SAA7146 / TT1.3 stream corruption

Hello
I am using Technotrend Rev1.3 for many years. But last time suddenly i find out strange problem. Seems that in some circumstances the card can't decode stream from satellite properly. I don't know for sure but it could be a driver problem, firmware or even (worse) a hardware problem.
I know that for now the problems occurs for example on HotBird (13.2E) Transponder 11158 V, SR 27500, PID 0x010B - but also for others from this ECM provider.
The stream is ECM stream - most of the time the stream is OK, and the ECM's have FF's at the end, but from time to time the ECM's seems be wrong - are longer and have a garbage instead of FF's. The garbage is same for whole "ECM period".

two ECM examples:
saved from TT1.3:
00037ff0h: 47 41 0B DD 00 81 30 8F 07 8D B4 8D 2E FE 3F C1 00038000h: 9B 14 74 AD 17 FC AA B4 3F F4 86 5D 74 E2 B9 AD 00038010h: 56 52 93 4D E3 59 79 44 E3 87 1B CF 6B 41 3E 76 00038020h: 18 8C 07 62 B0 93 AA D4 24 0B 5B E0 00 F1 53 45 00038030h: 85 CA E1 C4 99 D0 81 5B 59 65 22 A1 4C 2B 2C 6C 00038040h: FB A4 97 60 DF B2 0B 6B 46 A9 D8 EF 6E AB DA E3 00038050h: 91 8E 6D 08 37 73 DE E0 AE 15 80 19 24 EF 3C A3 00038060h: 3B 30 D1 56 D5 90 D0 5A DF 76 29 2A 16 E6 0A 77 00038070h: 5A 46 E8 28 46 A2 12 24 43 F6 DE CE 08 3C EE 3E 00038080h: 4E C9 43 5E 4E 75 94 B8 A7 22 61 A8 88 C9 9E 94 00038090h: CD 60 AC 6C 69 35 28 F2 03 91 B1 A2 5B 26 EA 58 000380a0h: 53 11 F1 EE AF 8A 66 9D F7 47 D7 DA

saved from different DVB-S card:
00037ff0h: 47 41 0B 1D 00 81 30 8F 07 8D B4 8D 2E FE 3F C1 00038000h: 9B 14 74 AD 17 FC AA B4 3F F4 86 5D 74 E2 B9 AD 00038010h: 56 52 93 4D E3 59 79 44 E3 87 1B CF 6B 41 3E 76 00038020h: 18 8C 07 62 B0 93 AA D4 24 0B 5B 91 14 8F F4 83 00038030h: 40 6B 0D 91 EC 74 FB 06 4C 8C 4A 90 36 BA 67 11 00038040h: 7B D8 22 2B 77 C2 8B 61 C7 ED 4A 25 21 6B DB 1D 00038050h: 5F 40 AF 12 1C F5 53 2D 2B 84 00 E1 AC 63 9A 76 00038060h: 57 A2 E6 46 25 BC D6 C6 89 D1 40 07 AA 27 30 58 00038070h: A0 81 91 E2 B3 8A BD D8 F8 A3 6B C5 98 77 E4 81 00038080h: A7 52 E5 56 D9 A6 58 FF FF FF FF FF FF FF FF FF 00038090h: FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF 000380a0h: FF FF FF FF FF FF FF FF FF FF FF FF

saved from TT1.3:
00055c8ch: 47 41 0B D6 00 81 30 8F 07 8D BA 7D 24 B2 E6 3A 00055c9ch: 86 4E FA E3 75 80 B6 E6 03 E6 C7 55 62 10 7C 92 00055cach: 4A DE 56 AD 86 45 6F 70 43 8B 00 D0 56 85 67 6F 00055cbch: AC 05 97 F8 1D AB 6F F5 C8 FE 20 8B E2 2C B0 59 00055ccch: 0F C5 80 38 38 BF 9B DA 66 25 E3 BB A0 E3 39 7E 00055cdch: CB 45 7F 62 AD E0 D0 60 88 90 90 AA EE 1D 6E FE 00055cech: 17 9F 81 E2 52 2A 4E 74 3B 8F C0 3E 5B 0E 57 4C 00055cfch: 8F EC CE EA C2 EA 10 F3 4A 57 F2 C2 48 C5 34 0D 00055d0ch: CB 2E 2C CB 62 84 1F 9D 90 3B 7C C4 37 F1 1D 1D 00055d1ch: 82 90 FB DA 15 0C C0 C7 8F 33 1E 5A 1A EE 44 43 00055d2ch: B3 B3 AF D5 97 D6 45 2E 39 2D E6 02 45 07 13 20 00055d3ch: BE 5A EF 7D 21 E6 4E 31 7E 8F F8 AD

saved from different DVB-S card:
00055c8ch: 47 41 0B 16 00 81 30 8F 07 8D BA 7D 24 B2 E6 3A 00055c9ch: 86 4E FA E3 75 80 B6 E6 03 E6 C7 55 62 10 7C 92 00055cach: 4A DE 56 AD 86 45 6F 70 43 8B 00 D0 56 85 67 6F 00055cbch: AC B7 66 86 D8 FF 41 28 8E 02 FC 0E E7 B2 E5 5C 00055ccch: 9E 81 8F 06 BE AD C0 FC 49 9F 3A 23 88 98 F7 EF 00055cdch: C9 F5 37 4C 57 31 B9 C2 1B BE BC 65 BE 60 CC E6 00055cech: F3 E5 A3 AD 76 03 20 11 25 09 88 A6 42 32 5D 2B 00055cfch: 06 65 B8 B6 B8 4D 3E 0B 77 97 EE D1 FD A6 DE CC 00055d0ch: 6C B6 BB F4 B1 78 39 8E 6D 9F 14 3D 39 64 86 42 00055d1ch: 58 61 0D 78 DD C7 64 FF FF FF FF FF FF FF FF FF 00055d2ch: FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF 00055d3ch: FF FF FF FF FF FF FF FF FF FF FF FF

As you can see the begining of the ECM are same, but the problem with garbage occurs later (not always after the same byte in stream).
The above results comes from saving the PID in the same time on my TT1.3 and friend's other DVB-S card.
I also find out that "affected" are only ECMs which starts with: [81|80] 30 8F 07 8D and the following low-nible of byte is A or B.

I was trying to collect as many information as i can - and here they are:

1. the problem is not related with any specific dvb application i was fetching the "pure" stream using i.e. such command:
dvbstream -f 11158 -p v -s 27500 -o 267 > my_file.raw

2. the problem also occurs on different physical TT1.3 cards
- i was testing it on several computers and several TT1.3 cards with different Satellite instalations - and it occur always on TT1.3 cards.

3. the problem is also on Windows - i heard that firmware (dvb-ttpci-01.fw) is very similar/based on windows firmware so maybe the bug was "copied"

4. a screen from binary diff and same sample of ECM data from TT1.3 and other card are on my homepage:
http://manio.skyboo.net/dvb/

5. of course i am using latest dvb-ttpci driver from current stable kernel

so that's all what i could collect about it...
Please help!

regards,

manio
jabber/e-mail: [email protected]
http://manio.skyboo.net

Now i can reply myself to provide info for users with similar problem.

The parameter which i need is:
hw_sections=1
When i load dvb_ttpci module with this parameter the stream is correct. Just by the way: people on DVBN forum has similar issues, but this parameter was not sufficient - they also need to write a patch for select pid ranges for providers (in their case: BEV and DN) More info on dvbn topics#: 42822 and 42653

regards,

manio
jabber/e-mail: [email protected]
http://manio.skyboo.net

Hi,

If you have very good soldering skills, there is the ultimate solution for all kinds of stream corruption with full-featured cards:
http://www.escape-edv.de/endriss/dvb-full-ts-mod/

CU
Oliver

----------------------------------------------------------------
VDR Remote Plugin 0.4.0: http://www.escape-edv.de/endriss/vdr/
4 MByte Mod: http://www.escape-edv.de/endriss/dvb-mem-mod/
Full-TS Mod: http://www.escape-edv.de/endriss/dvb-full-ts-mod/
----------------------------------------------------------------