Empeg stuck in AC mode. Dead Transistor is the culprit.

Posted by: TRUNGLE

Empeg stuck in AC mode. Dead Transistor is the culprit. - 15/01/2004 05:01

I recently had a short in my cigarette lighter and now the Empeg is stuck in AC mode. With the help of my backup unit, I was able to trace the problem to a transistor. It is located below the upper right hand corner screw on the main board. It has a marking "MG A6" on it. Does any one know where I can get a replacement for this part or maybe a similar one? Any info is appreciated. Thanks.

- Tom
Posted by: Shonky

Re: Empeg stuck in AC mode. Dead Transistor is the culprit. - 15/01/2004 07:20

Sure it's a transistor? I'm guessing since I can't see inside my empeg right now. Is it a SOT23 package? It could be single or a dual diode which can quite easily look like a transistor

See here:
http://www.semiconductors.philips.com/pip/BAS16W.html
http://www.repairfaq.org/REPAIR/F_SMD_trans.html
Posted by: JBjorgen

Re: Empeg stuck in AC mode. Dead Transistor is the culprit. - 15/01/2004 07:31

Ah...maybe that's the same problem I have... I'm pretty sure mine is not the AC sensor switch. I had a similar short when installing in my wife's car and one time when I put the empeg in it went pop and turned off. Since then, it will only boot in AC mode. PLEASE let me know if that fixes the problem.
Posted by: tfabris

Re: Empeg stuck in AC mode. Dead Transistor is the culprit. - 15/01/2004 11:52

If replacement of this transistor solves the problem, please notify me or follow up on this thread. I should like to update the FAQ entry about the AC sensor switch if that's the case.
Posted by: alex25

Re: Empeg stuck in AC mode. Dead Transistor is th - 16/01/2004 01:32

Please also have a look at this thread:
http://empeg.comms.net/php/showflat.php?Cat=&Board=bigs&Number=151957&page=&view=&sb=&o=&vc=1
Posted by: TRUNGLE

Re: Empeg stuck in AC mode. Dead Transistor is the culprit. - 16/01/2004 05:48

Thanks for the info. You're right, it's a diode in a SOT23 package. Looks like a common component. I'm going to see if I can find one in my pile of surplus parts.

http://www.fairchildsemi.com/ds/BA/BAS16.pdf
Posted by: TRUNGLE

Re: Empeg stuck in AC mode. Dead Transistor is the culprit. - 16/01/2004 06:45

Fixed! My player is now switching between AC/DC mode. I found an "A6" diode on a Diamond Multimedia Supra Express 56i ISA modem. I'm glad I didn't chuck it when I switched to an all PCI MB. I'm attaching a picture of the component for those that would like to see what it looks like.

I hope this info will help you fix your player.

In DC mode, I was measuring 2.53V on pin 1 of the diode when it should be 0.5V
Posted by: TRUNGLE

Re: Empeg stuck in AC mode. Dead Transistor is th - 16/01/2004 06:50

So does this mean that if the player is stuck in DC mode then it is the buffer ic and if it stuck in AC mode then is the diode? Oh well, I'm just glad that your player and my player is working like it should.
Posted by: TRUNGLE

Re: Empeg stuck in AC mode. Dead Transistor is the culprit. - 16/01/2004 06:53

During the course of this little venture, I also decided to upgrade the HD to a Seagate Momentus 40GB. Couldn't have done it without the guidance of your excellent FAQ. Thanks Tony.

Can you edit the thread title to Defective Diode instead of Dead Transistor?
Posted by: JBjorgen

Re: Empeg stuck in AC mode. Dead Transistor is the culprit. - 16/01/2004 07:51

Could you elaborate (and use small words for the EE ignorant) on how to test to see if this is the same issue on my player? I have my player dissected and a VOM handy.

EDIT: Your picture was too large...it didn't attach
Posted by: JBjorgen

Re: Empeg stuck in AC mode. Dead Transistor is the culprit. - 16/01/2004 10:39

I think I've found it...is it between the +3v jumper and the top of the board? Also, how do I test it? Do I just put the red on pin one and the black on the player case and measure the voltage?

I have an Mk2...so I imagine it could be in a slightly different spot on my board.
Posted by: TRUNGLE

Re: Empeg stuck in AC mode. Dead Transistor is the culprit. - 16/01/2004 17:39

Low Res: http://www.trung.com/empeg/A6-640x480.JPG
High Res: http://www.trung.com/empeg/A6-1280x960.JPG
Pin 1 and 3: http://www.trung.com/empeg/A6-PIN1AND3.JPG

How to test the diode:

1) See the pictures to help you locate the component.

2) You need to power your empep like if it was in the car before running the test. Don't plug the AC adapter into the power jack. Power the unit through the sled connector and make sure that the 3A fuse on the Yellow line is good. I use a regulated 13.8V power supply but you can use the AC adaptor. If you want to use the AC adaptor then the center is Positive and the outside part is Ground. Connect the Yellow line to the Positive center and the Black wire to the Ground of the AC adaptor. Since you are dealing with power lines please take your time and verify all the connections before applying power.

3) Connect the black lead of the VOM to the chassis of the player.

4) Now find the floppy power connector. It has a Blue, Orange, Black, and Red wire on it. Use the red lead of the VOM to check for power on the Red line. Double check your connections if you are not getting power.

4) Use the red lead of the VOM to measure the voltage at point 1. (see picture) The work area is very tight, so be careful you don't create a short.

5) If the diode is working, then the reading on the VOM should be about 0.5V. My defective diode was reading 2.53V.
Posted by: JBjorgen

Re: Empeg stuck in AC mode. Dead Transistor is the culprit. - 20/01/2004 15:49

Thanks for the great info!

I've been putting off posting back until I can get some pictures.
Unfortunately, my MK2 board does not have that diode (at least in the same place.) Evidently the rearranged some of that stuff between the MK2 and the MK2a.

The only pic I could find that looks like mine is this lower-res image I cribbed from mrmunsell's fuse replacement guide.



I'm going to snap some higher res pics soon and hopefully figure out what could possibly be the issue.

Ideally, if PCA could comment on what would be the likely suspect...

Stay tuned...
Posted by: JBjorgen

Re: Empeg stuck in AC mode. Dead Transistor is the culprit. - 22/01/2004 11:35

I've got my empeg on the test bench, but am not having any luck figuring out what's wrong. (Mostly from lack of knowing what to check.)


The camera I borrowed doesnt seem to be much better than the one above...here's the best shot I could get, even in macro mode.



I'm afraid I'll have to send it off to someone who knows what they are doing.
Posted by: altman

Re: Empeg stuck in AC mode. Dead Transistor is the culprit. - 26/01/2004 03:48

The component in there is only present in the mk2a, and helps to protect a 74LVX04 (IC33); on a mk2, pin 9 of one of these (the one nearest the unused internal LCD connector) connected directly to the power jack switch.

If AC/DC sense dies on a MK2, and the fault is *not* with the plug itself, then you need to swap out IC33 which is a bit harder than a SOT23 but not too bad. The biggest issue is getting a TSSOP 74LVX04, in fact.

Hugo
Posted by: JBjorgen

Re: Empeg stuck in AC mode. Dead Transistor is the culprit. - 28/01/2004 13:22

[blank stare] Oh. [/blank stare]

/me shuffles off to the back of the support line.
Posted by: maczrool

Re: Empeg stuck in AC mode. Dead Transistor is th - 28/01/2004 13:29

Seems easy enough to get a TSSOP 74LVX04. Quite easy to find and purchase actually. So the diode protects IC33 in a MK2A? And if the diode goes out, apparently the layout is such that IC33 still survives. Replacing TSSOP and SOT is a piece of cake.

Stu
Posted by: JBjorgen

Re: Empeg stuck in AC mode. Dead Transistor is th - 28/01/2004 13:33

You've got mail via your website.
Posted by: TRUNGLE

Re: Empeg stuck in AC mode. Dead Transistor is the culprit. - 28/01/2004 16:59

Which one is IC33? This is a picture of a RIO/MK2a board.

The IC is available at Mouser Electronics
http://www.mouser.com/index.cfm?handler=displayproduct&lstdispproductid=304482&e_categoryid=97&e_pcodeid=51218

Posted by: altman

Re: Empeg stuck in AC mode. Dead Transistor is the culprit. - 29/01/2004 06:05

The top left one, ie the one closest to the connector labelled LCD

Hugo
Posted by: TRUNGLE

Re: Empeg stuck in AC mode. Dead Transistor is the culprit. - 29/01/2004 13:47

I thought that was the one you were talking about but was not quite sure. Thanks for the info.

Tom

Posted by: maczrool

Re: Empeg stuck in AC mode. Dead Transistor is the culprit. - 26/08/2005 01:22

Which one is IC33 in an MK2? There are buffer chips on either side of the LCD connector pads.

Stu
Posted by: JBjorgen

Re: Empeg stuck in AC mode. Dead Transistor is the culprit. - 15/09/2006 12:28

Seeing as much of the brain trust is in one location this weekend, I thought I'd bring this thread back from the long dead.

What else can affect the ability to dectect AC or DC power? I had Ian replace the LVX04 chip at the empeg meet in Cincy two years ago. That didn't fix the problem, so Mark replaced the power jack for me a couple weeks ago. Still no dice. Empeg seems to work quite well other than this.
Posted by: maczrool

Re: Empeg stuck in AC mode. Dead Transistor is the culprit. - 15/09/2006 15:42

Pretty much you have only the level shifter IC (LVX04) and the GPIO on the Strongarm on an MK2 and the MK2A has a diode. I think maybe there is a voltage divider resistor network with two 12K resistors as well.

Nothing else there.
Stu
Posted by: tfabris

Re: Empeg stuck in AC mode. Dead Transistor is the culprit. - 16/09/2006 05:43

Quote:
What else can affect the ability to dectect AC or DC power?

Damage to circuit traces or solder joints somewhere on the PCB?

Fried code in the power-control PIC?

Disclaimer: These are wild guesses from someone who knows nothing of the hardware.
Posted by: maczrool

Re: Empeg stuck in AC mode. Dead Transistor is the culprit. - 04/12/2006 15:51

Quote:
What else can affect the ability to dectect AC or DC power? I had Ian replace the LVX04 chip at the empeg meet in Cincy two years ago. That didn't fix the problem, so Mark replaced the power jack for me a couple weeks ago. Still no dice. Empeg seems to work quite well other than this.


Did you ever get this sorted out?

We've got an MK2 unit which we replaced both those parts in also and it did not fix the problem either. In addition to not sensing AC properly, this one also has trouble starting up on occasion. The display light will turn on and flash a couple times, but the player doesn't boot and the display remains blank except for the heating wires. The only solution is to pull the power and reapply.

The problem doesn't seem to be any of the resistors in the path. Does anyone know if these problems could be caused by the power control PIC?

Stu
Posted by: JBjorgen

Re: Empeg stuck in AC mode. Dead Transistor is the culprit. - 04/12/2006 17:13

Nope. Never did solve it. If I have Hijack Force DC mode, everything seems to work as it should, so I never bothered. I'd be very interested if you find a solution. Hugo? Patrick? Anyone out there?
Posted by: altman

Re: Empeg stuck in AC mode. Dead Transistor is the culprit. - 05/12/2006 01:42

Hmmm. Usually when the display light goes out and nothing happens it's the 60v PSU trying to start but taking too much current and causing the system to brownout (and cycle). Did you have a look at the 60v supply? This should come on the instant the standby light stops flashing.

Don't think it'd be the power PIC (it's the PIC that turns the display on).

Hugo
Posted by: maczrool

Re: Empeg stuck in AC mode. Dead Transistor is the culprit. - 05/12/2006 12:49

Quote:
Hmmm. Usually when the display light goes out and nothing happens it's the 60v PSU trying to start but taking too much current and causing the system to brownout (and cycle). Did you have a look at the 60v supply? This should come on the instant the standby light stops flashing.

Don't think it'd be the power PIC (it's the PIC that turns the display on).

Hugo


It's not the display, we get 58+ volts. The hard drives even spin up and stay spinning. I am trying running with only one HD at a time and see if this could be the issue. I'll report back if this fixes the non-booting issue.

Any ideas about why a number of units can't detect AC/DC properly even after replacing the buffer and power jack?

Thanks,
Stu
Posted by: mlord

Re: Empeg stuck in AC mode. Dead Transistor is the culprit. - 05/12/2006 13:28

Quote:
this one also has trouble starting up on occasion. The display light will turn on and flash a couple times, but the player doesn't boot and the display remains blank except for the heating wires. The only solution is to pull the power and reapply.


This often happens on perfectly good units when rebooting via the Hijack menu. But it should not happen on any fresh power-up.

-ml
Posted by: mlord

Re: Empeg stuck in AC mode. Dead Transistor is the culprit. - 05/12/2006 13:30

Quote:
Quote:
What else can affect the ability to dectect AC or DC power? I had Ian replace the LVX04 chip at the empeg meet in Cincy two years ago. That didn't fix the problem, so Mark replaced the power jack for me a couple weeks ago. Still no dice. Empeg seems to work quite well other than this.


Did you ever get this sorted out?

We've got an MK2 unit which we replaced both those parts in also and it did not fix the problem either. In addition to not sensing AC properly, this one also has trouble starting up on occasion. The display light will turn on and flash a couple times, but the player doesn't boot and the display remains blank except for the heating wires. The only solution is to pull the power and reapply.

The problem doesn't seem to be any of the resistors in the path. Does anyone know if these problems could be caused by the power control PIC?

Stu


The thing to do next would be to check for continuity (and also for shorts to GND or +5V) between the power jack's sensor pin and the 74LVX04 chip (pin 9). Then do the same checks again between the 74LVX04 (pin 8) and the SA1100 CPU (pin 53). The fault pretty much has to be on one of those two traces, if the power jack and 74LVX04 themselves are okay.

-ml
Posted by: maczrool

Re: Empeg stuck in AC mode. Dead Transistor is the culprit. - 05/12/2006 14:30

Funny thing is I checked all those things. The switch definitely works fine and there's continuity between all those pins and no continuity between +5V and ground. I've put two different 74LVX04s in the unit and neither one has helped. Pin 53 of the Strongarm goes high and low as it should so the 74LVX04 should be working, but still the unit remains stuck in AC mode. Doesn't the power control PIC play some part in how this all works?

Stu
Posted by: maczrool

Re: Empeg stuck in AC mode. Dead Transistor is the culprit. - 05/12/2006 15:27

I tried using only one HD at a time, but the problem of not booting remains on occasion.

Stu
Posted by: mlord

Re: Empeg stuck in AC mode. Dead Transistor is the culprit. - 05/12/2006 16:27

Quote:
Pin 53 of the Strongarm goes high and low as it should


I believe that (pin 53) is all that is necessary for this stuff to work.

The player s/w uses pin 53 to detect AC/DC. So if you see the player thinking "AC" when it's really on "DC" (or the other way round), and pin 53 has been verified to be getting the correct levels, then that input on the SA1100 is probably fried.

Just to recap: is that the actual problem here: player uses DC when on AC (or vice-versa)?

??
Posted by: maczrool

Re: Empeg stuck in AC mode. Dead Transistor is the culprit. - 05/12/2006 16:59

Quote:

I believe that (pin 53) is all that is necessary for this stuff to work.

The player s/w uses pin 53 to detect AC/DC. So if you see the player thinking "AC" when it's really on "DC" (or the other way round), and pin 53 has been verified to be getting the correct levels, then that input on the SA1100 is probably fried.

Just to recap: is that the actual problem here: player uses DC when on AC (or vice-versa)?

??


Player on DC equals 3.3 volts at pin 53, but player is in AC mode. Player on AC equals 0 volts at pin 53 and player is in AC mode. So you vote for dead GPIO? I was leaning towards that too, I was just hoping there was some other component that might be to blame.

The other problem is that the unit does not boot reliably. Perhaps they are both from a sick Strongarm?

Stu
Posted by: mlord

Re: Empeg stuck in AC mode. Dead Transistor is the culprit. - 05/12/2006 17:41

Just for fun, try this:

(1) Go to the Hijack "Force AC/DC Power Mode" menu, and ensure it is set for Normal.

(2) Reboot the player using the Hijack menu to do so.

(3) What power mode does the player now think it is using?
(4) What is the "Current Mode" shown by Hijack on the "Force AC/DC Power Mode" screen?

-ml
Posted by: mlord

Re: Empeg stuck in AC mode. Dead Transistor is the culprit. - 05/12/2006 17:42

Of course, try all of that when using DC power from the docking connector (but not from one of my docks).

-ml
Posted by: maczrool

Re: Empeg stuck in AC mode. Dead Transistor is the culprit. - 05/12/2006 18:17

Quote:
Just for fun, try this:

(1) Go to the Hijack "Force AC/DC Power Mode" menu, and ensure it is set for Normal.

(2) Reboot the player using the Hijack menu to do so.

(3) What power mode does the player now think it is using?
(4) What is the "Current Mode" shown by Hijack on the "Force AC/DC Power Mode" screen?

-ml


When connecting to the docking connector, sometimes I get a DC reading, sometimes an AC and sometimes it refuses to reboot all the way. Most of the time I get AC.

Stu
Posted by: maczrool

Hijack preventing AC/DC Detection and Booting? - 11/01/2007 18:18

I think I have at least a little bit of an idea about why the unit would not detect DC or boot reliably. It seems to have something to do with Hijack or some setting associated with it. I have attached a boot log (no HDs attached) which first shows a successful boot and then a nonsuccessful boot. I noticed that the unit would always fail at the same spot which is when it loads the custom animated boot logo. I do not know what this has to do with not reading the state of the powerjack, but it certainly seems they are related. Mark, do you have any idea why a boot logo or anything involved with Hijack would cause the two problems of unreliable booting and no sense of power jack? I also noticed it says tuner: loopback=1, ID= -1 even though the Hijack menu was set for a normal boot.

Thanks for any help,
Stu
Posted by: mlord

Re: Hijack preventing AC/DC Detection and Booting? - 12/01/2007 02:48

Quote:
I noticed that the unit would always fail at the same spot which is when it loads the custom animated boot logo.


I think one obvious thing to do, is to delete the boot animation from that player. I think JEmplode (or LogoEdit ?) should be able to do that. Reinstall Hijack after doing so, and then see if it behaves better or not.

-ml
Posted by: maczrool

Re: Hijack preventing AC/DC Detection and Booting? - 12/01/2007 11:47

Hi Mark,

Thanks. That's exactly what I proceeded to do (with JEmplode) and yes without the logo it detects and boots consistently with Hijack installed. It just seems really strange these problems were caused by a boot logo. Anyone know why it would cause this?

This is an MK2 player, so is it possible it doesn't have enough memory to display the logo and carry out the boot process simultaeously? The logo was kind of long although I don't remeber exactly how long.

Tony, if you are reading this maybe you can add something in the FAQ about removing boot logos as a solution to not detecting AC/DC and or incomplete booting.

Thanks,
Stu
Posted by: JBjorgen

Re: Hijack preventing AC/DC Detection and Booting? - 12/01/2007 11:53

Hmm. Maybe I'll give that a try with my player (also an MK2). Mine has upgraded memory (16 meg), so if that fixes the problem, we can count out the memory.
Posted by: mlord

Re: Hijack preventing AC/DC Detection and Booting? - 12/01/2007 12:07

Quote:

This is an MK2 player, so is it possible it doesn't have enough memory to display the logo and carry out the boot process simultaeously?

Nope. That's always okay, regardless of memory.

Quote:
The logo was kind of long although I don't remeber exactly how long.
That's what I was suspecting. We don't really have any safeguards against super-long animations, and if they are too long, they'll overwrite parts of the kernel[EDIT] (actually, the "initrd" gets overwritten -- which is just as bad) [/EDIT], which is never really that good.
Posted by: mlord

Re: Hijack preventing AC/DC Detection and Booting? - 12/01/2007 12:12

Mmm.. and there's also this note I placed in the kernel source code years ago:
Code:

// Mmm.. bit of a race condition here: we shouldn't be accessing
// the flash memory directly (playing the animation), in case empeg_state
// wants to write to it when accessed via hijack_init() from display_animation().
// So what we do about it is.. nothing. Probably not an issue anyway.


I think a long animation would greatly increase the chances of this situation arising and causing problems. But there's not enough info to know if that was the case here (or is there? I didn't see any sign of a "bad flash checksum" message in the logs).

-ml
Posted by: mlord

Re: Hijack preventing AC/DC Detection and Booting? - 12/01/2007 13:26

I need a copy of that HUGE animation file, please.
Posted by: maczrool

Re: Hijack preventing AC/DC Detection and Booting? - 12/01/2007 15:03

Let me see if I can get a hold of it. It's not my player, and I was more concerned (at the time) about getting it working. I will post it if the owner sends me a copy.

Stu
Posted by: mlord

Re: Hijack preventing AC/DC Detection and Booting? - 12/01/2007 15:22

Ok, thanks.

What I think happened is this:

The animation was *way* too huge, spilling over into the final flash memory block of the kernel. When uploading Hijack, it overwrote the tail end of the animation, which just happens to contain the "frame index table" of the animation. This caused the animation player to "go wild" while trying to display it, which could make just about anything happen..

Of course, that's just my theory for now. If the actual animation was not "too huge", then something else must have been wrong.

Cheers
Posted by: maczrool

Re: Hijack preventing AC/DC Detection and Booting? - 12/01/2007 15:29

Well I got a copy of the logo. In my memory it sure seemed longer than this; guess I need an upgrade . No telling why it would cause a problem. I guess I could try reloading it and see if the problem comes back.

Stu
Posted by: tfabris

Re: Hijack preventing AC/DC Detection and Booting? - 12/01/2007 17:16

How many frames is that animation?
Posted by: maczrool

Re: Hijack preventing AC/DC Detection and Booting? - 12/01/2007 18:04

28 frames. I just put it back on and it still kills the player.

Stu
Posted by: wfaulk

Re: Hijack preventing AC/DC Detection and Booting? - 12/01/2007 18:23

I can't imagine that the number of frames has any bearing on the issue. The total binary size should be the relevant datum.
Posted by: maczrool

Re: Hijack preventing AC/DC Detection and Booting? - 12/01/2007 19:06

Funny, when I reloaded Hijack and removed the animation, the player remained unhappy. This time once Hijack was removed the player returned to normal. I now get the attached boot with Hijack and no animation.

Stu
Posted by: tfabris

Re: Hijack preventing AC/DC Detection and Booting? - 12/01/2007 21:19

Quote:
This time once Hijack was removed the player returned to normal.

That would be consistent with Mark's hypothesis. Rememeber that "removing" hijack actually isn't removing anything, it's replacing the kernel with a stock kernel.

If I'm understanding the thread properly, Mark's hypothesis is that an overly large animation file (hence my question about number of frames) will overwrite kernel code because the two things are intermingled in the flash RAM. So if you remove the offending animation, the kernel code is still hosed. Replacing the hosed kernel code would fix it.

Of course, the above makes a lot of assumptions.

EDIT: No, I was reading Mark's hypothesis wrong. Never mind.
Posted by: maczrool

Re: Hijack preventing AC/DC Detection and Booting? - 12/01/2007 21:37

Yep. For some reason the player will not work properly even with a fresh copy of Hijack and no animation.

Stu
Posted by: mlord

Re: Hijack preventing AC/DC Detection and Booting? - 12/01/2007 21:50

Could somebody please convert that .gif animation to Empeg format for me?

Thanks
Posted by: mlord

Re: Hijack preventing AC/DC Detection and Booting? - 12/01/2007 22:00

Quote:
Yep. For some reason the player will not work properly even with a fresh copy of Hijack and no animation.



How are you "removing" the animation? Hijack probably thinks it's still there.

Here, try this copy of Hijack: http://rtr.ca/v466_noanim.zImage
It *ignores* any user-supplied animation, and should boot up if that's really the problem here.

If it does boot up, then FTP to that player and GET /proc/empeg_kernel, and send it to me for analysis. You could also get /proc/empeg_kernel via the web interface, if that's any easier for you.

Thanks.
Posted by: maczrool

Re: Hijack preventing AC/DC Detection and Booting? - 12/01/2007 22:42

I was removing the animation in Jemplode using the clear animation command.

Okay I tried the new Hijack, but the same problem is presenting itself. It will boot every so often, but not much. As soon as I reinstall 2.01, all is well.

Stu
Posted by: maczrool

Re: Hijack preventing AC/DC Detection and Booting? - 12/01/2007 22:50

Mark, I have the Empeg kernel, but it cannot be attached due to size constraints. Would you like me to e-mail it to you?

Thanks,
Stu
Posted by: drakino

Re: Hijack preventing AC/DC Detection and Booting? - 13/01/2007 00:29

Quote:
Mark, I have the Empeg kernel, but it cannot be attached due to size constraints. Would you like me to e-mail it to you?


I just bumped up the attachment limit to something that should be large enough for it here.
Posted by: maczrool

Re: Hijack preventing AC/DC Detection and Booting? - 13/01/2007 12:37

Thanks Tom. Here's the empeg kernel I got during one of the times the unit was able to boot (obviously).

Stu
Posted by: mlord

Re: Hijack preventing AC/DC Detection and Booting? - 13/01/2007 13:03

Quote:
Thanks Tom. Here's the empeg kernel I got during one of the times the unit was able to boot (obviously).



Okay, this one was with the animation "removed". But it's really still there -- removal just means blanking out the animation "signature": 'ANIM' at the end of the file you posted. I put those chars back, and it boots and runs for me (with or without the ANIM).

The kernel in there is the special one that ignores the user-supplied animation, so I next replaced it with stock Hijack v466, and now it boots and plays the animation quite reliably.

So.. there is a hardware problem somewhere.

Cheers
Posted by: maczrool

Re: Hijack preventing AC/DC Detection and Booting? - 13/01/2007 13:34

Thanks. Fundamentally it makes sense that there is a hardware problem, it's just weird that it works correctly with the stock kernel and that it takes the Hijack kernel to uncover the hardware problem. Certainly there is something different about this player from others as I've never seen one that wouldn't run Hijack.

Thanks again,
Stu
Posted by: mlord

Re: Hijack preventing AC/DC Detection and Booting? - 13/01/2007 14:05

Mind you, I only have Mk2A units to test with here, no Mk2 units.
Posted by: mlord

Re: Hijack preventing AC/DC Detection and Booting? - 13/01/2007 14:17

Here's a special kernel to help debug this (maybe).

http://rtr.ca/v466+.zImage

Just capture the boot log and post it here.

Thanks.
Posted by: maczrool

Re: Hijack preventing AC/DC Detection and Booting? - 13/01/2007 18:46

Did you want a bootlog with a successful boot or unsuccessful? So far all I can get is unsuccessful which is attached. If I can ever get it to boot with Hijack I will post a bootlog from it.

Thanks,
Stu
Posted by: maczrool

Re: Hijack preventing AC/DC Detection and Booting? - 13/01/2007 18:51

Here's one with a no boot immediately followed by a good boot which I now see is what you wanted.

Thanks,
Stu
Posted by: mlord

Re: Hijack preventing AC/DC Detection and Booting? - 13/01/2007 20:29

Okay, same drill again, but with this kernel:

http://rtr.ca/v466++.zImage

Cheers
Posted by: maczrool

Re: Hijack preventing AC/DC Detection and Booting? - 14/01/2007 15:01

I was unable to get the unit to boot after 20 some odd attempts, but here is the output without boot.

Thanks,
Stu
Posted by: mlord

Re: Hijack preventing AC/DC Detection and Booting? - 15/01/2007 02:16

Quote:
I was unable to get the unit to boot after 20 some odd attempts, but here is the output without boot.
...

DEBUG: empeg_state_restore: loop a=5

DEBUG: empeg_state_restore: done a=5

DEBUG: empeg_state_restore: exit

DEBUG: done empeg_state_restore(), failed=0

DEBUG: calling hijack_read_tuner_id()



Mmm.. that's interesting. It shouldn't be getting stuck *there* !
Does it appear to *always* die at roughly the same spot, hijack_read_tuner_id() ?

If so, then here's the next one to try:

http://rtr.ca/v466+++.zImage

Cheers
Posted by: maczrool

Re: Hijack preventing AC/DC Detection and Booting? - 15/01/2007 13:30

Thanks for your continued interest in this. Yes it stalled at the same point every time. Let me run this new Hijack and see what we get.

Stu
Posted by: maczrool

Re: Hijack preventing AC/DC Detection and Booting? - 15/01/2007 14:49

Here is an unsuccessful boot log. It runs without end, so I posted a portion of it.

Thanks,
Stu
Posted by: maczrool

Re: Hijack preventing AC/DC Detection and Booting? - 15/01/2007 14:50

Here is a successful boot log. Both are with the newest Hijack you posted.

Thanks,
Stu
Posted by: mlord

Re: Hijack preventing AC/DC Detection and Booting? - 15/01/2007 20:31

Quote:
Here is a successful boot log. Both are with the newest Hijack you posted.

Thanks,
Stu


Okay, when it fails, it's because it sees an endless stream of "data" coming in over the tuner serial port. Which is somewhat peculiar.

The failed boot appears to have happened while "docked", right?

Do you recall if the "successful" boot was also from a home dock" ?

Here's a kernel that should fix it, by putting an upper bound on how long we're willing to listen to the "tuner" sending us serial data:

http://rtr.ca/v467pre.zImage

Let me know what happens.. if it boots up, then reinstall the animation and verify that all is working again. I'll then release that as Hijack v467.

Cheers
Posted by: maczrool

Re: Hijack preventing AC/DC Detection and Booting? - 15/01/2007 23:17

I'll try the new Hijack, but no the unit was running straight from the AC jack on the back- no dock for either good or bad boot. I'll report what happens though.

Many thanks,
Stu
Posted by: maczrool

Re: Hijack preventing AC/DC Detection and Booting? - 15/01/2007 23:32

It seems to boot reliably now (at least so far) but AC/DC detection is still iffy. Running from the docking connectors on back, It wants to stay on AC most of the time from a hard reset. If I tell Hijack to reboot for normal operation, it seems to detect correctly. Anything separate from this in Hijack that would prevent proper detection of power source?

Thanks,
Stu
Posted by: mlord

Re: Hijack preventing AC/DC Detection and Booting? - 16/01/2007 13:39

Quote:
It seems to boot reliably now (at least so far) but AC/DC detection is still iffy. Running from the docking connectors on back, It wants to stay on AC most of the time from a hard reset. If I tell Hijack to reboot for normal operation, it seems to detect correctly. Anything separate from this in Hijack that would prevent proper detection of power source?

Thanks,
Stu


Okay, at least now we're back to debugging the hardware again.

There's something *really noisy* on the tuner serial port signals. Maybe there's a short in/near the docking connector, or perhaps the line-driver chip for the tuner serial port is fried/flakey.

I think it's quite likely that the same problem is also what is causing the AC/DC detection to flake out. So there's two paths for investigation now, and hopefully they'll lead to a common component or two on the board (?).

It would be useful (again) to see the boot logs for both success and failure now.

Cheers
Posted by: maczrool

Re: Hijack preventing AC/DC Detection and Booting? - 17/01/2007 00:04

I can no longer make it fail to boot, but I can give you a boot log with successful DC detection and without. The strange thing is even when Hijack thinks it's in AC mode, the boot log reads DC. I am wondering why the disconnect? The first one in this post is a successful boot with proper detection of DC and with DC behavior.

Thanks,
Stu
Posted by: maczrool

Re: Hijack preventing AC/DC Detection and Booting? - 17/01/2007 00:06

Here is a "bad" boot with improper detection of DC. The boot log says DC on the second line and yet Hijack and the player behavior says AC.

Thanks again,
Stu
Posted by: mlord

Re: Hijack preventing AC/DC Detection and Booting? - 17/01/2007 03:44

Quote:
I can no longer make it fail to boot, but I can give you a boot log with successful DC detection and without. The strange thing is even when Hijack thinks it's in AC mode, the boot log reads DC. I am wondering why the disconnect?


That's simple: When it's on DC and Hijack thinks it's on AC, this happens because the home-dock loopback test succeeds, and the player has Hijack set to automatically pretend it's on AC whenever a dock is detected.

When it fails, it's probably the noisy tuner serial port that causes the failure (try with Hijack v467 to find out for sure), and thus the home-dock loopback test also fails.

Is the home-dock's loopback wire loose in the docking connector? If so, then that could explain all of this trouble..

Cheers
Posted by: mlord

Re: Hijack preventing AC/DC Detection and Booting? - 17/01/2007 03:48

Quote:
Quote:
I can no longer make it fail to boot, but I can give you a boot log with successful DC detection and without. The strange thing is even when Hijack thinks it's in AC mode, the boot log reads DC. I am wondering why the disconnect?


That's simple: When it's on DC and Hijack thinks it's on AC, this happens because the home-dock loopback test succeeds


By the way.. if you really don't care about "home-dock" detection (via the tuner serial loopback wire), then you could go into the Hijack menu (long knob press), and change the Force AC/DC Power Mode setting to No-Loopback. It should then boot up in DC mode every time while docked.

But one should really try and figure out why the dock detection works sometimes, and fails other times due to noise on the loopback wire.

-ml
Posted by: maczrool

Re: Hijack preventing AC/DC Detection and Booting? - 17/01/2007 11:35

Quote:
Is the home-dock's loopback wire loose in the docking connector? If so, then that could explain all of this trouble..


I am using a special sled connector we set up for testing tuners and AC/DC detection. It only has one set of RCA outpus for the front, the tuner wiring, ignition sense and power with a female power jack on the end to accept an AC adapter. The noise you speak of must be internal to the player because when Hijack observed the noise, the connector was not being used and the player had the AC adapter directly plugged into it. By the way which wire is the loopback wire? I've not heard much about it.

Thanks again,
Stu
Posted by: mlord

Re: Hijack preventing AC/DC Detection and Booting? - 17/01/2007 12:40

Quote:
which wire is the loopback wire? I've not heard much about it.


Pins 6 & 7 on the docking connector are the Tuner serial port signals. If you tie them together, then anything sent out from the player is looped back into the player. This is how home docks are wired so that Hijack can detect them and use AC power settings even when really on DC power.

Cheers
Posted by: maczrool

Re: Hijack preventing AC/DC Detection and Booting? - 17/01/2007 12:53

Makes sense. As an FYI, the player with the no loopback and original boot animation is operating fine now in all respects- save for the noise on the serial line.

Hopefully this excercise made your already great accessory even better! If I find something funny on the serial lines, I'll post it here.

Many thanks,
Stu
Posted by: maczrool

Re: Hijack preventing AC/DC Detection and Booting? - 23/01/2007 01:50

I went in to the player and tested the serial lines for shorts or loose connections. There were none to be found. Is it possible that the RS232 transceiver has a fault even if the other serial functions (loading firmware) work fine? We've replaced the tranceivers before, but there were more overt signs of malfunction than here.

Thanks,
Stu
Posted by: mlord

Re: Hijack preventing AC/DC Detection and Booting? - 23/01/2007 16:59

Quote:
I went in to the player and tested the serial lines for shorts or loose connections. There were none to be found. Is it possible that the RS232 transceiver has a fault even if the other serial functions (loading firmware) work fine? We've replaced the tranceivers before, but there were more overt signs of malfunction than here.


Yeah, it's using different gates of the transceiver than those needed for software loading (as we both know, this is the tuner serial port, not the back panel serial port). Could be a bad driver chip (kinda behaving like a internal driver short, isn't it?), or noise on the CPU side of the driver chip.