Prof 7301, 7500 and 8000 Tuners

[pendragon]

Being as this existing thread has evolved into something quite different than where it started, and the title isn't terribly apropos, I'm beginning another here as a follow-in in the more appropriate sub-forum.

To recap, Prof has introduced a series of PCI/USB/PCIe tuners based on the STV0903 chip. This should provide improved capabilities for S/S2 demodulation than previously available. Perhaps even more interestingly, the Linux driver has many capabilities that suggest this chip could provide a decent PC-based blindscanning function. SatelliteAV has announced they are going to be supplying this unit, which up until now was only available for direct import from Hong Kong.

I imported the 7301 (PCI) and 7500 (USB) units and have done some quick testing in Windows with both DVB Dream and DVB Viewer. While I haven't run a stressful set of signals through yet, on the surface both the 7301 and 7500 are doing better on high rate signals than my other PC devices (DVB World 2104 USB and TechnoTrend S2-3200 PCI). I will continue working with these on Windows, but I'm switching my emphasis towards Linux in search of the holy grail of PC-based blindscanning.

Many thanks to tester239, who in the previously mentioned thread was the one originally making the connections between the blindscanning capabilities of the STV0903 and also the inclusion of this chip in the new Prof tuners.

 

[pendragon]

Hmm. I suppose I was wrong about support Looks like it's got a cx23883 PCI interface chip ("broadcast decoder"). Need to add a card def in the cx88 drivers. Do you know if we can just use some existing cx23883 firmware from another card, or are they all card specific? Fun times ahead! I'm at a loss right now, since I don't have hardware to test with yet. I've just been trying to more familiarize myself with the DVB-API v5.

That's basically the problem I ran into. However as the Prof 7300 also employs the CX23883 and there appears to be driver support for it, my plan was to see how far I can get in comparing the two.

 

[tester239]

The cx2388x based tuner cards of any flavor are pretty simple to add base support for since the cx88 driver is fairly mature/stable and there are quite a few cards (not just DVB-S) cards that use that interface chip. In the case of the Prof 7301, the integrated tuner/demodulator module has a chipset in it that also has driver support, so THEORETICALLY one could just write a card definition in cx88-cards.c and map the inputs and GPIO stuff correctly (using windows driver inf data and regspy). In practice, every card typically has their own implementation quirks, but we're in really good shape to get this sucker working.


pendragon: since you have a card already, I encourage you to look at cx88-cards.c and see how the definitions are done (if you haven't already) and see if you can obtain the necessary information to add it (PCI ids, gpio info, what inputs are used and connected, etc etc).

EDIT: in fact, if you look at the output of dmesg, you should see the kernel yelling at you because it knows that you have a cx23883 but no card definition matches, so the driver never attaches.

 

[zamar23]

Which driver never attaches? Linux? How then this card works? Keep in mind, a lot of people of various background read this forum, hence more clear info is always welcome.

 

[pendragon]

At the moment we're talking Linux. The Windows drivers seem to work fine with the Prof devices. Yesterday afternoon I did look at cx88-cards.c for the 7300 (older unit) definitions and they're there, although they are rather generic. This weekend is supposed to be pretty good weather-wise, so I hope to be working on my toroid dishes and master switch, but if I get a little time I plan to start pushing the 7301 farther along.

 

[zamar23]

pendragon

Given your contribution to this forum , attached below are Prof 7301 Linux Driver files by Igor Liplianin, which are still work in progress, but working well and close to release version, and you're invited to contribute. You have 2 options:

a) Download and install Prof 7301 unofficial Linux driver of July 2009: prof7301_ready_to_install.tar.gz
Instructions are available to install Prof 7301 as a 2-nd sat tuner card, but need translation effort;

OR

b) Download and install DVB-S2 Linux Drivers by Igor Liplianin
Apply 2 patches of Oct. 2009 attached. There are still minor issue with showing correct signal levels, etc, but everything works.

 

[zamar23]

pendragon

Counting on even bigger contribution from your side , below is given today's maestro Igor Liplianin's reply to your issue with 2 identical DVB S2 cards in Linux:

"I suggest to open 4 consoles, start szap in two, and mplayer or ffplay in the other two as show below, and see what happen..."

szap -a1 mychannel1 -p
ffplay - </dev/dvb/adapter1/dvr0

szap -a0 mychannel2 -p
ffplay - </dev/dvb/adapter0/dvr0

Not sure, if he tried similar setup, but we can continue this dialog, and with his help the problem can be resolved for sure, as he wrote most of sat cards current Linux drivers himself and appears to be key expert on the issue. Usually he asks for detail system reports.

 

[pendragon]

pendragon

Given your contribution to this forum , attached below are Prof 7301 Linux Driver files by Igor Liplianin, which are still work in progress, but working well and close to release version, and you're invited to contribute. You have 2 options:

Big thanks! I had looked around but didn't find anything. Sunday's weather doesn't look as good, now, so maybe that's the day for inside play.

pendragon

Counting on even bigger contribution from your side , below is given today's maestro Igor Liplianin's reply to your issue with 2 identical DVB S2 cards in Linux:

"I suggest to open 4 consoles, start szap in two, and mplayer or ffplay in the other two as show below, and see what happen..."

Actually I have no problem running two or more 2104s on Linux, and that is the long term destination for those units. At the moment the problem is having two Sage network encoders, for which Windows is the only current option. I solved my problem with a DVB World 2104 (USB) and a TechnoTrend S2-3200 (PCI). mkanet wants to use two identical units for the same purpose, so he has few if any proven options. I can get the two Prof units I have to work together on Windows, but one is USB and the other PCI. I'm hoping to try two of the same type soon.

Someday I'll write a Sage network encoder for Linux and that will let me choose just about anything, but it will have to be able to control my 44x8 switch matrix, have the smarts to run motors, select the best option from multiple dishes and resolve conflicts. It's not at the top of the list for now.

Thanks again for your research!

 

[zamar23]

I have no problem running two or more 2104s on Linux

I can tell you, these drivers are very hard to find, mainly because Prof and TT don't want to pay Igor for his contribution, and due to other marketing considerations.

Do you use the same method to run two 2104s on Linux as Igor suggested? I don't have 2 identical cards at the moment to try, but will be glad to provide him some feedback. Did you try the method he suggested? If you want to improve Prof 7301 drivers in any way, you may want to communicate with Igor in basic English directly here.

 

[pendragon]

I can tell you, these drivers are very hard to find, mainly because Prof and TT don't want to pay Igor for his contribution, and due to other marketing considerations.

That's the unfortunate side of open source. I've contributed a lot over many, many years (more like decades) and still am waiting for the first penny.

Do you use the same method to run two 2104s on Linux as Igor suggested? I don't have 2 identical cards at the moment to try, but will be glad to provide him some feedback. Did you try the method he suggested? If you want to improve Prof 7301 drivers in any way, you may want to communicate with Igor in basic English directly here.

I tried this several months ago. There's no magic. You plug in two 2104s, they get assigned as different devices and you tell the app which device you want it to use. I hacked together a quick and dirty app that I'm experimenting with and now use a lot for alignment, but it will need to be rewritten before it will useful for anyone else.

 

[zamar23]

pendragon

I think you may want to contribute with Signal scales calibration and some other outstanding issues, since Prof and TT haven't sent Igor the cards samples, while other people having the cards at hand also contributed some code to the drivers above.

P.S. It was reported on another forum that Linux driver for TBS PCI-E DVB-S2 6920 and PCI DVB-S2 8920 sat cards works well with Prof 8000 PCI-E Card tuner.

 

[pendragon]

I was chasing feeds today instead of hanging on the roof with my toroids or messing with drivers. This is only an early data point, but I wasn't particularly impressed with the Prof 7500 (USB) unit vs. a DVB World 2104 (USB) on Windows. The link margin was a little thin and we had very high winds. The 2104 did the better job holding lock out of a splitter feeding both. However the only fair comparison in my mind would be with Linux drivers tweaked for best signal performance, assuming it makes a difference.

 

[zamar23]

Some problems with DVB-S2 feeds locking seems to be a known issue, said to be caused by pure MS filters performance in the BDA driver graph in Windows versions earlier than Windows 7. Its suggested to use a WDM driver instead. There may be other issues involved not identified yet due to limited so far customer base for this new device. Prof recommends also to use AltDVB or DVBDream applications with these devices as optimal choice. It was reported also that connecting a high quality DiSEQC Switch improves reception by shifting noise pattern created by 12/18V impulse voltage converter due to higher load. Some mods were also suggested to improve reception. Do you want to try?

 

[pendragon]

Yesterday was plain DVB-S, but at a high rate. According to sources there were issues with the uplink that on my spectrum analyzer were very apparent. That wasted most of my link margin, and the high winds were more than my Ku dishes could tolerate. Still with the carriers jumping up and down, the DVB World tuners did an excellent job holding lock. I only had one Prof 7500 running as a comparison off the same splitter as one of the DVB World units, and while it did well, it lost lock more frequently.

I've used both DVB Dream and DVB Viewer on Windows with both Prof units and they work about the same. Both apps have pros and cons, but yesterday I had to use DVB Viewer because it can do something I needed that DVB Dream cannot. I had just run a 24 hour test on another high rate signal with the Prof/2104 units and both were perfect (no transport stream hiccups). But that signal had an easy CNR.

I read about the mods a few days ago on some Russian forum, which looks to be the same link you provided. None of this applies to yesterday because the Prof unit was 'slaved' off the splitter through a DC blocker. I was somewhat dubious about the discussions, and it may have little to do with my system where the receivers at most power some switches and not any LNBs/motors. That means my receivers' current draw is always low. Nevertheless I set up a quick test where I powered a chain of the same types of switches I use all the way up to a LNB. I then compared a couple of signals on the spectrum analyzer using the receiver power and then using an external device to supply the switch/LNB power. It didn't make any difference so my links went into the circular file.

I haven't looked at the Linux drivers to see if there are some tweaking parameters that might make the Prof units perform better under adverse conditions. AGC loop time constants come quickly to mind. This was a particularly stressful test under conditions I rarely experience. Normally I have booming signals, carefully matched levels and a well equalized signal chain. 'Deaf' tuners seem to perform well for me.

Still I don't subscribe to fan boy clubs that declare any technology more than six months old to be garbage. The CX24116s in the DVB World units perform very well for me and newer units will have to establish their reputation by direct comparison with real signals, not by spec sheets and not by 14-year-olds posting on forums. At the moment the score on acid signal tests is only DVB World 1 and Prof 0. This will become clearer in time with more tests and more Prof units to compare. No two of my four DVB World devices perform identically and none clearly performs better than the others under every condition. I select the serial number most appropriate for a particular application, and I wouldn't be surprised if the same will apply to the Prof units.

The Prof BDA interface for DVB Dream has a ridiculous bug in it that prevents using at least DVB World units on the same Windows machine, but other than that I am very impressed with the stability and functionality of the Prof Windows software. It was able to command all the pieces of my complex switching chain out of the box, something the DVB World drivers are finally achieving after a long wait for fixes. I can't say that my TechnoTrend S2-3200 is even there, yet, and it isn't clear that it ever will be.

 

[tester239]

Getting closer... custom kernel with patches backported to 2.6.31.5

[    8.964510] cx88[2]/2: subsystem: b034:3034, board: Prof 7301 DVB-S/S2 [card=83]
[    8.964517] cx88[2]/2: cx2388x based DVB/ATSC card
[    8.964521] cx8802_alloc_frontends() allocating 1 frontend(s)
[    8.986364] <6>stv0900_init_internal
[    8.986372] <6>stv0900_init_internal: Create New Internal Structure!
[    9.171086] <6>stv0900_st_dvbs2_single
[    9.176178] <6>stv0900_set_ts_parallel_serial
[    9.193205] <6>stv0900_set_mclk: Mclk set to 135000000, Quartz = 27000000
[    9.197570] <6>stv0900_get_mclk_freq: Calculated Mclk = 135000000
[    9.221839] <6>stv0900_get_mclk_freq: Calculated Mclk = 135000000
[    9.221843] stv0900_attach: Attaching STV0900 demodulator(0)
[    9.259813] stb6100_attach: Attaching STB6100
[    9.261494] DVB: registering new adapter (cx88[2])
[    9.261512] DVB: registering adapter 2 frontend 0 (STV0900 frontend)...

Having some tuning issues right now.. will advise. I have various engagements today so I may not get back to this until Monday.

 

[zamar23]

CX24116 was introduced in 2004, and went through multiple revisions, fixing numerous problems widely reported in Europe. Glad if it was finally sufficiently improved. However last year CX started to advertize its replacement CX24118A, which already went through a couple of revisions. Jury is still out on the new one for most STB device makers. On the well known forum, you mentioned, that discussion and a few others were btw several electronics engineers, not teens, but they usually tend to criticize everything. Many problems with CX chip were of course attributable to traditional deficiencies of CN consumer sat cards and PC power supplies in general, as well as lower SNR in self-installed systems. Responding to complains, Prof RU has developed a new sat tuner series, based on STMicro STV0903 chip you're testing now, reported to be professionally designed, lock and switch channels much faster and better (S2 and HBR in particular, provided sufficient card quality within targeted cost range), and new driver series.The cards are met with great interest by consumers. Its interesting, what's your opinion about this new chip and the card you are testing, given high SNR advantage of your knowledge based reception setup?

 

[pendragon]

I wasn't referring to that forum as teenagers. They were knowledgeable to a point, and they understood a lot about electronics, but I wasn't all that impressed. It seemed to me an ill-defined search for a solution to an ill-defined problem. Not a lot of hard facts, well designed measurements and clear cause-effect associations. I wasted too much time proving to myself that my intuition was correct. My one error of omission was not putting the Prof tuners on dummy IF and DC loads and looking at the L-band noise spectrum on the SA. I will try to do that sometime soon.

In terms of the CX24116, it is an older chip, but I'm not aware of any dramatic revolutions in the world of DSP since that time (that's my specialty). In terms of PC tuners released five years ago vs. brand new ones, I've only seen a limited number of specifications that can be compared. It seemed the most obvious improvement was in NF, and I can believe that one. However in a well designed sat reception/distribution system, that specification often has little to do with performance. I want to see if there are real improvements that I can quantify on the bench and down the link. Shiny new covers don't do a lot for me unless there's something worthwhile under the hood.

Part of the problem I'm having is I've spent so much time tuning, aligning and tweaking my dishes, switches and distribution system that for the most part there aren't many feeds that are on the edge for me. I have a Pansat 9200HD which has a reputation for being deaf on DVB-S. Works fine for me. The TechnoTrend S2-3200, DVB World 2104 and Prof 7301/7500 seem to get nearly everything I throw at them. I initially seduced myself into believing the Prof units were locking better than the older tuners, but I'm not sure any more. If I pulled out my distro amps, ignored the frequency response of the system paths, put in cheap, unpowered switches and slammed my dishes with a sledgehammer, it's very possible the newer units would provide a better a better lock and a lower BER.

High winds and a weak uplink had me licking my chops. But the 7500 was undisputedly worse than the 2104. It's not the end of the world because this could be a pathological case, but it was slightly disturbing. But this was my best 2104 for locking high rate, low CNR signals and the 7500 was just a random unit from Hong Kong. I am slowly bringing my toroids on line with improvements that have been long in coming. It's possible that the reduced CNR off these dishes may be good diagnostic tools for evaluating new vs. old PC tuners.

So far I've seen negligible improvements in my real life world, but I'm still happy because the Profs perform fine and the drivers work. I would expect if I hit them with a 45000 S2 SR, they would be undoubtedly better than the 2104s. I'm just not running to much of that on the sats I can see. I'm hoping to have a couple of the 7500s and a couple of the 7301s soon so I can get a rough idea of what the unit-to-unit variations are. If the Profs can blindscan, this project will have been worth every penny and minute. Even if they don't, they will give me options I didn't have before.

 

[zamar23]

Of course you're right, the forum talk was at hobbyists level, and they lack testing equipment and advanced reception systems. It almost sounds, you need an average Joe's cheapy Sat reception system separately setup to do new tuners comparison testing when ordered. The second condition - local HBR signal availability - might come faster that we expect. Its unfortunate, true FTA in N/A seems to be mostly hobby for retirees due to luck of content, but I'm sure you managed to involve your own teens in this challenging hobby.

 

[zamar23]

Attached is a table of chipsets used in some popular DVB-S2 sat cards.

 

[enb141]

Hey great bable of chipsets, now about the prof revolution 8000 PCI e and the TT-Budget S2-1600 one is PCIe and the other one is old PCI, here is the link of the specs for the STB6100


now what's the difference between STB6100 and STV6110A?