No channel waiting for time fix!!

[ikeabuchi]

Had a little spare time on my hands. I had a intermittent box and another 3that showed the same fault but remained in that fault state indefinitely,
I did a Bootlog of working and non working and compared both of the log files,
line by line till I got to the error,
[camd] CA_SEND_MSG: Remote I/O error
this is refering to the avia_gtx it is not attaching at all and not giving a ID to microprocessor from there it will get its memory resource, and irq etc
what it should have done was sucessfully attach ie
avia_gt_core: masking unhandled irq reg 0 bit 7
i.e from working one
looking through the source code of the gt_core.c. you can see how it does this, but to cut a long story short, the problem, is the small copper plated feed throughs underneath and around the avia chip this get pretty warm, and causes heat fatigue, also the feed through should from the manufacturacting stage be plug with resin to cap them off this prevent oxidisation,
I heated and freezed the board in various locations, and concluded that the most stress was around the boundaries of the gtx chip this is where the heat differential is the greatest so using a scalpel I gently placed into the feed through rotated it carefully to expose a sliver of copper then apply solder till the solder permeating thoroughly into it, I know it is tedious but have patience,
after getting into a routine i did this to about 40 feed throughs ,even the ones on the underneath under the avia 500 ( the board);
I then turned the box on Presto !! first time changed to a an encrypted channel
waiting for about ten seconds and it came up checking the epg informations, they also updated succesfully .Doing the same repair procedure to the intermittent box, and now that ok!! I am not saying that this will work with all boxes but should have a high sucess rate (I have a perfect sucess rate has of now), if it isnt then you will need to scope some various points, I happen to have a circuit diagram for the box, if you want me to go into more detail then I will cheers

 

[scooby do]

Nice,

Could you post some pictures

 

[ikeabuchi]

Sorry for the delay I have been looking deeper into this particular problem, I
I have manange to get the data sheet for the vs1820 dvb channel reciever and there is a strange quirk, I have a working box and I have been taking voltage reading and waveforms on various points of the tuner and SEC chip has well as fifo memories it seems that when every thing is up and running pins 19-30 of tuner the waveforms are time varying signals with the displayed picture ,but on a non working box they are static, I will post the pictures of the signals laters, I have a sneaking suspicion that when replacing the Tuner that the feed throughs have become damaged and even though its a replacement new tuner, it fails to work, (be very careful when desoldering and resoldering this part, one inmportant signal is the signal pulse on 19 of the Tuner (with the aerial input facing towards you and looking underneath the main board pin 1 is closest to you) this Pin should be at a level of 3 volts and have narrow pulses 23.5uS apart of amplitude 3.0V, Do a resistance measurement (box off) from ground to this pin and if it is lower than 1.0k on the diode range of meter with common lead of meter to ground, then disconnect
it and the corresponding pin of the SEC chip or track which is adjacent to surface mount capacitor cc8 right hand side running up then resolder it back this is after I opened the cannister (Tuner Can Shiled ) snipped the leg from the inside and rejoined and it corrected the 36ohm leakage from the ves1820, still there was no channels I came across another leaking component in the tuner the most important signal is the 36.125MHz carrier from the bandpass filter measure across pin 13 and ground on the diode range, one way it should measure a value of 1k and the other way measures around 400 ohms (this measurement is also across cf18) if not then the intermediate frequency isn't at a sufficient level is being attenuated and cant be processed by the bandpass filter type x6966d to lock onto the 36.125MHz carrier hence no signal, this is what I am in the process of doing and will update you when I have ,

 

[ikeabuchi]

after exhaustingly changing the sec chip still the same problem, changed the gtx chip same problem then notice that the epg are store in the non volatile memory guess what!!! changed it bingo it works,
Can of freezer spray freeze IG01 it and reboot, DRAM 1Mx16 not that this will fix all problem since other problem may be tuner,
this will fix the majority of them,

 

[ikeabuchi]

repaired another box, fix looks good at the moment, left two boxes on all night still working, Oh forgot to say took the good chip back out and soldered the old one back in to confirm, And yes no epg waiting for time lol, good chip back in, And up and running,

 

[Fergalsworld]

Can you just explain in laymans terms what you had to do
did just use freezer spray on it
Thanks

 

[ikeabuchi]

Yes certainly, It looks like various symptoms will cause this problem so you need to methodically go thorugh various procedures, When the problem occurrs try freezing various areas, The one that worked for me and on 4 other stubborn boxes was to freeze the Dram chip IG01 type VG26V18165cj
this is the one on the right hand side of the gtx chip freeze it and hopefully your channels will come back and epg are stored into this non volatile so internal failure in the chip,
You may find that the box will work for days on end then come back to the fault, just done another box

 

[plasterer]

Are The pics on the way
As they will be much appreceated .....................Paul

 

[ikeabuchi]

I will write alittle report with pictures voltages and waveforms, but not sure what details I must go into, Since I Dont want to make it overly complicated, If you have a hot air gun be careful when you take the chip out, I have the necessary equipment so no damage to the Board,

SDRam GTX - [email protected] - Shop

 

[ikeabuchi]

Also one things comes to mind that the proximity of the Sdram to the powersupply that perhaps a glitches on the powerline causes a failure in the chip,

 

[Killer_K]

thanks for all the info you have posted here mate, a lot of people will be thanking you for this ;-)

looking forward to photos and instructions.