The debug-port 1CV5 is unpopulated but I soldered in wires on TXD-SERVICE and TXD-STANDBY so I could monitor them with serial adapters.
Is there any difference between those two serial ports? While I can get clear communication from TXD-SERVICE (with expected parameters of 115200-8bit-1stop-no parity-no flow control), I just get garbage from TXD-STANDBY. I assume that signal level is 3.3V in both cases.
Is there any difference between those two serial ports? While I can get clear communication from TXD-SERVICE (with expected parameters of 115200-8bit-1stop-no parity-no flow control), I just get garbage from TXD-STANDBY. I assume that signal level is 3.3V in both cases.
Yes, the levels are 0-3.3V TTL on both ports.
It said in the service manual that the main processor was set to 115200 8N1, and by simply testing other speeds I discovered that the standby processor was set to 57600 8N1.
I used a USB-to-TTL serial-adapter: https://www.adafruit.com/product/954
This one with the blue casing from Adafruit uses an original chip, so the device works with the official Windows-driver.
If you find one that's half the price or cheaper, then it's most likely using a pirated chip, which the signed drivers in Windows will detect and refuse to work with.
The pirated ones still work in Linux, though.
I learned this the hard way.
I also got a 5-pack of small test hook clips with a male Dupont connector on them: https://www.electrokit.com/produkt/t...klamma-5-pack/
I soldered a short piece of bodge-wire onto the relevant pads on the PCB and then hooked the test clips to the wires I wanted to listen on.
I left the bodge-wires on the PCB with the ends tucked between a couple of layers of kapton-tape, ready for use the next time.
I discovered that the standby processor was set to 57600 8N1.
Thanks. I tried other lower speeds without luck, but not this one!
I use a USB serial adapter with FTDI FT232BM IC and a Sipex SP213ECA IC for standard serial voltages. I disabled the second IC (Enable to GND), and changed FT232BM signaling voltage from 5V to 3.3V connecting VCCIO (p13) to its own 3V3OUT (p6).
The problem is that I tried TXD-STANDBY first, and as it wasn't good looking (due to the speed difference you point out), I blamed my modifications of serial adapted, but after giving up with that port I tried the SERVICE port and it worked fine!
Re: Philips 60PFL8708S/12 - QFU1.2E LA - Error code 53
Why you're trying to access standby ser.port? I think it has no useful info in it, i have tryed with a Chinese board theese days, st.by serial only prints few characters at boot and 2 char. at shut down. It is a the port of the st.by processor that has a simple software, not much to communicate..
Re: Philips 60PFL8708S/12 - QFU1.2E LA - Error code 53
Well, I just wanted to check that I get exactly the same log outputs in both ports than Maalobs. Also, I wanted to make sure that my mods on the standard serial USB adapter were working fine.
I have ordered a programmed NAND and I’ll report results after replacement.
Re: Philips 60PFL8708S/12 - QFU1.2E LA - Error code 53
I still have the dumps of all the flashroms from the TV, even the NAND which is too big to post in the dump-thread on this forum.
Let me know if you need it, I can arrange an sftp link for you.
Re: Philips 60PFL8708S/12 - QFU1.2E LA - Error code 53
Error 2 / 53.
Thanks Maalobs. I have an XGecu TL866II Plus that can read/write these NANDs, but I don’t have a TSOP-48 adapter, so I’ve prefered to order an already programmed NAND for this chassis.
Re: Philips 60PFL8708S/12 - QFU1.2E LA - Error code 53
But, what caused the first improvement in Maalobs TV? CPU reflow or NAND replacement? Since he did both things at the same time, it wasn’t clear to me which one was the root cause. I admit that I’ve chosen the NAND part first because CPU reflow is out of my reach.
Re: Philips 60PFL8708S/12 - QFU1.2E LA - Error code 53
Does your TXD-SERVICE output end on <020>K like it did for me and Alpengeist?
If that's the case then you probably need to reflow the Fusion processor, and I'd suggest that you wait until after the Fusion reflow before you replace the NAND.
With the Fusion processor running, you can record and inspect the output from TXD-SERVICE to see how bad the existing NAND is.
Then if you replace the NAND, you can record the output from TXD-SERVICE again and compare it to the previous output.
It's easy to compare the outputs with a diff-tool like WinMerge, the boot sequences will be a lot of mostly similar lines, this way it's easy to spot the differences.
The reflow was what got my TV working, as described in this post.
It was when I got that far and had the Fusion booting properly, that I discovered in the proper TXD-SERVICE output that the NAND had multiple bit errors too.
So I decided to replace the NAND to extend the useful lifetime of the TV.
Re: Philips 60PFL8708S/12 - QFU1.2E LA - Error code 53
Have you been able to gain root access on any TV with a current software image installed?
Back in April 2019 I read through a several years long comment thread on neophob.com about it as I mentioned in this post.
That page only remains in the Internet Archive now, but as I recall, it had been fairly trivial in the early days to gain root and then pick out the crypto keys from the filesystem for the upg-releases (the publicly downloadable encrypted archives that contain upgraded factory images to all internal memories in the TV).
Over the years however, Philips incrementally locked down the security until eventually no one in the thread was able to get root again.
RXD-SERVICE and RXD-STANDBY were totally dead on my TV so I couldn't write anything into the terminal, and I think I remember from the neophob-thread that the last thing Philips had done was to encrypt one of the filesystems in the NAND, so it wasn't even possible anymore to manually binwalk the NAND-dump to identify and mount that particular filesystem, and Philips hid the crypto keys in there for the upg-files.
I don't remember the details anymore about all the hacker stuff, but I read through all that for a possible way to get hold of a factory clean NAND-dump from the upg-file, since my running theory for my TV's 2-53 problem at that time was bit errors in the Flashmems or the EEPROM.
Does your TXD-SERVICE output end on <020>K like it did for me and Alpengeist?
My Service log is identical to yours (except the very first 32bit hex number, that looks like a boot cycle counter). This determinism lead me to rule out broken CPU solder points, because something like that is not very deterministic, but rather random (broken solders will be in a given zone, but you can’t tell which one will broke first).
Re: Philips 60PFL8708S/12 - QFU1.2E LA - Error code 53
I'm afraid I don't follow your reasoning; if you have the identical symptom in the TXD-SERVICE output as me and Hermann Alpengeist, how do you come to the conclusion that your problem is not identical to ours?
I don't think it has to be the same BGA-pads that disconnect when the PCB flexes from the overheating; just any data-, address-, control-, power- or whatever pads get loose and now the processor won't boot anymore, or the 8051 standby-processor doesn't allow the Fusion to boot if it doesn't pass the power-on-self-test as a consequence of any of these loose pads.
Reflowing a big BGA-chip like that was daunting to me too, I was just lucky in that I got help from an experienced professional who did the reflow for me.
Re: Philips 60PFL8708S/12 - QFU1.2E LA - Error code 53
Yes Maalobs i did it in some Chinese boards like Akai and brothers.. often if you press enter or ctrl+c you can access prompt. Obviously if the failed partition is the system or something with programs in use the part. is locked and the job gets bigger a bit. Chungalin i've seen so many of them failing with cpu, usually the fault worsen with time, first there are random crashes or no boots, this is a proof of soldering prob. Last but.not least also the software problems are random.
Chungalin is the name in Spain for the cartoon of your icon?
Comment