test Suite ROM too.

KingQuack wrote

You can also try to use sverx's SMS Test Suite ROM too.

but the Test Suite uses a completely different approach to detecting if it's 50 or 60 Hz, not 'cycle count' based but detecting the vcount 'roll-back' count.

sverx wrote

but the Test Suite uses a completely different approach to detecting if it's 50 or 60 Hz, not 'cycle count' based but detecting the vcount 'roll-back' count.

ah right, I admit I didn't look at how the test was carried out.

For my part, here's what I used.
I just draw a 0 or a 1 on the screen depending on the result.

Problem solved, nothing to see here. It was human failure after all. Console, Everdrive and detection code are all working as they should. Turns out it was just me being stupid.

Apparently, I accidentally commented out a di, which in turn messed up the detection. It’s actually weird it even worked in Emulicious and on the Mega SG in the first place.

Not saying that is the case, but....

The last time I was scratching my head why some ROM was working on emulicious and was not working on the everdrive, the root cause was....that I didn't copy it to the sd and I was testing the wrong one!

Kagesan wrote

Apparently, I accidentally commented out a di, which in turn messed up the detection. It’s actually weird it even worked in Emulicious and on the Mega SG in the first place.

Interesting! I suspect the difference could be that the Master EverDrive leaves the interrupts enabled before starting the ROM while the others might start that with interrupts disabled.
In either case, never assume any previous status, always set some initial setting yourself.

Author	Message
Kagesan Joined: 01 Feb 2014 Posts: 878	Can the Everdrive mess up TV type detection? Posted: Tue Feb 06, 2024 5:57 pm
Kagesan Joined: 01 Feb 2014 Posts: 878	I made a little program for the upcoming competition and I encountered a strange bug I don’t really know what to make of. When I test it on real hardware via Master Everdrive X7, the TV type detection seems to fail. The program runs in 50 Hz mode even though the system runs at 60 Hz, so everything is going too fast. On the other hand, the same program works absolutely fine in Emulicious and also on the Mega SG hardware. I can’t rule out that it’s my Master System, as that was originally a PAL machine that’s now permanently stuck in 60 Hz mode after a botched attempt at installing a switch. However, I used the exact same detection routine in Flight of Pigarus and that have always the correct results. Any thoughts or ideas?

KingQuack Joined: 15 Jul 2022 Posts: 29	Posted: Tue Feb 06, 2024 6:05 pm
KingQuack Joined: 15 Jul 2022 Posts: 29	Hi, I also have an everdrive X7 with a modded Master System 2 PALto switch from 50 to 60hz. Did you use the Master System region detection to perform your test ?[/url] If you have a code example, I might be able to take a look and test it on my own.

Kagesan Joined: 01 Feb 2014 Posts: 878	Posted: Wed Feb 07, 2024 8:22 am
Kagesan Joined: 01 Feb 2014 Posts: 878	KingQuack wrote Did you use the Master System region detection to perform your test ? No, in this case I use this code. I’ve also used the exact same code in the past, both in Flight of Pigarus and in my WIP Mermaid game, and it has never failed me in either. Hence my question if maybe the Everdrive writes something to System RAM. The byte that stores the TV type information is located at $C000, right at the beginning of RAM. In the other two games, it’s positioned somewhere else.

sverx Joined: 05 Sep 2013 Posts: 3828 Location: Stockholm, Sweden	Posted: Wed Feb 07, 2024 8:57 am
	I wouldn't expect it to disrupt your RAM, but... did you try moving that byte in some other place?

KingQuack Joined: 15 Jul 2022 Posts: 29	Posted: Wed Feb 07, 2024 3:07 pm
KingQuack Joined: 15 Jul 2022 Posts: 29	I perform a test myself, with the everdrive X7 on a SMS 2, and I can confirm that the TVTypeDection code is working well. I can switch from 50 to 60hz and the code detects the change. You can also try to use sverx's SMS Test Suite ROM too.

sverx Joined: 05 Sep 2013 Posts: 3828 Location: Stockholm, Sweden	Posted: Wed Feb 07, 2024 4:08 pm
	KingQuack wrote You can also try to use sverx's SMS Test Suite ROM too. but the Test Suite uses a completely different approach to detecting if it's 50 or 60 Hz, not 'cycle count' based but detecting the vcount 'roll-back' count.

KingQuack Joined: 15 Jul 2022 Posts: 29	Posted: Wed Feb 07, 2024 6:22 pm
KingQuack Joined: 15 Jul 2022 Posts: 29	sverx wrote but the Test Suite uses a completely different approach to detecting if it's 50 or 60 Hz, not 'cycle count' based but detecting the vcount 'roll-back' count. ah right, I admit I didn't look at how the test was carried out. For my part, here's what I used. I just draw a 0 or a 1 on the screen depending on the result.

Kagesan Joined: 01 Feb 2014 Posts: 878	Posted: Thu Feb 08, 2024 6:27 am
Kagesan Joined: 01 Feb 2014 Posts: 878	Problem solved, nothing to see here. It was human failure after all. Console, Everdrive and detection code are all working as they should. Turns out it was just me being stupid. Apparently, I accidentally commented out a di, which in turn messed up the detection. It’s actually weird it even worked in Emulicious and on the Mega SG in the first place.

kusfo Joined: 29 Mar 2012 Posts: 886 Location: Spain	Posted: Thu Feb 08, 2024 6:55 am
kusfo Joined: 29 Mar 2012 Posts: 886 Location: Spain	Not saying that is the case, but.... The last time I was scratching my head why some ROM was working on emulicious and was not working on the everdrive, the root cause was....that I didn't copy it to the sd and I was testing the wrong one!

sverx Joined: 05 Sep 2013 Posts: 3828 Location: Stockholm, Sweden	Posted: Thu Feb 08, 2024 8:23 am
	Kagesan wrote Apparently, I accidentally commented out a di, which in turn messed up the detection. It’s actually weird it even worked in Emulicious and on the Mega SG in the first place. Interesting! I suspect the difference could be that the Master EverDrive leaves the interrupts enabled before starting the ROM while the others might start that with interrupts disabled. In either case, never assume any previous status, always set some initial setting yourself.

Forums

View topic - Can the Everdrive mess up TV type detection?