v1.2d test 1 - Diseqc switch tests needed

Postby **rel** » Fri Sep 22, 2006 10:16 am

Needs following tests:

- Diseqc switch functionality (If it is working fine?)

- "Auto Retune" / "Force to re-send diseqc cmds" options in preferences/performance (If they do their jobs? their jobs are just what their captions say)

- Initial Diseqc settings (If LNB Type and Uncommitted switch settings retain after the first restart following the installation ?)

Download

You can ignore other bugs. (You will see some unnecessarry labels specific to test versions and there will be something else in the place of CPU usage on the status bar)

If you see any problem, please reply under this forum thread. I will probably release it tomorrow. Thanks.

genpix · Postby **genpix** » Fri Sep 22, 2006 11:29 am

Diseqc:

looks like there are two different behaviour styles.

1) RAW diseqc field is EMPTY
DD does send diseqc command (default "E0 10 38 Fx" - depends on the input selected) the number of times set by the Diseqc count in the preferences. If in addition, uncommited switch is selected, DD sends uncommited command as well.
For instance, I have "Always force to resend Diseqc command = ON" and count = 2 :

Code: Select all

[3240] DISEqC: 4 bytes - E0 10 39 F0 

[3240] DISEqC: 4 bytes - E0 10 39 F0 

[3240] DISEqC: 4 bytes - E0 10 38 F2 

[3240] DISEqC: 4 bytes - E0 10 38 F2

But, I would expect them interleaved. Don't you think so?

Though, when I have "Always force to resend Diseqc command = OFF", DD does not ever send any Diseqc commands. Even when I change sat.
I guess, you have to force to send at least single Diseqc command, when different Diseqc input is selected (no matter what setting in the preferences).

2) RAW diseqc field is NOT empty
Uncommited switch settings are ignored (uncommitted command gets never send)
"Always force to resend Diseqc command" is ignored
number of resends is ignored (only one command is send)

and threre two sub-sets again:
2a) "Always retune" = OFF
error is still the same
tune is always sent when V/L polarity is selected
NO retune is sent if select different channel within H/R tp
Diseqc command is sent ONLY when tune command is send.

2b) "Always retune" = ON
Tune is always send.
within H/R tp: Diseqc is sent only once (when I switch from different TP), then does not get send if I stay within TP.
within V/L tp: Diseqc is always sent, no matter what.

2a and 2b are closely related, though.

If I stay withing single Sat, but change TP, dd sends Diseqc command anyway.

NoOne1 · Postby **NoOne1** » Fri Sep 22, 2006 11:41 am

Good afternoon,

Rain forecast drastically changed, my outside work is cancelled.

Just dl'd and am about to start tests.

====
NoOne1
Your humble servant for the next 36 hours if needed.

genpix · Postby **genpix** » Fri Sep 22, 2006 12:02 pm

forgot to mention:
"Auto retune" error for V/L tps is laways present (with or without RAW disecq command).

Postby **rel** » Fri Sep 22, 2006 12:55 pm

found several bugs in diseqc routines (some broken recently, some were never working before probably) I will fix them and try to release another test version .

thanks

NoOne1 · Postby **NoOne1** » Fri Sep 22, 2006 1:34 pm

Well Rel, this is kind of meaningless... VBG

Had a phone call interruption and you posted about addtional opportunities...

Here is my initial response

====
v1.2d test 1 - Diseqc switch tests...

1) LNB Type and Uncommitted switch settings are retained...

HOWEVER, these are currently present...

2) No routing beyond Uncommitted switch value zero occurs.
i.e. only the first four basic DiSEqC ports are available.

3) a Signal Strength of 23% and Quality of 100% is always displayed on non addressable ports.

4) Actual change on a Uncommitted address does not occur until exit/close restart of app even though osd changes.

More, but got to get by number 2 for me to go further.

====
NoOne1

genpix · Postby **genpix** » Fri Sep 22, 2006 3:27 pm

Rel,

On a second thought, you are probably right about different scenarios (with or without RAW commands).

If I enter raw command, I could make it as long as I want, put as many repeats as I want, choose the order of commands I want, etc.
And I should not care about "force repeats" in preferences.

The only problem: is there special symbol for pause?
for instance, I'd like to enter something like this:

Code: Select all

E0 10 38 F2 W E0 10 39 F6 W E0 11 38 F2

where W is a wait/pause (25ms or so).

Let me know if you have some special symbol for that.
Otherwise, I could send the command like:

Code: Select all

E0 10 38 F2 E0 10 39 F6 E0 11 38 F2

and inspect the Diseqc string in my driver (for instance, insert pauses in front of E0 or E1 bytes).

Unless you already do this

Snuffer · Postby **Snuffer** » Fri Sep 22, 2006 4:28 pm

@ Genpix

To see the commands take a coil 20 turns 4 mm and a C 10 Pf in serie of Coax.

Put this signal to the mic in of the audiocard and see the commants as amplitude best to do is Cool 96.

You can see the commends and pauze from the commands,

genpix · Postby **genpix** » Fri Sep 22, 2006 4:37 pm

LOL

I use scope to see the real DISEqC curves.
But for this thread I don't even need this.

Cause I see all DISEqC dumps in DbgView window

And I don't even need to insert a pause between DISEqC packets (my own firmware does it automatically). All I need (in driver) is to break long DISEqC string into packets.

And I bet, Rel is already doing this (I'll check it when I get home).

genpix · Postby **genpix** » Fri Sep 22, 2006 11:39 pm

I just checked. DD already breakes long DISEqC comand into separate elementary commands

Snuffer · Postby **Snuffer** » Sat Sep 23, 2006 12:28 am

Ok i do not now you can see it by a scoop and for a simple feedhunter it's nice to check for me.

Postby **rel** » Sat Sep 23, 2006 5:14 am

TEST 2: (PATCH FOR TEST 1)

Download

I've tested with SS2 and a 4 port Diseqc switch. everything seems to work fine.

Now some replies:

2) RAW diseqc field is NOT empty
Uncommited switch settings are ignored (uncommitted command gets never send)
"Always force to resend Diseqc command" is ignored
number of resends is ignored (only one command is send)

should be ok now

"Always retune" = OFF
error is still the same

tune is always sent when V/L polarity is selected

Problem was happening when one of DP LNB types is selected. fixed.

If I stay withing single Sat, but change TP, dd sends Diseqc command anyway.

Probably the thing I will say is meaningless for a DP LNB (because it uses just one polarization):
DD could resend the diseqc command if polarization changes (and also if band changes when Ku-Linear / Universal is used)
Because diseqc specification and some switches need the commands to be resent when pol or band is changed. (Different commands will be produced for different polarizations and bands)

i.e there are different commands for Diseqc Switch Port 1:
0xF0 = Vertical, Low Band
0xF1 = Vertical, Hi Band
0xF2 = Horizontal, Low Band
0xF3 = Horizontal, Hi Band

ofcourse this might be meaningless for DP LNB.

[3240] DISEqC: 4 bytes - E0 10 39 F0
[3240] DISEqC: 4 bytes - E0 10 39 F0
[3240] DISEqC: 4 bytes - E0 10 38 F2
[3240] DISEqC: 4 bytes - E0 10 38 F2

But, I would expect them interleaved. Don't you think so?

I thought it should ensure that the first switch functioned well (repeating commands), then should send the

commands for 2nd switch. Other wise repeats could be unsuccessful.

2) No routing beyond Uncommitted switch value zero occurs.
i.e. only the first four basic DiSEqC ports are available.

3) a Signal Strength of 23% and Quality of 100% is always displayed on non addressable ports.

4) Actual change on a Uncommitted address does not occur until exit/close restart of app even though osd changes.

Now everything seems to work fine with my SS2 . DD sends right commands to the device. I've also tried setting the uncommitted values to see if it is sent. (just checking b2c2_direct.log), all fine. If there is still problem, reason might not be the commands sent by DD.

NoOne1 · Postby **NoOne1** » Sat Sep 23, 2006 6:38 am

Good morning Rel,

This is a quick note as I have to leave town in about an hour. Will contact you Monday to do more testing at that time.

w2kp, p4-2.4GHz
==
ss2-2.3p drv4.4.0
Correct routing via Basic DiSEqC (4x1) only.
Uncommitted switch values are stored but operate as though always set to zero.

th1020a-03t
Correct routing of all DiSEqC switching commands tested.

Did not test any other cards at this time.
====

Just a little food for thought on version 12dtest2.
The manner in which it was provided invalidates any inital install tests.

====

We need to get you at least one Spaun 420 to test with.

What else do you need i.e. cards.

Stabs in the dark are beneath you.
====

Leaving in about 20 minutes...

======
NoOne1

Postby **rel** » Sat Sep 23, 2006 7:24 am

ss2-2.3p drv4.4.0
Correct routing via Basic DiSEqC (4x1) only.
Uncommitted switch values are stored but operate as though always set to zero.

I think I've just fixed this for test3 (but probably I will just release v1.1d)
it was particular to SS2 device. Solution will work only with MIBI's driver patch.

We need to get you at least one Spaun 420 to test with.

What else do you need i.e. cards.

Stabs in the dark are beneath you.

oh, heh, the things I lack while testing are a Dishpro LNB (impossible to find in Europe), an uncommitted switch (not sure about Europe but none of the sat shops in Turkey sells them)

PS: If you don't use a committed swtich (cascaded with an uncommitted switch), just select "None" as Diseqc Type and specify the port of uncommitted swtich on the satellite selection dialog.

I dont have time to struggle with the complexity of the graphical support of uncommitted switches in diseqc tree at the mooent.

genpix · Postby **genpix** » Sat Sep 23, 2006 8:51 am

in Test2 version, with
"Alway retune" = OFF
"Always force to resend Diseqc command" = ON
"Repeat count" = 1

DD works with DP LNBs as I expect

Only minor problem:
if RAW command is present in Diseqc configuration, it does NOT overwrite "default" diseqc command.
Both commands are sent, RAW is followed by "default" (essentially killing what RAW was supposed to do).

v1.2d test 1 - Diseqc switch tests needed

Who is online