Sometimes, TinyG changes aam parameter by itself

[jkuusama]

Sometimes, TinyG changes aam parameter by itself. I have no idea what causes this. So far, a few people have reported it and only one more than once. As a workaround, there is a check at connection (commit 1d826ad, 14.11.2022). If the value is not right, the correct value (1) is written. This is really a problem hiding workaround, not a fix!

[ril3y]

https://github.com/search?q=repo%3Asynthetos%2FTinyG%20cm_set_am&type=code

This is odd. cm_set_am is only called in this one place. Wonder if there is some sort of memory corruption going on?

[jkuusama]

Maybe. Or a timing issue associated with"set defaults" command, which writes all settings. In other words, lots of eeprom activity in a row. I suspect the latter, as even when this happens, a rewrite fixes it. Even before I hid this, nobody reported seeing the issue again. So it could be kind of "weak" write that didn't take or faded away fast

[ril3y]

Yes, the xmega needs small delays when writing to the EEPROM. I don't recall the exact amount but we had to do this in UI's when applying lots of settings in a row.

…

On Thu, May 18, 2023 at 1:33 PM Juha Kuusama ***@***.***> wrote: Maybe. Or a timing issue associated with"set defaults" command, which writes all settings. In other words, lots of eeprom activity in a row. I suspect the latter, as even when this happens, a rewrite fixes it. Even before I hid this, nobody reported seeing the issue again. — Reply to this email directly, view it on GitHub <#176 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AABYSM2IVGQVEIYFB6MZHVDXGZMNBANCNFSM6AAAAAAR74N5KQ> . You are receiving this because you commented.Message ID: ***@***.***>

[jkuusama]

The software has that, with some margin. But wouldn’t be the first time that a part doesn’t always and in all cases work exactly as the data sheet says - and the workaround is already there, so this is on the open list more for principle than for need of fixing. Riley ***@***.***> kirjoitti 18.5.2023 kello 20.59:  Yes, the xmega needs small delays when writing to the EEPROM. I don't recall the exact amount but we had to do this in UI's when applying lots of settings in a row.

On Thu, May 18, 2023 at 1:33 PM Juha Kuusama ***@***.***> wrote: Maybe. Or a timing issue associated with"set defaults" command, which writes all settings. In other words, lots of eeprom activity in a row. I suspect the latter, as even when this happens, a rewrite fixes it. Even before I hid this, nobody reported seeing the issue again. — Reply to this email directly, view it on GitHub <#176 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AABYSM2IVGQVEIYFB6MZHVDXGZMNBANCNFSM6AAAAAAR74N5KQ> . You are receiving this because you commented.Message ID: ***@***.***>

— Reply to this email directly, view it on GitHub<#176 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/ABDYRZSEN3DKEV7C4GEMF7TXGZPO5ANCNFSM6AAAAAAR74N5KQ>. You are receiving this because you authored the thread.Message ID: ***@***.***>