Atlas JP1.3 Extender beta - Topic CLOSED

[Capn Trips]

Similarly to Jeff, I am no longer certain that I was using enough rigor earlier when I had non-functioning extended IR images. It may been me running into the "Absent setup code" problem.

Should that symptom recur, I will save and upload the offending IR image for analysis.

As for the shifting active devices, I will try to document accurately what occurs, but for starters, 2 specific circumstances (the ONLY macros currently assigned are the default device-selection macros):

(1) After extender activation, I selected the AUD device, and the first time that I use F1, OnDemand, or F2, the assigned (AUD) command executes, but subsequently the remote is in "CBL" device mode. I had only 1-command functions assigned to those buttons in device upgrades (not even keymoves, since they are in the CBL device buttonmap)

(2) again while in AUD device mode, when I pressed PIP ON/OFF (to which I had keymoved the extender deactivate function in TV mode) would execute the assigned funtion in AUD, but throw the remote into "TV" device mode.

Don't have my IR file with me, but will upload it later if necessary for analysis.

[speaker.guy]

F1, On Demand, and F2 have some hidden macros to control specific cable functions depending on the cable provider. They may still be loaded, and would surely leave the remote in cable mode after execution.....

[Capn Trips]

F1, On Demand, and F2 have some hidden macros to control specific cable functions depending on the cable provider. They may still be loaded, and would surely leave the remote in cable mode after execution.....

My experience has been that these Cable-company-installed macros are not "hidden" but are clearly visible in the macros tab of IR when one does an initial download of the remote, and I have most certainly deleted them. If there is some ADDITIONAL hidden function assignment to those buttons, I cannot imagine where and how, since I also have this symptom following a 981 reset and subsequent extender build.

When those pre-installed macros executed (in my case they were all 3-digit macros, apparently to select a particular TV channel) one could see three flashes of the CBL device button. In THIS symptom, there is NO flashing of the CBL device button until the NEXT button is pressed. It acts as if the INITIAL function I assigned to the button executed, and then a "X_CBL" command executed invisibly.

[vickyg2003]

Actually it sounds like perhaps UncleMilite is executing the wrong keymove macro search. The multi-macro group of keys are CBL keys, the PIP keys TV.... Or perhaps some of those unknown keys, govern which device is associated with those groups.

[unclemiltie]

There is no multimacro support in the extender. I'm doing a "classic extender" search of the advance code area on my own and deciding if it's a macro/keymove, etc.

However, I *AM* using the main loop in the remote to do all of the idle/backlight processing and that could also be playing into this if the F1-F3 keys are being processed prior to the call into the extender. I'm going to have to look into what happens prior to the extender being called. (of course, when I get some time)

For those who are adventurous, you could try moving the device table. I think I used RA0 as the starting address since I thought those were unused. It also appears that the R90 block is unused as well, so you could try to change that and then rebuild the extender to see if you get the same behavior.


-bill

[binky123]

I was able to get SKP/LKP working. I believe the timeout value was being placed into the wrong register. Here are the unofficial changes. unclemiltie will do some testing and release an official one later.

Code: Select all

		InclLDKP	= 1		;include LDKP protocol

		LD	W1,R03			;get control byte (first byte of keymove)
		LD	R_T0+1,RC1		;Get the control byte from the KeyMove
		SWAP	R_T0+1			;put Timeout value in low nibble
		AND	R_T0+1,#0F		;Set timeout
		CLR	R_T0			;clear upper byte of timeout counter
		CALL	StartT0

[greenough1]

Thank Binky123!

Best,
jeff

[Capn Trips]

Well, following up on Vicky's and Bill's comments, I tested a little more accurately, and there is something to what Vicky said.

The F1, OnDemand, F2 and PIP group of buttons all send the "active" device function ONCE and then shift the remote into TV device mode.

Similarly, the transport buttons all work once in the selected device mode and throw the remote into VCR mode thereafter!

The magic device re-selection is a serious problem from where I sit.

[greenough1]

Can anyone say why the phantom keys are not available to put a DSM on for example?
I like to define the keysets for a device as the short side of the LKP and put it on a phantom.

TV: Phantom1: DSM: M_TV;C_TV;V_TV;T_TV;O_TV
TV: Phantom2: LKP: [Short]: Phantom1; [Long]: other_stuff

You can choose the phantoms as part of the macro, but you can't put a special protocol on the key. The phantoms don't show up in the key drop-down menu. I hope that makes sense.

Where in the rdf does this get setup?

jeff

[vickyg2003]

Capn_Trips Wrote:

The F1, OnDemand, F2 and PIP group of buttons all send the "active" device function ONCE and then shift the remote into TV device mode.

Similarly, the transport buttons all work once in the selected device mode and throw the remote into VCR mode thereafter!

Ah that makes perfect sense. Typically on a remote that has punch through there are 2 device registers, the "active" device register, and the "saved" device registers. It sounds like one of these registers isn't getting set, so when it pops out, the device index isn't getting restored properly.

Since UncleMiltie is porting his Kameleon extender, punch through probably wasn't an issue.

[Capn Trips]

Curiouser and curiouser: As I continue pulling this string, I have noticed that each of the remotes behaves DIFFERENTLY regarding this device shift. Each remote has the exact same extender IR file uploaded.

When I press F1 on the extended Sunflower (silver, backlit learner), it shifts the remote into TV mode. Pressing PLAY shifts it to VCR mode;

When I test F1 on the TWC backlit learner, it also shifts device to TV. Pressing PLAY shifts it to VCR mode;

But on the TWC non-backlit non-learner, pressing either F1 or PLAY shifts the device to CBL!

So there appears to be something hard-coded in the non-writable area of the Flash-ROM that (a) controls this behaviour, and (b) may vary from provider to provider.

What a pita these cable company remotes can be!

Can the extender be altered to deal with this?

P.S. Great news about the progress on LKP. I can't wait for the next distro of the extender!

[greenough1]

Capn,
I think, in principle, the punch-through stuff can be dealt with . There are lots o'options in the unextended remote for dealing with the punch through stuff (general tab on the right side). I don't remember the details, but I had an issue with that at first, unextended, getting the remote to behave like I wanted.

I see in the asm vpt enable bits and an initial vpt device. We might need to add some extra logic to deal with it. (I wish I was more of a wiz with the extender software )

I tested binky123's changes and they seem to work... very nice indeed.

Best,
jeff

[vickyg2003]

Captain, its not THAT curious. I read the extender ASM code, and UncleMiltie is saving the last HT key into the saved Device register AND the active Device Register. But when he processes a key, he sets the Active Device Register. After a punch through key, most remotes restore the Saved Register to the active register, so I would think the restored device will vary depending on the last PseudoHT key was processed. I don't have much time available these days. When I'm at home I'm very busy. Too bad I can't do this at work, where I'm currently bored to tears.

[vickyg2003]

Jeff, I believe its the AllBind that is preventing your keys from being in the DSM dropdown list.

[speaker.guy]

So there appears to be something hard-coded in the non-writable area of the Flash-ROM that (a) controls this behaviour, and (b) may vary from provider to provider.

Did I suggest that?
[/code]