BRT Community

Hi Rudolph, you can add MT25QL01GBBB to the list (specifically tested MT25QL01GBBB1EW9-0SIT):

Manufacturer   Chip   Size(MBit)   Detect   Erase   Read/Write   CMD_FLASHFAST
Micron   MT25QL01GBBB   1024   Yes   Yes   Yes   Yes

Hi, did you get any new information about that compatibility topic?

Thanks for the fast answer.
I am using 47K on CS, SCLK, IO2 and IO3. No pull ups on MOSI/MISO. I also tried it one time without any pull ups and one time with 10K.
The pins are not multiplexed by something else but are just connected to the BT817.

On the exact same board, the "W25Q128JVEIQ" is working fine. Both have the same package so I replaced them for testing.
A special thing about the "W25Q01JV" is, that it has stacked dies (2x 512M-bit), but they are both linear addressable.

Any information about differences and especially compatibility of the BT815 and BT817 blob file from Eve Asset Builder would be appreciated.

Hi,

I am trying the W25Q01JV flash with 1Gbit size and have problems while flashing. When I do a crc calucation after I wrote a sector (4K) I get errors on very specific sectors. All of them are the odd indices. F.e. Sector 0 (0-4095), sector 1 (8191) fails, the next one is successful again.
Just to show an example, those are the first few sectors which fail: 37, 43, 45, 67, 69, 73, 75, 77, 79, 97, 99
It seems to only happen, when I switch to FLASH_STATUS_FULL after writing the Blob. So I stopped to switch to FLASH_STATUS_FULL when writing on the flash.

But whenever I don't switch to FLASH_STATUS_FULL I cannot write more than 16 Megabyte. I get address errors back when using CMD_FLASHUPDATE (invalid address). This is the error I get when reading out RAM_ERR_REPORT:
illegal flash address in cmd_flashupdate()

This error I would usually get when my Flash would have a size of <= 16 Megabyte, but I am writing after that address. The BT chip seems to get the correct size, because I get "128" back for REG_FLASH_SIZE.

Is there a compatibility issue with that Winbond flash? I checked this list and did not find any flash with 1Gbit size, which I need: http://www.brtcommunity.com/index.php?topic=77.msg1658#msg1658

I also changed the flash on my board to make sure that it is not damaged, but had the issue with 4 different pieces. Do you have any suggestions how to debug that or any information if that Winbond flash has been tested by someone already?

Hello,

there are 2 different blobs: "bt815.blob" and "bt817.blob" in the EVE Asset Builder. What is the difference between them? I am interested in a feature list, which requires the bt817 blob instead of the bt815 blob when used with a BT817 chip.
If my tests were correct it seems that the bt815.blob is compatible with BT815 and with BT817 (both go successfully into full mode), but bt817.blob seems to just be compatible with BT817 and if used with a BT815 it does not go into full mode.

I am asking that because I am using both BT815 and BT817 in production and would like to share the same external flash data for both. So it seems to work fine for be when just using the bt815.blob, but I would like to know if there are any disadvantages for the BT817 in that case.

I am using the width values for each letter which can be found in the xFont file. I read them from RAM and then add them together to get the width of the whole string.
In the programming guide (https://brtchip.com/wp-content/uploads/Support/Documentation/Programming_Guides/ICs/EVE/BRT_AN_033_BT81X_Series_Programming_Guide.pdf) you can find the information in "5.4.3 Extended Font Metrics Block". It is the last line of the table.
Those values work very well for me.

Just a comment about "It does not seem to do anything though, at least not in regards of saving the pixel engine from crashing when there is not enough time per line.":
I tried enabling and disabling REG_ADAPTIVE_FRAMERATE on a project which needs to display a lot of text. It uses also a custom font in the flash.
I analyzed it with an oscyloscope and could see the differences. If you enable it (= using the reset value) you can see the Display clock pin having breaks/pauses in between the usual frequency which you setup.
I was also able to view the influence of that setting right on the display: In a displaylist which is using the processing limits (so again a lot of text loaded from flash) my display started to flicker a lot when I enable REG_ADAPTIVE_FRAMERATE. If I disable it it just stops showing the most right part of the display because it cannot handle it in time.

And I aggree that having multiple font caches for different fonts / sizes would be great. The font cache of BT817 is a huge advantage compared to the BT815 in my opinion.

I was able to test it with a BT817 now and the updating of the firmware works for me. So it seems to be an issue just on the BT815. I was able to run my old test code with fixed values and also to test it with another touch controller (to read out a register). So I am fine with just using the BT817 in the future then. For the BT815 I will fallback to host mode.

Thanks for the answer. I will try to test it with a BT817 in the future. In the past I just used the BT815 for testing the touch firmware update. It would be fine for me, if it just works with BT817. I will switch to the newer version then.

This really gets frustrating at some point, I am waiting for an answer for close to one year now. Besides the initials answer that you could reproduce it on your side I never got something back. Could you please check that? If it is reproducable on your side, it would help a lot to know if this feature is not working at all or if there are just specific cases. I provided test code without any specific hardware needed to reproduce the issue.

I used the Ili2511, but also tried without any touch controller and just used fixed numbers instead of i2c connection which seemed also not to work (Described it in the first post of this thread).

Quote from: BRT Community on July 02, 2021, 11:29:29 AM
Hi David,
Just to update you, we tried your code and could see the same result, we're just investigating and will post an update soon,
Best Regards, BRT Community

Did you get any response about that?

That's what I am using:


BEGIN(RECTS)
if (roundedCorners) {
LINE_WIDTH(10 * 16)
} else {
LINE_WIDTH(1 * 16)
}
VERTEX2F(x * 16, y * 16)
VERTEX2F((x + width) * 16, (y + height) * 16)

And I get completely sharp corners (if you use the !roundedCorners version).
You said you would need to shrink it, do you still get rounded corners when you use 1 * 16 as width or did I misunderstand the question?

Hello, thanks for the information.

Option 1 is actually not a workaround for me, because I want to have the cache enabled just on specific subviews and on others I need a lot of RAM for other things (image streaming). I think I would see some black screen or daly when switching between those views, if I would do a cmd_coldstart in between.

Option 2 is also problematic, because I am using the same font on both of those views, but just on one of them I need caching (because there is more text). So just not using it on the other view would not be an option.

I found a workaround which looked promissing for me: I am setting the cmd_fontcache to a font handle which does not exist (some lower value like 30, so inside of the allowed handle range). Would you recommend that or could other problems on the BT817 happen with that, if the font handle is not registered by CMD_SETFONT2?