Audioniek Images: download

[Audioniek]

I already spotted that the Sony CXD2820 is a DVB-T tuner. The problem is the driver. For the this Fortis model two different DVB-T tuners were sold. The early model does not support DVB-T2. It has a Sony CXD2820 tuer and a Samsung 5SH1432 demodulator. Both the driver and the datasheet for the latter are confidential and for that reason are not part of the image I can offer. As far as I know a driver can be found in Bpanthers' Neutrino images, but this will not solve your problem, as only the later DVB-T tuner supports DVB-T2. At present I do not know which chips it is based on.

In the past I have started work on a driver for the older model, but my tuner module broke down while working on the driver. These tuner modules are hard to find, as the majority of the receivers were sold and used with 2x DVB-S(2) as tuners and this receiver model is slowly coming of age and re-/disgarded as obsolete, although it is my favourite Fortis model.

Regards,

Audioniek.

[mrspeccy]

I already spotted that the Sony CXD2820 is a DVB-T tuner. The problem is the driver. For the this Fortis model two different DVB-T tuners were sold. The early model does not support DVB-T2. It has a Sony CXD2820 tuer and a Samsung 5SH1432 demodulator. Both the driver and the datasheet for the latter are confidential and for that reason are not part of the image I can offer. As far as I know a driver can be found in Bpanthers' Neutrino images, but this will not solve your problem, as only the later DVB-T tuner supports DVB-T2. At present I do not know which chips it is based on.

In the past I have started work on a driver for the older model, but my tuner module broke down while working on the driver. These tuner modules are hard to find, as the majority of the receivers were sold and used with 2x DVB-S(2) as tuners and this receiver model is slowly coming of age and re-/disgarded as obsolete, although it is my favourite Fortis model.

Regards,

Audioniek.

Hi Audieniek,

thank you for your fast reply. I just tried the original Octagon 1028P software from 2017 and found that it allows me to tune in to DVB-T2 channels! So I seems I have the "newer" variant of the Sony CXD2820 tuner - sometimes this is possibly referred to by CXD2820R (https://www.linuxtv.org/wiki/index.php/Sony_CXD2820). The latter link also refers to open source Linux drivers but I'm not sure it helps. Do you have a link to Bpanther's Neutrino images?

(The situation is that I am located in Denmark where more and more channels are moved from the DVB-T to the DVB-T2 standard, so this is why I'm interested in receiving the DVB-T2 channels. In the worst case, I have to move back to the original Octagon software or buy a new receiver.)

The other issue I mentioned is that with your build of the atevio7500 image, the DVB-S tuner does not allow me to tune in to vertically polarized channels on Astra 19E - which works with the latest HDMU image, though. I know I am demanding a lot, but can you reproduce this issue with your DVB-S tuners on the Octagon/Fortis?

[pop1234]

hello mrspeccy & Audioniek

https://git.linuxtv.org/media_tree.g...23689a487970f2

https://git.linuxtv.org/media_tree.g...83e5c8c8e1da16

[Audioniek]

Do you have a link to Bpanther's Neutrino images?

Bpanthers Neutrino images are here. You to register to be able to download them.

The situation is that I am located in Denmark where more and more channels are moved from the DVB-T to the DVB-T2 standard

Same here. The Dutch provider KPN has been so kind to abolish DVB-T too soon, rendering millions of usable receivers worthless,

I know I am demanding a lot, but can you reproduce this issue with your DVB-S tuners on the Octagon/Fortis?

On my Rebox RE-8500's (I have several), all DVB-S(2) channels tune fine, so I cannot repoduce the problem.

Regards,

Audioniek.

- - - Aktualisiert - - -

hello mrspeccy & Audioniek

Thank you, but as said above, The Sony CDX2820(R) is not the problem, the elusive Samsung 5SH1432 is...

I will try and check if my RE-8500's can receive DVB-T2 with the newer tuner.

Regards,

Audioniek.

[Audioniek]

Identifying the plugin tuners for the Fortis HS8200.

This done easiest by the back plate writing. The DVB-S(2) is course easily identifyable by its F-connectors as the others have IEC connectors. It is also the only plugin tuner having a heatsink.

As far as I know, of the DVB-C tuner only one variant was available. Its backplate has the texts CABLE IN and CABLE OUT. The tuner module was made by Samsung has a sticker on it with the number DNQS44PP1010A.

The old and new DVB-T plugin tuners look exactly the same, and can only be distinguished on the outside by the sticker on the tuner module.

• Old (DVB-T only): DNOD44SQH102A.
• New (DVB-T and DVB-T2): DNOQ44QCH106A.

To complicate things further, the DVB-C, DVB-T and DVB-T(2) tuners are all built on the same PCB, only the components differ.

Regards,

Audioniek.

[mrspeccy]

The old and new DVB-T tuner look exactly the same, and can only be distiguished on the outside by the sticker on the tuner number.

• Old (DVB-T only): DNOD44SQH102A.
• New (DVB-T and DVB-T2): DNOQ44QCH106A.

Hi Audioniek,

thank you for your great support. First of all, for the DVB-T(2) issue: On my tuner it says DNOQ44QCH106A, so it seems DVB-T2 ready. This is confirmed by the original Octagon software, which indeed does tune in to channels in the new DVB-T2 standard. Unfortunately, this works with neither HDMU, Audioniek nor Bpanther's Neutrino images, which all return "no channels found" on the DVB-T2 frequencies. In the end, all these images seem to rely on the same driver from the Duckbox project (stored in the cdx2820.ko kernel module), so I guess this driver is the issue. Even though the source code (e.g. "cdx2820.c") refers explicitly to DVB-T2, the corresponding implementation seems buggy or is not active. The driver posted above by pop1234 seems to be different from the Duckbox project, and it is not obvious to me how to migrate it to sources relying on Duckbox.

Please correct me if I'm wrong.

Now for the DVB-S(2) issue: Thank you, Audioniek, for your extensive tests. I really don't understand what is going on. I can tune in to vertically polarized Astra 1E channels (e.g. KiKa HD) with the original Octagon, the latest HDMU from 2017 and Bpanther's image from 2020. It does not work with your precompiled atevio7500 Enigma 2 image and compiling it manually did not solve the issue either. I will try whether I can overwrite the avl2108.ko module from your builds with versions from other images.

Edit: Just to make sure it did not get lost in the noise:

I did not make any progress on this except that I found the following link from this board that mentions a "tuner 12V bug":

http://www.hdmedia-universe.com/boar...l=1#post120927

Could this be related? If I understand correctly, 12 V is the voltage used to select vertically polarized channels. Does anyone know how this bug was fixed?

After building a DEBUG version using Audioniek's build system (Enigma 2 from Jan. 28), I called dmesg and found that the kernel log was filled with errors of the kind

avl2108: ERROR: Could not get lock status!

like 10s of these messages in a row. Any ideas?

[pop1234]

@Audioniek
http://www.tbsdtv.com/download/docum...rs_v150728.zip
cxd2820r_c.c
cxd2820r_core.c
cxd2820r_priv.h
cxd2820r_t2.c <--- T2 !
cxd2820r_t.c

[Audioniek]

I have studied some things somewhat further and opened up the tuner modules in my possession. This revealed the following:

DVB-S(2): AVL2108 demodulator and STV6306 tuner (works)
DVB-T: S5H1432 demodulator and MXL5007T tuner (old, DVB-T only, tuner module number DNOD44QSH102A,work on factory firmware, not on E2, no driver available)
DVB-T(2): CXD2820R demodulator and TDA18212 tuner (new, tuner module number DNOQ44QCH106A, works on factory software, not on E2)
DVB-C: TDA10024 demodulator and MXL201 tuner (tuner module number DNQS44CPP101A, not tested by me)

If you have look at the driver code things match up regarding DVB-S(2) and DVB-C. The drivers are indeed for the chips mentioned.

With DVB-T(2) it is a completely different story. As mentioned before the DVB-T only module (...102A) is not supported, due to the missing driver. With the DVB-T2 module, there is a surprise. The driver is for the combination CDX2820R plus TDA18272, not the TDA18212 I found in my module. I looked three times with a magnifying glass, and I did not misread the number, as a 1 and a 7 can be mixed up easily in small print. The TDA18212 also exists and seems to be an earlier member of the NXP DVB-T2 compatible tuner chips of which the TDA18272, TDA18273 and the TDA18292 also are members. A quick hunt on the net did yield a brief product description, but no detailed data sheets. Most promising was the site of Antti Palosaari who has drivers for most of them

This is quite revealing, assuming that the TDA18212 and TDA18272 are not compatible with each other. Frontends are made around the demodulator, the part that is actually visible in hardware. The tuner is controlled through the demodulator, in many cases through a so called I2C gate. This explains that the CDX2820R/TDA18212 module is recognized as a frontend, because the TDA2820R is detected (and reported) correctly. The tuner however is different from that assumed by the current driver and hence: no reception and the lock errors found in the logs.

This leads to two possibilities. The first is that the chip number TDA18212 was misread as TDA18272, and the driver was built for the latter. If so, this driver could never have worked on the module I have (...106A). The second possiblity is that there are modules around with the TDA18212, but also with the TDA18272. Technically it is possible this can work if you drive the TDA18272 as a TDA18212 in case the '7' is a superset of the '1'; both ICs have the same housing and are mechanically compatible.

The fix? I have to study a driver for the TDA18212 and figure out the differences with the TDA18212 and intgrate that in the current driver. This however is time consuming and I am very busy with two other projects. To pull this off one would also need the full datasheets (not the 10 page product briefs) of the two devices.

Also see Samsung-S5H1432-demodulator-help-wanted

Regards,

Audioniek.

- - - Aktualisiert - - -

The 12V tuner bug...

is not a bug. For reasons that are beyond my comprehension, somebody mixed the LNB voltages up with the 12V output that can be found on only one SH4 receiver I know of: the ABcom IPbox 9900HD. The two (12V out and LNB power) have nothing to do with each other, but there are STV090X tuner drivers around that do strange and unjust things with the LNB power driver and the 12V output (which does not exist on almost all models), the latter can be controlled through procfs.

A long time ago I removed all 12V output related stuff from the frontend drivers, so this "bug" does not exist in my images.

In addition, in my images the entry /proc/stb/misc/12V_output is missing as the corresponding hardware is simply not there (I do not offer an image for the IPbox 9900HD as I do not have one and therefore cannot test things). E2 tests for this entry and reports that something is missing if it is not found (which is not true, the message is wrong). Regarding the previous, on SH4 this is all normal, and you should not bother about this "bug".

Regards,

Audioniek.

[mrspeccy]

I have studied some things somewhat further and opened up the tuner modules in my possession. This revealed the following:

DVB-T(2): CXD2820R demodulator and TDA18212 tuner (new, tuner module number DNOQ44QCH106A, works on factory software, not on E2)

With the DVB-T2 module, there is a surprise. The driver is for the combination CDX2820R plus TDA18272, not the TDA18212 I found in my module. I looked three times with a magnifying glass, and I did not misread the number, as a 1 and a 7 can be mixed up easily in small print. The TDA18212 also exists and seems to be an earlier member of the NXP DVB-T2 compatible tuner chips of which the TDA18272, TDA18273 and the TDA18292 also are members. A quick hunt on the net did yield a brief product description, but no detailed data sheets.

Hi Audioniek,

impressive research. I can add that the DNOQ44QCH106A indeed works with E2, including your image; however, only DVB-T channels are found while channels in the DVB-T2 standard cannot be found. With the stock firmware, both types are found. I also made some tests. With your E2 firmware (the same holds for HDMU from 2017), I tried the following

cat /proc/bus/nim_sockets
NIM Socket 0:
Type: DVB-S2
Name: Availink AVL2108 DVB-S2
Frontend_Device: 0
NIM Socket 1:
Type: DVB-T
Name: Samsung CXD2820 DVB-T/T2
Frontend_Device: 1

That is, the tuner is recognized as DVB-T only, not DVB-T2, so it seems plausible that the channel search does not even attempt to tune in to channels in DVB-T2 format. On the positive side, it seems that the tda28278 driver (mind the 7) is compatible with the tda18218 to some extent, when it comes to DVB-T type channels.

I hacked the file driver/player2/linux/drivers/media/dvb/stm/dvb/e2_proc/e2_proc_info.c to force the type of the tuner to DVB-T2 (confirmed by nim_sockets). As a result, the Enigma2 menu mentioned DVB-T/T2 in the headline of the channel search while it previously only showed DVB-T there. Unfortunately, this time no channels were found at all - neither DVB-T nor DVB-T2 ones. Maybe the tuner crashed this time due to the wrong driver being used.

One thing I find curious: the build system includes the file driver/frontends/pace7241/tda18218.c (and related headers). However, I am not sure to what extent these files are useful.

[mrspeccy]

DVB-T2 works (kind of) with TDA28218/CDX2820!


The driver for the TDA28278 finally works on my box. However, it seems that I could not tune in to DVB-T2 channels before due to a timing issue. For example, in the source file cdx2820.c there are lines such as

msleep(30); /* fixme: think on this */

and similarly for the tda18272.c. I increased these pauses to 300 milliseconds and, after recompiling the related modules, I could receive DVB-T2 channels that I selected manually in E2. It is not stable yet, though; a full channel search sometimes misses some DVB-T2 channels. I will play around with these pauses/timeouts a bit. I also think that the issue with my DVB-S2 tuner that fails to select vertically polarized channels is a timing issue. This could explain that it works on Audioniek's hardware - mine is compatible but production tolerances for the hardware could make a difference. Will look into that later.

Finally, I found this URL: https://patchwork.kernel.org/patch/3757251/ It indicates that TDA28278 is a superset of TDA28218; as far as I understand it just includes some additional modulations that E2 may not use anyway.