DISEqC MultiSwitch failover switch

[David Jashi]

Hello, everyone.

I have a task of assembling receiving headend in quite a remote location with limited access for most of the year. That is - it should not be serviceable. There are 4 directions (13E, 36E, 42E, 46E), which I will connect with fiber LNB/IRS (Triax, or Global Invacom, whichever will be available at the moment), 17x32 multiswitch and whole bunch of receivers.

Making everything in N+1 or double quantities is the best way out, but not always possible. Antennae (and, therefore, LNBs) are out of question - there is simply no place for 8 antennae there, I'm struggling to fit 4 of a descent size (1.8 m with heating in lower part). Then I'm putting 1x2 optical splitter to connect 2x4 fiber IRS, whom I then connect to two 17x32 Polytron Mutiswitches

So far so good. Now that's where it all goes haywire. There are 32 inputs on 16 receiver cards, and I need to connect each input to both multiswitches. 1x2 DISEqC is out of question - receivers must control multiswitches with 14/18V+0/22Khz+DISEqC 2.0 and multiswitches have committed DISEqC, so they won't ignore commands, sent ot switches before them.

What I am looking for is technically something like manual switch, but controlled electronically, like relay, so that I could send a signal from Arduino's GPIO and switch it back and forth.

Is there any sane way of doing it other than using solenoids to move the damn switch? I really don't want anything mechanically moving inside my headend.

 

[rolfw]

Hi, welcome to the forum. The setup you propose is outside the scope of this forum, it is really for domestic systems, as per the title.

However, on reading your post, I can't quite understand why a receiver would require access to both multiswitches, as the outputs should be identical, with four satellites available on both.

 

[David Jashi]

Sorry, did not notice word "domestic". If you can move this thread to more appropriate branch, it will be very appreciated.

All receivers must have access to both multiswitches for redundancy. In case the first one dies, they should be switched to second.

 

[rolfw]

Hi David, there isn't a section for commercial systems, as we are primarily a hobby forum, even though many of our members may be involved in the commercial aspects.

I'm not sure whether there is actually anything made to do what you require remotely and what the chances are of a mutltiswitch failing rather than a quattro GTU, or perhaps even fibre MDU.

You may need to use manual rocker type mechanical switches with some sort of solenoid as you mention. It's a shame that smart priority switches couldn't be adapted, as they were primarily designed for switching between receivers rather than LNBs.

 

[David Jashi]

Thanks for your answer. It's not commercial system, more like community one, but nevertheless you are right - quite off-topic here.

I'm browsing through DISEqC 2.0 specs right now to find a combination of commands, that helps me in this matter, but as you mentioned, I might resort to solenoid.

There is another thing I should try (very carefully) - what if I will keep redundant multiswitch powered off until needed (and powering it on only after first one is shut down) and receivers connected to both of them via dumb T-splitters

 

[PaulR]

Would such a switch not introduce a further source for potential failure? And given the number of different cables it would have to switch even a partial failure would be annoying at best.

 

[David Jashi]

Well, that's why I want to do without mechanical parts. I'm taking apart old 1x2 DISEqC switch tomorrow to see, if I can do something with modification to standard one.

 

[rolfw]

There is another thing I should try (very carefully) - what if I will keep redundant multiswitch powered off until needed (and powering it on only after first one is shut down) and receivers connected to both of them via dumb T-splitters

On that basis, you may be better off looking at standard splitters like this one

http://www.satelliteandaerialsupplies.com/product/?s=konig-2-way-f-type-splitter-5-2400mhz-vhfuhfif

 

[David Jashi]

On that basis, you may be better off looking at standard splitters like this one

Will it pass signal in opposite direction? It's not "1 in/2 out" in my case, it's "1 of 2 possible in/1 out".

 

[rolfw]

Sorry, got that wrong, the bi-direction dc pass is only on the RF splitter combiners, you could probably get some IF units made or adapt to suit.

 

[David Jashi]

Thanks for a tip. Those are 1GHz max, though. So its more like something like this, with 5-2000MHz screening factor.

But first of all - DISEqC autopsy on weekend.

 

[rolfw]

Good luck. Seen some L band splitter combiners which may suit. ETL COM04L1P-2504 4-Way Passive L-Band Splitter/Combiner

 

[a33]

What I am looking for is technically something like manual switch, but controlled electronically, like relay, so that I could send a signal from Arduino's GPIO and switch it back and forth.

Is there any sane way of doing it other than using solenoids to move the damn switch? I really don't want anything mechanically moving inside my headend.

Maybe I am missing something, but why not use UNcommitted diseqc switches (diseqc 1.1)? The multiswitches would not care..., as they function on diseqc 1.0 commands.

Or else: a 0/12Volt switch would be possible, I guess. I guess it wouldn't be difficult to "command" it with a computer?

If you want to make something like that yourself, google on [ Frankenswitch diseqc ] and find the posts of a person called mlord. His Frankenswitch v3 is beautiful . He modded a 4/1 switch, though.

Greetz,
A33

 

[David Jashi]

Maybe I am missing something, but why not use UNcommitted diseqc switches (diseqc 1.1)? The multiswitches would not care..., as they function on diseqc 1.0 commands.

Or else: a 0/12Volt switch would be possible, I guess. I guess it wouldn't be difficult to "command" it with a computer?

If you want to make something like that yourself, google on [ Frankenswitch diseqc ] and find the posts of a person called mlord. His Frankenswitch v3 is beautiful . He modded a 4/1 switch, though.

Greetz,
A33

Thanks for a Frankenswitch tip, A33, I will look into it.

About uncommitted ones - all multiswitches I ever saw in action or heard of are committed ones, so that's not quite an option.

As for 0/12V switch - thought of that, but how will I stop it from passing further? Remember, the switch in question is a first one in chain.

 

[hairybadger]

Does the installation have network access? Like @a33 says something like a Raspberry Pi could be used for remote control.

I'm a bit unsure about the splitters - wouldn't they introduce signal loss? If you're searching for every last dB (you mention a heater so I'm assuming that snow will be present at least some of the year) then that could be an issue.

 

[David Jashi]

If you want to make something like that yourself, google on [ Frankenswitch diseqc ] and find the posts of a person called mlord. His Frankenswitch v3 is beautiful . He modded a 4/1 switch, though.

Thanks again, I got a proof of concept myself (frying 2 DISEqC switches in process). Looks like I will go by 4th or 5th version, as I want it to be controlled by 5V GPIO on Arduino with an expander like PCA9555.

 

[David Jashi]

Does the installation have network access? Like @a33 says something like a Raspberry Pi could be used for remote control.

Well, its industrial grade headend for a small (couple of hundreds of end-users) community of my friends in remote mountainous location and it contains 10 small NUC-sized (mostly) computers for receiving, descrambling, transcoding and one 19" 1 RU for recordings. I think I can handle an Arduino or two for switching purposes. And switch-over is done by software, so there is no need for Internet for this purpose.

I'm a bit unsure about the splitters - wouldn't they introduce signal loss? If you're searching for every last dB (you mention a heater so I'm assuming that snow will be present at least some of the year) then that could be an issue.

Frankenswitch mentioned by A33 (God bless him for that) is based on standard DISEqC, so I'll simply use a bit expensive ones to have less attenuation (14 dB on a cheapest one). And as a matter of fact, I intentionally use 1.8m dishes and fiber to have signal budget for things like this. Heaters will be installed due to the fact, that when it snows there - it covers dishes and shuts off signal entirely, not just stealing a couple of decibels here and there.

 

[a33]

Do I, or don't I, understand your problem correctly?:

For continuity reasons you have doubled the parts from splitter (at dish-end) till (inclusive) multiswitch. You need the receivers to be able to choose between the two multiswitches, and are looking for a type of switch that can do exactly that.

Is this problem-description right or wrong?

Second question: do you want ONE switch for the whole community, or do you want separate switches for all individual participants? Or is that indifferent?


A diseqc1.1-switch as I described it comes before the two multiswitches. No problem for the multiswitch, to stay switching on committed diseqc commands.

A 0/12volt switch needs a separate command-cable (so the 0/12 volt doesn't go through the coax cable!). By the way, as I read somewhere, the 12volts aren't critical; it should work on 6volts, so maybe also on 5 volts????
The need for a separate command cable also applies to the frankenswitch.

Greetz,
A33

PS Did you also read this google link of the Frankenswitch?: Switching antennas automatically, using modified DisEQc hardware - Canadian TV, Computing and Home Theatre Forums
Both threads (this one, and the one you mention) are very informative!

 

[David Jashi]

Do I, or don't I, understand your problem correctly?:

For continuity reasons you have doubled the parts from splitter (at dish-end) till (inclusive) multiswitch. You need the receivers to be able to choose between the two multiswitches, and are looking for a type of switch that can do exactly that.

Is this problem-description right or wrong?

Yes, and Frankenswitch v4/v5 in conjunction with Arduino will do just fine.

Second question: do you want ONE switch for the whole community, or do you want separate switches for all individual participants? Or is that indifferent?

RF is present only in headend, clients are fed with HLS streams. There are 16 (14+2 spare) TBS6910 on multiswitch outputs and 8 (4+4 spare) TBS6909 cards on multiswitch's cascade (those cards have internal switches, so they need direct quattro input) for FTA channels.

A diseqc1.1-switch as I described it comes before the two multiswitches. No problem for the multiswitch, to stay switching on committed diseqc commands.

I don't really need to switch BEFORE multiswitches (UPDATE: "before"="dish side", "after"="receiver side"), as they are constantly fed from IRS. I will need to switch outputs. All 32 of them.

A 0/12volt switch needs a separate command-cable (so the 0/12 volt doesn't go through the coax cable!). By the way, as I read somewhere, the 12volts aren't critical; it should work on 6volts, so maybe also on 5 volts????
The need fo a separate command cable also applies to the frankenswitch.

Yes, of course, I will drop control voltage to 5V, as described here to comply with GPIO voltage and will control switching process with Arduino or similar device. I will still have to choose, how to trigger switchover, whether by controlling multiswitches powerstate with Arduino's inputs or programatically analyzing the state of carrier - if signal lost on all receivers, them multiswitch died, switchover. I think I will do both.

PS Did you also read this google link of the Frankenswitch?: Switching antennas automatically, using modified DisEQc hardware - Canadian TV, Computing and Home Theatre Forums
Both threads (this one, and the one you mention) are very informative!

Nice one too, but I prefer to control it from my application.

 

[a33]

A diseqc1.1-switch as I described it comes before the two multiswitches. No problem for the multiswitch, to stay switching on committed diseqc commands.

I don't really need to switch BEFORE multiswitches (UPDATE: "before"="dish side", "after"="receiver side"), as they are constantly fed from IRS. I will need to switch outputs. All 32 of them.

Ah, sorry!
I meant 'before' as receiver side/output side (the side from where you control the switch). You are correct that that is called 'after', conventionnally ; as convention follows satellite-signal to receiver....

Diseqc 1.1 switches are normally used OUTPUTside of a diseqc 1.0 switch, and would therefore also be applicable in your case (but probably not for the TBS6909? Not sure how they manage diseqc control), assuming the receivers (TBS6910) can give diseqc 1.1 commands (that could even be easier than building some kind of controler?)

But it seems you have chosen your preferred way of solving, it seems. Glad to have been able to help .

greetz,
A33