239 replies to this topic

[Robinson]

...... but once it arrives it still shows 0% signal strength for half a minute or even more.

Does this also happen when the box has just booted up?
I ask because I have a box that sometimes shows a similar behaviour; when I reboot the box the problem has gone (for some time).

I will check this evening.

 

 

I'll check these also, for the moment when I see the following video and what the potential of the tuner is I wanna punch my screen..

Indeed. Totally different quality of drivers.

[Robinson]

 

...... but once it arrives it still shows 0% signal strength for half a minute or even more.

Does this also happen when the box has just booted up?
I ask because I have a box that sometimes shows a similar behaviour; when I reboot the box the problem has gone (for some time).

 

 

Yes, the same happens right after reboot.

It took my rotor 32 seconds to travel 40 degrees and then it took 55 seconds to pick up the signal and show the picture.

In total: almost 90 seconds.

[MCelliotG]

In this case it looks the voltages aren't up to par. It's an extension to the behaviour I inspected. I should maybe try a different power feed but I don't think it's the case and neither it should be. Perhaps even a change in the voltage parameters in the rotor menu or an activation of the sequence repeat function won't help either.

I also noticed from several DM900/920 videos that the signal is also measured in smaller increments, not as small as I'd like but vastly better than HD51. There are values that end with .25 plus I don't see that issue where the tuner first locks a much less signal (8.7, 9.5db) and then goes to the maximum when tuned to a channel or the 0% indication. I'm pretty sure that the percentages stop at 100% (unimportant in our case since we want SNRdB) and not continue after like 110% or so...

 

But... a good source (on OpenATV) was promised blindscan for HD51. I know we are full of promises concerning blindscan on E2 but at least Mutant is not Octagon for all it matters....

[Robinson]

 

I quickly monitored the behaviour when moving a dish with HD51.

- When the satellites are adjacent (less than 3 degrees) the rotor works fine, no problem. This goes both ways.

- When the satellites are further apart this is what happens. When you first select the channel the dish does move, but after a second it stops moving. Like the power gets cut off. When you have a normal behaviour the rotor is first sent a signal with a certain voltage to move, then a second one with higher voltage to move faster. It looks like the driver will not send the second signal, this means the dish will move up to three degrees and then needs a new command (this happens when you change channel, sometimes twice) to resume the motion and reach its final position.

This is where the bug lies, the receiver does not send the correct voltages sequences to keep the rotor moving until it reaches its desired destination.

- For slow lock, I did not really notice it yet. Once the dish arrives in the satellite it locks right away.

 

For signal strength and tuner status reporting - I can confirm this 100%, unfortunately.

However, as regards satellite lock, the behaviour here is different although adjacent satellites do indeed work correctly. Here (Moteck SG-2100), the dish does not stop after a second or so. It keeps moving as fas as the target satellite but once it arrives it still shows 0% signal strength for half a minute or even more. I have not measured the time and I'm not sure if it is always the same.

By the way, can somebody please confirm all the 3 bugs (rotor lock, tuner status reporting, and signal strength) can also be observed on other images, such as OpenATV?

I can hardly believe in Blindscan promise if they cannot deal with the 3 bugs mentioned.

[Robinson]

When you have a normal behaviour the rotor is first sent a signal with a certain voltage to move, then a second one with higher voltage to move faster. It looks like the driver will not send the second signal, this means the dish will move up to three degrees and then needs a new command (this happens when you change channel, sometimes twice) to resume the motion and reach its final position.

It looks like in my case the second signal with more voltage is sent correctly.

 

 

- For signal status. Yes, it's quite annoying. The signal is measured in very large increments. There is no logic to have 15.2db and the next step to be 15.7db, especially if you try to finetune a position or even use your STB to adjust your arc.Compared to what the BCM73xx tuner that OSmini has, that is so precise and in such small increments it's ages behind. For this issue though, I cannot tell how Silabs behaves in general. Perhaps all Silabs tuners measure the signal this way.

 

Basically, I zapped from some channels today and I saw the following dB values: 12.  7   - 13. 0  - 13.2 - 13.5  - 14.0 - 14.2 - 14.5 - 14.7 - 15.0 - 15.7 - 16.0  

This clearly shows that dB value is rounded to only the following digits after the point: 0, 2, 5, 7. I guess noone observes any other values? This is very poor for those who want to fine-tune their dish.

Also, 1dB is equal to 5,0-5.1%.
 

[MCelliotG]

It's confirmed that all Silabs demods measure in 0.25db increments. Dreambox drivers are a absolutely correct displaying the right values but they are the only ones (except Octagon). The driver used in Mutant and Zgemma is simply atrocious. 0.25db is not bad for accuracy but the current status in Mutant is. This issue is even worse in Multistreams. I've seen the SNR jump from 11.2 to 14db! Let's see when blindscan gets released if the signal problem gets fixed too!

Στάλθηκε από το LG-D855 μου χρησιμοποιώντας Tapatalk

Edited by MCelliotG, 4 February 2018 - 22:36.

[Robinson]

Let's see when blindscan gets released if the signal problem gets fixed too!

I don't really believe in blindscan but it would be nice to be surprised.

Tthe lock issue has been around for a year or so and nothing has been done so far. Perhaps because they are so busy working on blindscan.

[MCelliotG]

What was a bit surprising for me is that this driver problem, along with the inexcusable non-real time tuner report (still shows 10600 and 9750) is present in all receivers using the Si2166 tuners, except Dreambox and Octagon.

Plus the values are hardware-rounded. I use a modified FrontentInfo that displays two integers (15.20 for example) and see no 0.25 increments at all! What a shame!

Edited by MCelliotG, 5 February 2018 - 11:13.

[Robinson]

So it looks like the drivers for the cheaper boxes (Mutant, Zgemma) are deliberately stripped of their capabilities.

[Rob van der Does]

So it looks like the drivers for the cheaper boxes (Mutant, Zgemma) are deliberately stripped of their capabilities.

I don't think that's how it works.

An STB-manufacturer only gets a basic set of drivers from the chipset manufacturer. It's up to the STB-manufacturer to enhance the drivers to an optimum performance. Also e.g. the activation of codecs depends on the latter; and will be influenced by the amount of license-fees he's willing to pay.

So the above statement would only be true, if the same manufacturer is involved; in that case I could imagine the budget box drivers to be 'stripped'.

[WTE]

.

But... a good source (on OpenATV) was promised blindscan for HD51. I know we are full of promises concerning blindscan on E2 but at least Mutant is not Octagon for all it matters....

 

Source is MCelliotG ? I didn't read and heard any promises about blindscan.

[WTE]

So it looks like the drivers for the cheaper boxes (Mutant, Zgemma) are deliberately stripped of their capabilities.

 

Don't believe everything you read. Mutant use real values and doesn't recount the value which actually most manufactures do to get "fake" values so it looks like this tuners work better.

 

The Mutant use for LNB regulator from manufacture TI which has one of few good control for overcurrent when you use a motor. This kick in when the motor pulls more current then the maximum value. The problem comes by some customers when motor is damaged and so pulls more current. At moment they talk with TI how to handle with this safety check and find a good and safety solution.

[Robinson]

As regards motor, the main issue, reported on this forum and others, is the long time we have to wait for lock after arriving at a destination satellite. It is about 60 seconds. I have had motorised systems for over 20 years but none of the receivers has ever shown a delayed lock.

[MCelliotG]

.

But... a good source (on OpenATV) was promised blindscan for HD51. I know we are full of promises concerning blindscan on E2 but at least Mutant is not Octagon for all it matters....

Source is MCelliotG ? I didn't read and heard any promises about blindscan.

No source is not me, the last thing in my vortex is to spread fake news. Besides I only know what I read myself, I can't speak of others.

Στάλθηκε από το LG-D855 μου χρησιμοποιώντας Tapatalk

[humaxboy]

 

Let's see when blindscan gets released if the signal problem gets fixed too!

I don't really believe in blindscan but it would be nice to be surprised.

Tthe lock issue has been around for a year or so and nothing has been done so far. Perhaps because they are so busy working on blindscan.

 

 

This would never happens,  I am sure.

 

Just to inform you guys that the same bug exists on Octagon SF4008, which uses similar tuner. But for me is not so critical because it is happening only rarely...

 

Do you mean the lock issue when using a rotor? So on SF4008 you sometimes get a lock immediately and sometimes have to wait up to 1 minute after dish arriving at the target satellite?

Or do you mean you do not move very your dish much between distant satellites so there is not a lot of opportunity to experience the bug?

 

One positive information : with drivers (linuxtv.ko) from 03.2017 moving of the dish with Octagon SF4008 works perfect. I am using it for several days and I havent had any problem since then...
 

[Robinson]

humaxboy, but I don't think those drivers can be used with HD51.

[humaxboy]

humaxboy, but I don't think those drivers can be used with HD51.

 

Almost for sure not...

[Robinson]

I think driver update was Dec 2016. May 2017 was when it was first reported. OpenViX images before this driver change don't have the bug. Manufacturer was notified in July 2017.

 

So theoretically an image from 2016 should remedy the rotor lock issue? Although it might re-introduce some other bugs.

Are OpenPLi images available for HD51, dated October/November 2016?

[twol]

I think driver update was Dec 2016. May 2017 was when it was first reported. OpenViX images before this driver change don't have the bug. Manufacturer was notified in July 2017.

 
So theoretically an image from 2016 should remedy the rotor lock issue? Although it might re-introduce some other bugs.
Are OpenPLi images available for HD51, dated October/November 2016?

..... but this was before the s2x tuner was released, so an issue in terms of s2x drivers

[Robinson]

Do you mean the old drivers would not work with S2X tuners? If so, you are probably right.

Or do you mean that the issue concerns S2X tuners only? If so, this is probably not the case.