EZ-Flash Omega DE RTC not working randomly

[ghjfdtg]

So i got the ODE today brand new from a official reseller ("Dealers"). As far as i can tell everything works except RTC.

New EZ ODE with nearly full RTC battery (measured 3.01V).
GBA SP AGS-001 completely stock (no modifications).
Genuine (tested) 64 GB Samsung Evo microSD formatted as exFAT with SDFormatter.
Newest kernel and firmware from ezflash.cn downloaded yesterday.
All tested ROMs are verified against the no-intro database.

I set the RTC in the menu correctly and it's holding time and date as expected even on longer periods without power. No time jumping. "GAME RTC" is set to "OPEN" (blue square). The issue is if i boot any of the Pokemon games requiring RTC it sometimes displays the battery run dry message and sometimes not. It seems to be random. Same happens booted from NOR, mode B and it makes no difference if i clean boot or with addon. I will also try other RTC enabled games later.

I don't believe this is a hardware issue. It seems to me there is something wrong either with the software or the FPGA RTC implementation.

edit:
I just tried Boktai (also verified clean ROM) and sometimes the date and time resets to 1.3.2004 (DMY) 00:00 and sometimes it's the correct one i set in the menu.

 

[Shadow#1]

So i got the ODE today brand new from a official reseller ("Dealers"). As far as i can tell everything works except RTC.

New EZ ODE with nearly full RTC battery (measured 3.01V).
GBA SP AGS-001 completely stock (no modifications).
Genuine (tested) 64 GB Samsung Evo microSD formatted as exFAT with SDFormatter.
Newest kernel and firmware from ezflash.cn downloaded yesterday.
All tested ROMs are verified against the no-intro database.

I set the RTC in the menu correctly and it's holding time and date as expected even on longer periods without power. No time jumping. "GAME RTC" is set to "OPEN" (blue square). The issue is if i boot any of the Pokemon games requiring RTC it sometimes displays the battery run dry message and sometimes not. It seems to be random. Same happens booted from NOR, mode B and it makes no difference if i clean boot or with addon. I will also try other RTC enabled games later.

I don't believe this is a hardware issue. It seems to me there is something wrong either with the software or the FPGA RTC implementation.

edit:
I just tried Boktai (also verified clean ROM) and sometimes the date and time resets to 1.3.2004 (DMY) 00:00 and sometimes it's the correct one i set in the menu.

Replace the battery

 

[ghjfdtg]

It's very unlikely to be the battery. Most RTCs work down to around 2.7-2.8V. A fresh battery has around 3.3V so it's basically right in the middle. The RTC is working fine as said. Just not reliably in games which makes me believe this is a software or some kind of timing issue with the RTC implementation on the FPGA.

I have seen other threads where people had issues but it didn't work at all for them or they forgot to enable RTC. If my issue is related that makes me think there is a hardware dependent timing issue that could be solved with a new FPGA bitstream (or "firmware" how EZ-Flash calls it).

I should add that i obviously already tried removing the battery for a few minutes. Makes no difference. I have no replacement batteries in this size at home right now.

 

[EZ-Flash2]

So i got the ODE today brand new from a official reseller ("Dealers"). As far as i can tell everything works except RTC.

New EZ ODE with nearly full RTC battery (measured 3.01V).
GBA SP AGS-001 completely stock (no modifications).
Genuine (tested) 64 GB Samsung Evo microSD formatted as exFAT with SDFormatter.
Newest kernel and firmware from ezflash.cn downloaded yesterday.
All tested ROMs are verified against the no-intro database.

I set the RTC in the menu correctly and it's holding time and date as expected even on longer periods without power. No time jumping. "GAME RTC" is set to "OPEN" (blue square). The issue is if i boot any of the Pokemon games requiring RTC it sometimes displays the battery run dry message and sometimes not. It seems to be random. Same happens booted from NOR, mode B and it makes no difference if i clean boot or with addon. I will also try other RTC enabled games later.

I don't believe this is a hardware issue. It seems to me there is something wrong either with the software or the FPGA RTC implementation.

edit:
I just tried Boktai (also verified clean ROM) and sometimes the date and time resets to 1.3.2004 (DMY) 00:00 and sometimes it's the correct one i set in the menu.

Please wait we investigate it.

 

[ghjfdtg]

Yeah, turning it off games don't recognize it at all. Not even randomly. I also tried toggling this option a few times. It makes no difference.

As mentioned i think there is some timing issue. The tolerances of the clock and FPGA crystals may trigger this timing issue. It's only a guess since i don't know your GBA ROM and RTC implementation.

edit:
Additional infos:

I ran your recommended test setup letting the aging cartridge run for several hours. With the ROM in PSRAM it survived 5 hours and 26 minutes before it froze. The battery indicator was already red at that point but i think the flashcart should not stop working before the battery is fully drained.

edit2:
I just tested a little more and the RTC is alternating between the set time & date and an invalid time & date. To be clear, only in game. Not in the kernels menu.

edit3:
I tried the Real Time Clock Reader homebrew and can now exactly tell you what's going on. The month and week day increase above their limits and reset to normal exactly on second 0 of each minute. Possiblyi should catch this on camera. It's pretty weird. The time, day and year are read correctly.

 

[ghjfdtg]

@EZ-Flash2
I think this warrants a double post since i got a heap of new infos. I already added most of it to the previous post but looks like you didn't see it yet.

First of all i know exactly what's going on with the RTC. In game the month and day of the week values go out of bounds and increase even further until the minute counter reaches zero. Then it goes back to normal but will increase again. I made a video showing this behavior.

Also i tested the cart with the Aging Cartridge ROM and it seems to always freeze around the time the battery indicator goes red.

Something new i tried is making a 32 MB homebrew ROM that hashes itself with SHA1 to see if the PSRAM and NOR work correctly. I'm getting the correct hash loading the ROM to PSRAM (clean) but the hash is wrong writing it to NOR (clean) and direct booting it. No difference in mode B. Is the kernel doing any patches when writing ROMs to NOR?

So knowing all this what do you say? Should i get a replacement?

 

[EZ-Flash2]

When the battery turns red cause the card freezes, this is not a problem with the card itself, but the battery has been unable to output a strong enough current.

As for your concerns about the RTC, a replacement might be a good way to go. Please make a note with the card when you send it back to the dealer, we want to be able to clearly distinguish the card for diagnostic purposes when we receive the faulty cards from the dealer.

 

[ghjfdtg]

@EZ-Flash2
I contacted the reseller and i'm getting a replacement. I told them to forward the cart to you for diagnostic purposes and i included a note with the cart so it can be distinguished. Probably takes a while because of the holidays.

 

[ghjfdtg]

@EZ-Flash2

Update:
Got the replacement cart and everything works this time. The self hashing test ROM also gives the correct hash now when written to NOR (clean) and direct booted from there so the NOR flash or a solder ball of the NOR was also faulty. I don't know if this was also the cause of the RTC giving invalid values between minutes.

One problem that is not solved is the Aging cartridge test freezing roughly 15 minutes after the battery indicator goes red. Since the new cart has the same problem it might indeed be the battery but i repeated the test with the charger plugged in. It survived 6 hours and 29 minutes before it froze. I don't think it should freeze with the charger plugged in at all.

edit::
A small correction. The ROM hash test i was doing still fails if the test ROM is written to the upper half of the NOR flash. Can it be that i got a faulty cart again?

edit2:
Found out what's wrong. NOR erase and/or write is unstable. I can see single bit errors in the written ROM data.

edit3:
Attached my ROM test tool so you can experiment with it. It uses the fastest ROM wait states used by original games. The ROM hash is in its file name.

 

[EZ-Flash2]

I took some sample cards to do the test, including PSRAM and NORFLASH, hash results are 2ff at the beginning, padding are no mismatches found.

We have heard in past cases that NOR writes can be problematic with insufficient battery power, can you perform NOR write tests while plugged in?

Also, have you tested writing the commercial games like mother3? On cards that do have bad bit blocks in norflash, the game doesn't even pass the title screen properly.

 

[ghjfdtg]

@EZ-Flash2
I did test with the charger plugged in. I tested NOR write on a different AGS-001 and a DS lite both with charger plugged in. I get random bit flips on all 3 consoles during NOR write. Usually it's a bit that is supposed to be 0 but it flipped to 1 during write. And it seems to occur far more often in the lower half of NOR. Almost never in the upper half. I know my test tool is the worst case with 0x55 padding however i do think it should pass this test. Real game data might trigger bit flips too.

Mother 3 (Japan) checked against no-intro db, works and reaches in-game. Tested until the scene where you wake up in bed. Tested multiple times.

Also news on the freezes:
I tested on the 2 other consoles both unmodified stock. The second AGS-001 survived 6 hours and 26 minutes. Suspiciously close to the first GBA SP. The DS lite survived 16 hours and 31 minutes. The DS lite does have a very worn battery. All tests with charger plugged in. I might call the DS lite result a pass. Not sure. But this flashcart definitely doesn't pass this test on the GBA SPs for me.

 

[turkilmaz]

Hello,
I have the exact same issue, I have re-done all the tests using the RTC reader and my DS. I can see the month and weekday moving and having no-sense values. This values are reset to the correct ones when a new minutes starts, after few seconds the numbers started roaming again. So, I have few seconds every minute to start a game that requires clock without error messages, then I don't know if I will have issues during the game.

If I understood correctly, the solution for this issue is ask for a replacement, is it correct?

By the way, I received my card on November 9th 2021, so more or less 2.5 month ago. I suppose I am still covered by the warranty but I would like to have a confirmation from EZ-Flash and at the moment this thread looks the best way to contact them.

Thank you in advance

EDIT:

I used the ez flash omega DE until know with games that did not require the RTC. Today I decided to try Pokemon Emerald. I did not have any problem until today.

 

[EZ-Flash2]

Hello,
I have the exact same issue, I have re-done all the tests using the RTC reader and my DS. I can see the month and weekday moving and having no-sense values. This values are reset to the correct ones when a new minutes starts, after few seconds the numbers started roaming again. So, I have few seconds every minute to start a game that requires clock without error messages, then I don't know if I will have issues during the game.

If I understood correctly, the solution for this issue is ask for a replacement, is it correct?

By the way, I received my card on November 9th 2021, so more or less 2.5 month ago. I suppose I am still covered by the warranty but I would like to have a confirmation from EZ-Flash and at the moment this thread looks the best way to contact them.

Thank you in advance

EDIT:

I used the ez flash omega DE until know with games that did not require the RTC. Today I decided to try Pokemon Emerald. I did not have any problem until today.

According the Basic Repair Guide of Current EZ-FLASH Products #2

If your clock jumping abnormally in kernel interface. it means the crystal oscillator may have problem.

You can contact dealer for replacement.

 

[Barranqueiro]

Sorry for reviving this thread, but i was testing a NOR type cartridge with your rom test and im getting a different code but no mismatch on the pads. What could be causing that?
(Including a photo with the different hash code)

 

[ghjfdtg]

Either corruption in the code/data or patching. What kind of cart did you use? Also make sure hashing the file results in the hash in its name.

 

[Barranqueiro]

Either corruption in the code/data or patching. What kind of cart did you use? Also make sure hashing the file results in the hash in its name.

I'm using a old NOR flashcard (elink) that I got working, but don't know why I can't make games work on it, so i decided to test the NOR using your rom.
Is there a way to fix the corruption? (Sorry for the delay, busy week on uni)

 

[ghjfdtg]

@Barranqueiro
Unless the writer software you use patches the ROM it's probably broken. You can retry writing a few times and see if you get it working.

 

[timdmackey]

This is some awesome sleuthing @ghjfdtg!

@EZ-Flash2First of all i know exactly what's going on with the RTC. In game the month and day of the week values go out of bounds and increase even further until the minute counter reaches zero. Then it goes back to normal but will increase again. I made a video showing this behavior.

So did the RTC issue end up being fixed when you got your replacement? Or is it a software issue as you initially suspected?

 

[ghjfdtg]

@timdmackey
At the end i can't tell if this was a pure hardware fault or a combination of software and hardware tolerance. Replacing the cart made RTC work correctly. It's still working but the RTC chip/crytall is not very accurate. Drifted like 7 minutes in 3 months.

 

[timdmackey]

@timdmackey
At the end i can't tell if this was a pure hardware fault or a combination of software and hardware tolerance. Replacing the cart made RTC work correctly. It's still working but the RTC chip/crytall is not very accurate. Drifted like 7 minutes in 3 months.

Thanks for the info. Yeah, I figured the RTC wasn’t very accurate. 7 minutes in 3 months is actually way better than I was expecting though .

I’ve had some weird issues with the clock in the past, I’m pretty sure I’ve had to reset it a couple times. I haven’t played any games that depend on it though, so I’m not totally sure. I’ll keep your test ROM in hand in case I ever need to diagnose, so thanks for that!