TCL 55P607 / 55P605 tcon dump

[ziddey]

Reassuring Do you still have the failed tcon flash image? Would be interesting to compare it with the 2 failed ones I posted to see how similar.

seem to be many differences. At this point, I'm not sure if it's just a different version, or a bad dump.

[howardc64]

googled "25Q40 corruption" This TCL TCON isn't a unique case. Samsungs and some monitors also have failures as lotus mentioned. Found the attached document on NOR flash power up down sequence.

I guess its easy to expect these NOR flashes are simple non volatile storage devices. But in reality, its quite complex. Physics of storage mechanisms of storage devices are inheriently analog and often in the realm of quantum mechanics as designs aim to get more bits / area. These NOR flash, have pre-stored parameters called "configuration settings" that are loaded into internal registers during power up to start the device. These configuration settings include many manufacturing details including yield improvements (substituting random bad blocks from manufacturing with spares) and other manufacturing varying characteristics of analog devices manufacturing.

Document details out the proper voltage power supply sequence during power up, shutdown, brown outs etc. If not properly handled, can cause error during this initial "configuration settings" transfer to internal registers which then can result in stored data bit filps. The complexity involved seems quite high for a simple TV board designs to include all the cases.

Anyway, maybe power sequence of the TCON isn't robust enough for this complex device and encounter this error at a low random frequency. Maybe newer version of 25Q40 is more robust at handling less robust power sequences as designs improved over time to handle problems observed in the field. Probably need NOR flash design experts to have any real insight.

https://www.renesas.com/en/document/...-bO6Dr4Nzn8ZTR

[howardc64]

seem to be many differences. At this point, I'm not sure if it's just a different version, or a bad dump.

Yes, bad dumps have many difference compared to good.bin but many are just flipped bit 3 in a big endian nibble (bit 4 3 2 1) so quite a few numbers incrases by 4.

your bad.bin compared to my 2 bad dumps differ by only 2 bits. So these corruptions are happening in a very similar way.

[howardc64]

Found datasheets for GD25Q41BT and GD25Q40C*, can't tell much differences. Here are some info

Datasheets

https://www.alldatasheet.com/datashe.../GD25Q41B.html
https://www.gigadevice.com/product/f...flash/gd25q40c

Decoding Chip Markings

(pg 46-41 of GD25Q41B datahseet)

Q = 2.7-3.7V
40/41 = 4Mbit
B/C = Revision
T = SOP8 150mil (package spec)

Datasheet Dates

Data sheet revision history of GD25Q40C* shows initial date of 2/3/2016 (pg 63) which is before manufacturing date of these TCL TVs (2017) with this TCON board. So these 2 NOR flash is likely substantially the same spec wise.

Data Retention Lifetime

GD25Q40C datasheet feature set (pg 4) does say 20 year data retention typical. So maybe its good to replace the flash that is near 50% of its typical data retention duration. The GD25Q41BT in a 2017 TCL TCON would be 8 years old now while a new GD25Q40C is presumably much younger and with different semiconductor fabrication chacteristics.