Strange new issue that I noticed today in my 95 VW golf.
I put the player on standby, and then unstandby-ed it a few minutes later.
It took a couple of seconds longer than I expected to come back, and when it did come back it was displaying the battery icon. The icon went away after a few seconds, and the player responded normally.

This was with the car running. And is infinitely reproducable.
I also tried it with the car off, and with my spare player too.
Same results every time. Always the battery icon right off standby.
Never had this problem before.
Nothing has changed in my car as far as my install goes, just thought it was kinda weird.

Anybody else ever experience anything similar?

Oh yeah, and I'm also having a weird problem where the volume will adjust a few notches up or down on its own. I know its not the remote, cause thats in the house. It only happens occasionally and I dont know whats causing it.
Oh well...

Cheers,
Zachary V.

Do you have big external amps? Perhaps when you come out of standby it turns on the amps, and that current draw is enough to drop the voltage. As to why it is just starting now, perhaps your alternator or battery are about to go. You could disconnect the remote wire from the amps and see if the player still reacts the same way.

i noticed something similar to this also. although mine only does it when first inserting in the dock (car running already), letting it boot, then when i power it on i notice the battery icon there for a few seconds and goes away. This just started happening this past week or so, never did it before.

I've just got one Kenwood external amp.
Same one I've always had, and its not even all that powerful.

Your comment about the battery may be something though.

I've been wondering when my battery was gonna go.

Its the same battery the car came with back in November of 1994.
By all rights it should be dead as a doornail by now.
7+ years on a stock VW battery!?

Would a dying battery trigger the icon though, even if the car is running?

Z~

That does sound similar.

But do you not get the icon if you go to and from standby without un-docking the player?

Z~

Would a dying battery trigger the icon though, even if the car is running?

It could do.

The amplifier starting up (when the player comes out of standby) can be a pretty good "shock" to the system -- a sudden request for, say, an extra 25 amps of power. Think of your battery as a reservoir to supply this extra demand until the alternator can take over the load.

If the battery can't do it, then you will get the symptoms you describe.

Try coming out of standby with the engine at 2000 RPM or higher, see if that makes any difference.

tanstaafl.

i have noticed this new exact same problem. car's system hasnt changed at all since installing the empeg, and the empeg never displayed this behaviour, till recently. last thing i did was : updated hijack. cant remember what version i am currently running (still on 2.0b11) but now when waking the empeg from standby to on, it will delay a noticeable 3/4 secs and the battery icon comes on. never did it before. i have an Optima yellowtop 750cca 52amp/hr battery, so i do not believe current is an issue, especially since the unit was fine before. my voltmeter reported a steady 13.4V one time, and empeg still delayed its start and put up the battery icon. in fact i may downgrade to an earlier hijack and see if this goes away - will try this later tonite and post here - unless im wrong and this is not hijack related.

I just starting noticing this after I upgraded to 2.0b13 a few days ago. I also have Hijack v283 which I updated from v261 at the same time. The icon stays on for a few mintues after I start the van and the empeg and then goes away. I'm not sure about the delay, I'll watch for that in the morning. It never did that before the update.

I turned the empeg on while doing 75mph on the highway, RPM's around 3000 and the battery light still came on. There's a 3-4 second delay of nothing, then the display comes on along with the battery icon which stays on for few seconds and goes away.

Sure sounds like a wiring or supply-voltage problem to me.

this is definately not a wiring problem. As others have said, they have only just gotten this recently after upgrading. It wasn't doing this before and my wiring is just fine.

i thought it was at first and was very worried - since my empeg gets its power direct from the 4gauge feed straight from the battery (same feed that goes to the amps) - now that i see others have experienced it, i feel that its not an isolated case of poor wiring/unbeefy electrical system.

on my one, when you press a button to wake it, the slowly pulsing LED goes off completely, entire unit is fully dead for those 4 secs, before coming on - as if it needed to drip-fill a bank of capacitors before getting enough power to turn on - do you guys get the same thing (LED cuts off completely...)?

nah, I've been seeing this effect after upgrading to 2b13 as well. Haven't changed anything except the software.

I'll look for that today. I'll also try turning it on after I have been driving a bit and see if it happens. I always bring mine up from standby. It's easier to turn it off since I have to open the drivers door after turning the key off to get it to stop playing.. So far the display is coming on with the battery icon in the left corner and it disappears after a couple of mintues.

I get the battery indicator for about 10 seconds after starting engine. Before b13 I only used to get it if I turned the ignition off and then restarted car within my 60 second timeout.

As I get it for such a short time, I'm not sure if it is connected with the above posts or not.

ok here are my findings so far. i found that i am currently on hijack 280 (i had forgotten heh). problem happens all the time, reproducable. my last update was from hijack 278. reloaded this on. tested - problem gone! did all kinds of things to get it to happen - no go - it turned on instantly and with zest each time. ok, reloaded hijack 280 back on. instantly - problem again.

i notice that, with 280, when it happens, the display seems to "fade in" from black (more descriptively, it "fades in" center first, then it works towards the left and right sides, far left and right coming in last) - this 'fade' happens only in about 1 sec tho.

with 278 on, the screen lights up instantly. no fade, pops to full brightness straight away, no battery icon, no problem.

anyway thats what ive got for now. its late so im scheduling the latest hijack, and 2.0b13 emplode and developer image to download overnight and see what happens. i'll try 2.0b13 first with no hijack, then again with hijack, may also experiment with the different versions of 278 and 280.

can definitely say that somethings affected the unit in hijack 280 that 278 seems to leave alone.

however if im misinterpreting, or if someone else can back this up or argue against it im curious to find out!

Worth trying hijack 286 or any from 283, Mark Lord cleaned things up for beta 13. Seems smooth.

I just did the exact same test as lofreq, with the same results.
I flashed HJ 286 and did the standby/recover thing and I got the 3-4 second delay, the slow screen fade in and the battery icon.

I then flashed HJ 253 (unfortunately the only older version I had before 281. And it appears all the old HJ versions are no longer downloadable from the hijack website) and I got a much quicker return from standby, and no battery icon.

Something definitely changed somewhere.
If somebody wants to send me HJ 278 and 279 I can try those as well.
So it would appear that this is not a result of wiring (phew) but a change in the code?
Maybe it only effects certain cars though? Due to different wiring schemas?
I dunno... Just know its broke in my 95 VW.

Zachary

v278 on the way in mail.

With 282 I get about 6 seconds worth of battery icon - that's after booting to stand-by in a car that's already started.

/Michael

Hmm, can we narrow it down further? I'm guessing either v279 or v280.

v282 Updated for v2.00-beta13 kernel changes (very minor).
v281 Fixed LBA48 bugs.
v280 Rewrote much of the ButtonLED Illumination code.
v279 Fiddled with reboot code.

I just tried HJ 278 courtesy of mtempsch and I DO NOT get the battery icon.
If somebody wants to send me 279 Ill give that a try too.

Zachary

Well I tried it today. I get the 4-5 second delay with a blank screen. Another 6-7 seconds with the battery icon and no sound. Another 3-4 seconds with the sound starting up slowly. I tried this off and on all day from just starting the van to doing it driving on the expressway and it was the same every time. So I guess you guys have figured out that it is something in Hijack.

Hey Mark!!

Someone then please explain why it doesn't happen on my player with Hijack and beta 13?

Because you're lucky

I certainly am not.

The Force DC/Power Mode is not enough to reproduce it. The unit has to be on REAL DC power through the docking connector.

Happens for me too, as ov v2-beta13.

-ml

Right, doesn't happen in my car either.

But I take it that you know what's going on and I don't need to worry about it?

I think he said it does happen for him.
Didnt he?

>Right, doesn't happen in my car either.

Now that's odd, really odd.

Maybe it's just another name colour thing.

I don't really know what's going on, but I have a theory.

Attached to my next posting is the actual Hijack changes that were new in version 280 -- the problem presumably goes away when using Hijack versions older than v280.

Pretty much the ONLY change that could affect any of this is the new code in v280+ that prevents scrambling of the display controller by having two threads trying to talk to it simultaneously. As of v280, there is a locking mechanism around that code, forcing new writers to wait until the current writer finishes up, and then some.

Perhaps the new player software issues a whole bunch of display controller commands in a row after returning from standby, such that the locking code forces it to wait slightly on each command, adding up to 3-4 seconds or so.

Note that the AUDIO starts immediately, but the display controller -- brightness and buttons -- doesn't become fully functional for a few seconds.

Harmless, but somewhat distracting.

We could fix it by keeping a queue of display controller commands in the kernel, and having a hijack thread to service them over time or something, so that the original caller need not wait for conflict-free access.

Personally, I'm going to ignore it and hope it goes away.

-ml

Attached are the diffs from Hijack v279 to v280.

If anyone wants to confirm (1) no issue on v279, and (2) delay on v280, I have restored those versions to the Hijack site. You'll have to manually edit the URL to get them, but they're there.

Cheers

ok, after an annoying amount of running from room to garage and back, and flashing in between, heres what ive found

at last post i confirmed that with 2.0b11 and hijack 278 there is no prob. with 280 there is the prob. i then went up to 2.0b13 (no hijack) - problem gone! i do believe some people have posted saying they had this issue after going to 2.0b13 - not so for me.

then i put hijack 286 onto my new working 2.0b13 - problem back again. mark brought back 279 so i tried that - problem gone again.

so it is an issue from 279-280. i looked at marks code attachment and nearly went immediately insane so i closed the browser window promptly...

i also saw that there is now 288 as the latest since yesterday, so i grabbed that as tried it as well, no luck, still battery icon rah rah.

my amps have a long turn on delay so i cannot tell if music starts immediately as the unit is awakened (even tho the display/LED are both fully dead...) so i cannot comment on that - however after doing it over and over i do feel that the sound comes on LATER, when the battery icon problem occurs, than when all is working normally. this isnt concrete tho.

anyway it seems to be an issue between 279 and 280 that has not yet been resolved in versions up till the latest (288 at time of post)

oh yeah - i did the 'force DC" modes and all that, it doesnt happen on AC power - only when docked in car.

mark if that narrows it down for you and youre bothered to have a nosey into it, you now have the controls sir

I can also confirm that it happens between 279-280.

I never saw the scrambled screen that the changes were supposed to fix, so I would prefer it without

For me it is a full 10-11 seconds before the music starts up.

The scrambled screen fix was not a part of v279-v280. It stays.

-ml

Ill second that confirm.
279 OK
280 No Ok...

Okay, now I'm (almost) as confused as poor Tony.

To make it easier to poke at the problem, I dug out a spare docking sled, and just wired up the yellow+orange to +12V and the black to GND. The aim here being to reproduce the in-car behaviour exactly, without having to wire-up the serial port behind the dashboard of my car.

Well.. no problemo. Works perfectly, in/out of standby.

Ugh.

To add fo the confusion - I think I see this here at work, in a docking sled with a dummy plug to put the player in AC mode. So I don't think it's a car-only issue.

It's fine at bootup, but if I put the unit into standby, then wake it again, the display doesn't appear for 2-3 seconds, then it appears with the battery icon for another 7-8 seconds.

No messages appear on the serial.

Like you, I have yellow and orange both tied to +12V and black to 0V. Amp control output, and the mute and lights inputs are all unconnected. Audio out and serial are through the docking connector, Ethernet via the socket.

Hijack v288.

Anything else you want to know?

Could it possibly have something to do with the difference between 1-drive-units and 2-drive units? Or the difference between Mk2 and Mk2a units?

For reference, I have two drives in my player, it's a Mk2, and I don't have the problem even when in the car.

Good thinking. I have the problem in the car.

Beta 13, Hijack 288, Mk. 2a, 2 drives.

Also have the problem in the car.

Beta 13, Hijack 288, Mk. 2a, 1 drive(30GB).

Dunno.

However, I did just notice that with the version I'm playing with now (no significant difference in the related parts, though), I only see the battery icon and delay when the engine is off.

In/out of standby with the engine running is just fine, thank-you.

So, I'm dropping the investigation.

On the off-chance that my slight tweaks since v288 may have fixed something, I'll leave them in for v289 (5-10 minutes from now).

Cheers

Just tried HJ 289. Still have the same behavior. No change.
Battery icon and delay with car running or with car off.

My empeg is a mk2a also. Single drive 20GB.

Z~

Well, just for good measure I just tried my second unit (2 drive mk2a) in the car, and it works fine with/without engine running.

So I'm done with it, at least until it shows up again on my own units.

Cheers

i have a 2drive mk2a with the prob

anyway thanx mark for trying to help solve the prob

I've had another look at this, and I don't think it's related to the drives at all. I tried it in the car this morning, putting the player to sleep in tuner mode, and when I woke it up again I got the mysterious behaviour - blank screen for about 2 seconds and battery icon for a further 7 or 8 but without the disk icon appearing. Since the car's engine was running, the disks weren't active, and I never used to see this, I doubt that the supply voltage really is too low (though I don't have my multimeter handy).

I wonder if it's possible that the code that checks the supply level is confused by something. I don't know how it works in either hardware or software, but is it possible that thread scheduling or interrupt masking (including servicing higher-priority interrupts) is somehow making it difficult for the player to get a correct reading for a period of time?

Something else I thought of:

I seem to remember that Mark does not use initial=xxxx in his config.ini, do you lot that are having problems? I know I am.

I just wonder if there is some delay due to the initial codes???

....Nope just tried removing that and it doesn't help. It would make sense that it is something to do with the button illumination. I don't have lit buttons. Anyone else?

For the record, my player has

• disk icons on
• notify on
• no initial command
• restore Car visuals on
• stock non-illuminated) buttons

Mark,

Do both your units have the button hack??

One of mine does and one doesnt.
Both display the same behavior in exactly the same way.

Z~

Nope, just one of them.

Still waiting for more button kits from Brian for the others.

Cheers

For the record, I completely disabled (in the source code) the buttonLED illumination code, and it made no difference for me.

Cheers

Has anyone solved this problem or had theirs just up and fix itself or anything? I'm still getting this and I miss the days where it would just come out of standby and immediately continue where I left off without having to wait a good 5 seconds or so for the screen to fade in, battery icon to pop on and finally for the sound to come through.

I am running Version 2 beta-13 with hijack 292.

Still happening here.
I choose to ignore it.
Well, as best I can, which is not so good.

You could run HJ 279...
That version for some reason doesnt cause this issue to occur.
Z~

How did I miss this thread? I was this close to stopping in the shop to have what I thought was a failling battery replaced. The bbs saved me the cost of a new optima ... drinks are on me!

You could run HJ 279... That version for some reason doesnt cause this issue to occur.

And you're absolutely sure that 279 prevents the problem 100 percent of the time, while 280 causes the problem 100 percent of the time?

And you're absolutely sure that 279 prevents the problem 100 percent of the time, while 280 causes the problem 100 percent of the time?

I admit that this is the result of limited benchtop testing, but I did run HJ 279 for quite a long time after this thread ran its course. Couple of weeks at least. Never saw the battery icon once. Put 292 on a couple of days ago when I found that weird crashing bug with the auto changing visuals.
Have had the icon regularly since.
Conclude what you will.

Z~

Yeah, I almost bought a new battery too.
My empeg actually made me start to doubt my multimeter.

Must've been jealous.

Z~

Still happening here, and I'm just putting up with it.

It annoys me briefly every time it happens, but as soon as music appears(*) I'm happy again. I'm just like Homer Simpson, really. Ooo, look - shiny!...

(*) Is "appear" the correct word for audio? Bitt?

I did stop by the auto shop and had my battery checked.. but it checked out fine so I was puzzled. Now I know otherwise..

Well, since you asked ....

``Appear'' usually refers to a physical manifestation. So, out of context, someone might think that ``music appearing'' might refer to the sheet music being delivered. But, in your context, I think we understand what you mean. And I can't come up with a drop-in replacement. The closest I can think of would be ``emanate'', but I think it would sound a little awkward without a qualifying adverbial phrase (that is, ``as soon as music emanates'' vs. ``as soon as music emanates from the empeg'' or ``from the speakers'').

This is certainly picking nits, though.

How about "plays" instead?

Calvin

Fine. Just come up with a better, obvious word, Mr. With-The-Logic.

What can be simpler and more obvious than "plays" ?

Sheet music "appears" but music "plays."

Calvin

You misunderstood. I was agreeing with you and trying to denigrate myself for not seeing such an obvious word. (Sarcasm doesn't come off well in print, huh?)

doh!

Speaking of words, have you read Kim Stanley Robinson's Mars trilogy? Now that's a piece of contemporary fiction that is packed to the gills with vocabulary, new words, and in very innovative usages as well. I don't have any specific examples, but was reading the last book in the series last night, and was fairly impressed by Robinson's very poetic prose.

Calvin

He's a good writer. I haven't read all of them yet -- I'm about halfway through the first one and was quite impressed, but as with most books that are reasonably challenging, I keep getting distracted, since it takes longer than just a few hours to read.

Edit: One thing in it that I keep thinking about is how he took the four classical Greek personality types and combined them with another set of classical personality types that I forget and came up with modern psychological types. Very interesting. I like stuff like that, and that was just a tangent (as far as I could tell from that point -- it might turn out to be very relevant, so don't spoil it ).

Typically, a good science fiction novel takes a single strong idea and expands upon it. What are the social implications, what are the effects on humanity of such and such? So his Mars trilogy is quite the shocker because due to its cornucopia of new ideas, even new characters, introduced every few chapters. It's quite the departure from the usual run of the mill. He has economic theory, political theory, string theory, geology, biology, psychology, genetics, terrorism, ecology, nuclear fusion, planetary exploration, space elevators, solar sails, immortality, religion, self replicating industries, moholes, viruses, ... and that doesn't even touch upon the incomprehensible scale of the imagery and plot. It takes skill to combine ingredients into coherent fiction, and amazingly enough, none of the ideas seem far beyond what is possible, and it would seem at no point was Robinson tempted to throw in alien life and tip the story into fantasy.

I'm going to have to end up reading the trilogy twice just to make sure I got it right .

He took the four classical greek personality types and mixed them up with other stuff, and in that particular instance, he is doing it to develop a character, to expose Michel Duval's method of thinking. It creates the sense of rationality, and then he switches over to another external point of view, and stomps all over the psychological exposition, calling it alchemy and pseudoscience and bullshit. It's disconcerting, but effective in building up the many characters have their own systems of thought, many times at odds with each other.

I'll probably finish the last book today, hopefully. :-D

Calvin

Trying to keep the original topic here alive. empeg taking a while to "fade in" from standby and showing battery icon for a small period of time. Here's hoping a new version of hijack will randomly fix this error the same was it was randomly implemented.

I agree. I am still waiting.......

Anyone else STILL getting this? I'm thinking of going back to an old old version of HiJack..

I still am. Mark even took a look at my empeg while at the meet but was not able to figure it out. I'm not so sure it is Hijack doing it and was thinking about trying to run it without. Have you tried that yet?

I've noticed that I also get this. Not every time though.

I'm still getting this too, as well as the battery icon during startup, delaying amp switch on. Am still on Hijack v293 mind you, due to the ftp server not behaving either

Mine happens everytime in the car - only w. hijack.

Same here.
Has been happening since I think version 278 or 279.
I did a test a while back in this thread. Too lazy to look it up now though.

Z~

I get it every time I come out of standby - on both kinds of power, on both radio and MP3 sources, playing or paused/muted.

I haven't investigated deeply enough to find the cause.

[bump]

as I see HJ is getting some work done to it, thought i'd just re-remention this to possibly get a little glanceover

I thought Mark had given up on this particular issue...?

Z~

This thread always haunts me. We never got it sorted out and I've been living with that battery icon since it happened. I doubt that so many of us could have been losing our minds. Neither Tony nor Mark were able to reproduce the problem (but it appears Mark had it for a little while) so even though they were trying to help, they were limited in what they could do.

I guess Mark saw this behaviour in Laura's empeg too.

Could it be something in the Mk2a hardware that "wore out"? Was it only appearing in Mk2as? Mk2a with 2 drives? Mk2as with over X amount of songs? Maybe certain brands or models of drives (or in combination) draw more power at spin up than others? Was it really HiJack or was it B13? Combination of the two? Maybe the low voltage sensor was not working before hand (giving false negetives)?

Because of all the new stuff that keeps appearing in HiJack and its refinement, I'm willing to live with this, but I'm still amazed that we've never been able to figure this out.

I originally started this thread way back in July and have been very dissapointed that it appears unresolvable.
Lately though I've been able to come to terms with my battery icon, and can honestly say:
"Icon, what icon...?"

Besides, I really cant see it through the little piece of black tape.
Z~

Besides, I really cant see it through the little piece of black tape.

Works wonders for my car's "Check Engine" light, too.

Just keep bumping this thread once a week or so, and when other things become less important than this issue I'll tackle it again.

Cheers

I don't think it's a case of any hardware wearing out, because I sometimes have a "good" period of a day or two every so often, when this effect doesn't happen.

I'm pretty sure it's not temperature-related, as I get it at home, at work, and in the car.

Perhaps communication between the kernel and the watchdog voltage sensor is flakey?

Both my 2-drive empegs suffer the same problem. It all started happening about 8 months ago. I don't know if it was tied to software upgrade or not. I have a relatively new car, new Optima yellow-top battery, and I am absolutely positive it is not related to voltage (13.4 volts according to tester). I have learned to live with the delay and battery icon, but I'm awfully curious as to why it happens.

weekly *bump*

bi-weekly *bump*

...something that goes *bump* in the night....

didnt see thid thread till now if got the same problem.I guess if it doesnt hurt anything ill just put up with it.and I to took it to the shop and they volt tested everything redid the pos and ground.lol I cant believe I did'nt see this thread.

just removed hijack from one of my players and it's not doing it anymore.just to make sure.

Yeah, its a suspected hijack bug (feature?) but apparently it only effects some players which makes it kind-a hard to pin down.
Mark said he'd poke around in hijacks guts again sometime if we bumped it to the top once in awhile.
For now its a "live with it" situation.

Z~

Mine is still doing it. Only happens when the drives are spinning up. I'm wondering if it's related to what drives you have and how much power they draw?

Is anyone with only 1 drive getting this?

I have 2 players both with one drive. 1 is a 10GB the other is a 20GB. They both show the icon coming off standby.

both of mine doing it with one drive.only when i turn on from stand bye,not when it boots.

I have had the battery icon come on several times especially in the morning. So since my vehicle has a standard wimpy 65 amp alt. I had it upgraded to 195 amp G3? and for some reason still get the batt icon but just not as often, mostly in the morning. You could be suffering from short trip driving and it's not charging up the system each time or have the basic wimpy factory alternator. I do have 1200, 300 and 50 watt amps drawing off mine the guy that upgraded my alt. said if I needed more than 195 amps then somethings wrong, so I'm going to upgrade my stereo wires to 0/1 gauge, that should do it. Good luck with yours, check your wires.

both of mine do it only when hijack is on them.

Hmmm I've never tried my player w/o Hijack on it, I love it too much. I'll try it on my spare to check mine as well I could also check, since I have both, MK2 & Mk2a for any differences.

yes thats why I put it back on.I'm getting used to the icon now.I was just worried that it might hurt my drives.

definitely not wiring or alternator output.
Happens in my VW GTI and my Jeep identically. Jeep has an extra heavy duty alternator.
Problem disappears with the removal of hijack.

Oooo.. an idea: go into the Hijack menu, and set the "Hard Disk Detection" to "One or Two Drives". .. any difference in behaviour?

-ml

I'm not allowed that option because HiJack detects that I have 2 drives, but maybe that will help the 1 drive peeps.

still doing it.thanks for the idea though.

Yeah, here too.

Oh well I'm keeping Hijack on my player, small price to pay for cutting edge excellence. I think some of what I experienced was alternator related for it doesn't happen as often. The player doesn't pass out at stoplights anymore, just the icon appears w/no break in the sound. It'll soon be resolved I'm sure.
These guys are good.

bumping this for what its worth

im still getting this problem... for those semi-new to the thread, it is not a wiring problem or a current draw issue with your vehicle, or a 1/2 drive thing. or an issue with a component 'wearing out' since it is rectified if you go back to hijack 279 or earlier.

so it is some 'kernel talking to voltage sensor' thingy that got changed and hasnt been tracked down, or something like that.

anyway.. learnt to live with it like the others here.

Oh yeah, I've forgotten.. has anyone seen this on a Mk2a(RioCar), as opposed to a Mk2(EmpegCar) unit?

EDIT: seems to happen on mk2a's, now that I've read backwards in this thread.

Thanks

Okay, help me out with a theory here:

Can you reproduce this issue when the DIMMER settings are BOTH at 100% on the player (set dimmer too 100% with lights off, and again with lights on).

Next, try it again with BOTH dimmer settings at 50%.

Thanks.

I was able to reproduce the issue in both instances.

Is button illumination set to OFF in the hijack menu (regardless of whether you actually have illuminated buttons..)?

uh... no.
Hold on...

Still get same same.

Okay, I'll put out v333 shortly, with the fake_tuner fixed (I think), and a possible fix for this particular problem..

Cheers

and a possible fix for this particular problem..

I'm dying to know what you think might be causing the problem.

Well, the only possibilities are the diffs (posted earlier) from v279 to v280, which added synchronous access control to threads trying to write to the display controller. I suspect this "enforced serialization" is slowing things on in-car players for certain combinations of player setttings.

So, Hijack v333 now only enforces the access control when Hijack itself wants to write to the display controller -- to do this, it still has to track the player's accesses, and force those to wait when Hijack is in there, and vice versa.

Maybe this will help.

Cheers

Interesting!

Well, hijack v333 is out there now. Any change?

Thanks for all your help and time on this.it's still doing it on both my players.it not a big deal it just stays on for a few seconds then goes away.Thanks agian.

Just thought I'd bump this to see if there is any new news. I am having the same problems.

Bump.

Has anyone tried wiring the player directly to the battery to fix this?

I had my whole system rewired and removed the tunner and it is not doing it any more.

Okay, I have managed to reproduce this condition again (finally!) with a Mk2a in a Home Dock. But only when an Empeg Tuner is attached (special dock!).

Unplug the tuner, and all is normal. Plug in the tuner, and we get the battery icon off of standby condition. Happens with at least two of my three players (haven't bothered trying the third).

So.. I've been poking at the theory that the voltage takes longer to rise when a tuner is attached, and so if I simply disable powerfail detection a bit longer than the default of about 100ms after resume (from standby), the problem ought to go away again.

A 2-second delay definitely works for me! Now I'll try shorter delays. I may also revamp the entire "delay" mechanism, as currently the code (from original Empeg kernels) uses a 100ms in-line busy-wait loop, with interrupts disabled. Much better, at least for the coming from standby case, would be to simply schedule_timeout(HZ/10) for the same approximate delay in a friendlier fashion.

I'll poke at it some more, and release a fix (for my players anyway) in Hijack v352 later today.

Cheers

Hijack v352 is now avaliable.

This version incorporates a fix for the "battery icon after standby" issue that several people have experienced at various times over the past two years. Hijack now waits at least 250ms after turning on the display (and radio module) before reenabling powerfail detection. Previously, it only waited about 100ms (standard Empeg kernel value), but with other changes that have been made over the years, that value was no longer adequate in all situations.

The patchset for this version is larger than the theoretical one-liner fix, because I also hacked it to not do an inline busy-wait loop when invoked from the player process -- it uses schedule_timeout() instead whenever feasible, which will make it somewhat "nicer" to any other processes/apps that one may be running on the player.

EDIT: Also in v352: I removed the option to "power off the display" in the screen blanker, as this was never really satisfactory under all conditions, and is of dubious value. It also conflicted with the battery icon fixes in some combinations.

v351, released earlier but without fanfare, includes a fix for a small bug in the kernel whereby it was misdetecting the headlight/dimmer sense value on startup. Fixed in v351 (and v352).

Cheers

Wow, Thanks Mark!

Sweet thanks Mark, I was just going to ask about that problem, when I got my tuner the other day I noticed it having the battery icon problem.

>Wow, Thanks Mark!

You are very welcome, Brad. And your patience with this has been amazing!

Cheers

For the curious, the problem is that when coming off of standby, the display must get powered on again. And the tuner gets power only/whenever the display is powered (same hardware supplies the juice).

So, when the display/tuner are powered on, there is a momentary dip in voltage from the power supply, which can trick the built-in software power-fail logic into thinking it has to do an emergency "save-to-flash" operation and whatnot, the first step of which is to poweroff the display.. leading to..

Well, you get the idea. Now, the original Empeg code included logic to deal with this, by disabling powerfail detection for 100ms after the display is powered on. This worked great until Hijack started messing with the display control logic, slowing down certain internal operations to make them more reliable when used with the extra hacks Hijack provides. In the end, 100ms was not long enough for whatever reason, ESPECIALLY with a tuner drawing current from the same lines.

So, v352 simply waits longer (250ms is what I used, could probably get by with less, but..) before turning the powerfail detection logic back on.

Cheers

And you thought we were all crazy!

Mmm.. now that I've looked at it even more, I'm not so sure that any of this logic actually does anything useful any more.

The code USED TO turn off the display logic, to buy more time for it to write the player state to FLASH on powerfail. But because turning the display off is no longer a simple operation (as of MK2A players), the code (from Empeg) doesn't bother with the display until AFTER saving the player's state.

So why bother turning the display off at all now?

It looks like I could delete a few hundred lines of C code from the kernel without adversely affecting anything.

Anyone got an opinion on this?

Cheers

Less is more right?

Less kernel is definitely more buffer space/time for music, especially so on Mk1/Mk2 players (vs. Mk2A).

Cheers

Hmmm.. Im having a strange problem now, I no longer get the battery icon but now when I leave the car for a while and then come back and start it the empeg boots up and then goes into standby. I know the empeg would go into standby if it was in standby when the car was shut off but now it goes into standby regardless. To make it even more strange if it is playing and I remove it from the sled and then put it back it reboots and plays fine, the problem only seems to happen if the car is off for a while. Wierd....

Hi, I was wondering if the empeg booting into standby could be related to the screen blanker being removed from hijack, I read another thread about this problem and it was said by one person that turning off the screen blanker would fix it. Also it may be related to the tuner becuase before I was having this problem it would take along time for the empeg to turn back on with the tuner attached.

OK, I have found that it is hijack that is causing the empeg to boot into standby, I put hijack 350 on it today and the standby problem went away but of couse the batt. icon came back. Any ideas Mark?

Nope, no ideas. I'm waiting for more than a single player installation to demonstrate this problem.

Cheers

Mark, on the "less is more" vein, are there any features that could be "handed off" to empeg for future software releases to free up some room in HiJack? Such as having Volume Adjust, Bass and Treble appear with Volume, Fade and Balance? Sorry if this has come up before.

Bass and treble are currently implemented in Hijack by... um... hijacking two bands of the equalizer. So this is not an optimal way of implementing bass and treble. The right way to do it would be to use the DSP's built-in bass and treble features. So, in theory, the empeg guys could add that in version 3.0-something if we kept asking them really nicely.

As far as the voladj is concerned... The Karma has an automatic gain control feature, and since the 3.0 empeg software shares a code base with the Karma, it is feasible that a future (improved) version of the Karma AGC might possibly appear in one of the carplayer 3.0 alphas.

The Karma has an automatic gain control feature, and since the 3.0 empeg software shares a code base with the Karma, it is feasible that a future (improved) version of the Karma AGC might possibly appear in one of the carplayer 3.0 alphas.

Like the one it has now?

It's in 3.0 now? Did it just appear in Alpha 5, or was it in Alpha 3 as well?

I have to confess, I have not installed alpha 5 because of the database rebuild problems. The empeg is my daily driver and I need to be able to add and remove tunes without hassle.

I haven't installed Alpha 5 (because of the rebuild problem for the same reasons), so yes, it's in Alpha 3.

Dang, I completely forgot about that.

Now that I really think hard about it, I now remember it from Alpha 3.

They say the memory is the first thing to go... what was the second thing?

Memory.

So.... once we get a more stable 3.0 beta that we can presume a majority of people will be using, Mark could remove the volume adjustment from HiJack and free up some code?

I always wondered what the relationship was between HiJack and the empeg software. Or rather, how the empeg guys felt about it. I know that most, if not all, seem to be using it. But would they consider "borrowing" some of its features or would that not be fair to Mark? How do things like this usually work?

>How do things like this usually work?

Totally fair. Except they probably don't want to copy line-for-line from Hijack, since all kernel code (including Hijack extensions) is GPL'd, and excessive copying from it to the player code might "taint" the closed-source player with GPL issues. But featurewise, feel free -- especially anything on the audio pathway like Bass, Treble, L/R Alignment, ...

The built-in VolAdj in the player seems to respond differently from that in Hijack, so the Hijack one stays put for now -- especially since Hijack is tuneable for one's own preference of response/attack rates. But if the player were to expose it's parameters for config.ini tuning, then we could nuke the Hijack version.

Cheers

Thanks Mark. That makes sense.

Does anyone know any keywords we have to put in this thing to get it seen? Does "Visuals" still work?

Visuals only worked for Toby.

Bass and Treble is something we've bugged the empeg guys many times in the past about, even on the private alpha testers' list, and they haven't yet added it to any version. Suffice it to say that they know we want it.

Regarding the difference between Hijack Voladj and the V3 AGC... John has some ideas about how he wants to improve it, and there are plans afoot to change it. It will still use a somewhat different algorithm than Hijack, so leaving Voladj in place would still make sense.

Tony,

Have you heard anything about ReplayGain support on the alpha board?

--Nathan

The empeg is my daily driver and I need to be able to add and remove tunes without hassle.

I've not added any tunes for a while (haven't bought any CDs recently) -- why would you want to remove tunes? -- so I'm still happily running v3.0-nightly20030815, which I installed when I was still at Rio.

I'm thinking about installing v3.0-alpha5, but I'm worried that it'll be less stable than what I've already got. Fortunately, I found the upgrade file I used originally, so I'm now safe to try it.

Personally, I'm planning on installing v2.0-final, uploading whatever I need to, and then installing v3.0-alpha5. Alternatively, if I'm using Hijack (which I've never yet done), I'll just keep different player binaries in the /empeg/bin directory.

I've not added any tunes for a while (haven't bought any CDs recently)

Well, in my case I'm always buying music. On top of that, I'm always listening to studio sessions (which the emepg is GREAT for), but it means constantly updating new reveisions.

why would you want to remove tunes?

I often borrow CDs from friends and upload the tunes onto my empeg. If I decide not to buy the CD, I'll delete the tracks off.

As for this whole stability thing, I really haven't had many problems with aplha 3. I've had to reboot it maybe twice in the last three months. I guess I'm just getting lucky. The only thing I wish were fixed is the crossfading TO wma files. Was that fixed in alpha 5? Even if it was, though, I still need to add/remove tunes.

Personally, I'm planning on installing v2.0-final, uploading whatever I need to, and then installing v3.0-alpha5.

I don't know for sure, but I seem to recall that this won't work. 2.0 and 3.0 use different database types, yet 3.0a5 won't rebuild databases. So if you're going to go "down" to add content, it's gotta be down to 3.0a3 rather than 3.0a5.

There is also the manual database rebuild that others have described, so you could look up how to do that.

Ah, bummer. For some reason, I was hoping that empeg towers had an Alpha 6 floating around with the database thing fixed.

I was hoping that empeg towers had an Alpha 6 floating around with the database thing fixed.

Even if there were, I wouldn't qualify for a copy -- I don't work there no more, remember?

I am on the alpha team, but we've had nothing since v3.0-alpha5.

Have you heard anything about ReplayGain support on the alpha board?

No..

Tony,

Thanks much for the link to the Riovolution board. I for one am very pleased to hear the news.

--Nathan

So if you're going to go "down" to add content, it's gotta be down to 3.0a3 rather than 3.0a5.

Yeah, that'll probably work. I've come up with a cunning scheme to make it easier to switch between versions:

1. Install v2.00-final, and back it up to the drive0 partition, using:

# tar cvfz drive0/empeg-v2.00-final.tar.gz empeg

2. Install a bunch more music, while it's running a stable version.
3. Install v3.00-alpha3, and back it up similarly.
4. Install v3.00-nightly, and (again) back it up.
5. Install v3.00-alpha5, and (you guessed it) back it up.
6. Install the latest Hijack.

This'll make it easy to move between releases without faffing about with upgrade files, since only the player and libraries have changed between those versions.

v3.00-nightly

Hmmm, interesting.....

Hmmm, interesting.....

Nah, not really -- it's just the last build that I took away with me when I left. It falls somewhere between -alpha3 and -alpha5, and is more stable than either. I couldn't be bothered to remember the actual date of it.

There is a nightly build script at empeg, but it does nothing more than ensure that the code builds. It wouldn't be sensible to release that into the wild.

In reply to:

---

I no longer get the battery icon but now when I leave the car for a while and then come back and start it the empeg boots up and then goes into standby. I know the empeg would go into standby if it was in standby when the car was shut off but now it goes into standby regardless. To make it even more strange if it is playing and I remove it from the sled and then put it back it reboots and plays fine, the problem only seems to happen if the car is off for a while.

---

I've seen this (not necessarily every time) for several months. It's not often that I leave the player in the car - normally I shut off the car and remove the player (often in the other order) and it's fine. Most times, it resumes as expected, but sometimes it boots into standby. Not often enough to bug me, in all fairness.

Anyway, I returned yesterday and upgraded Hijack from v340 to v362, so I'll look out for it occurring on the new one. I wonder if it's a general issue with saving state on power fail? (and therefore dependent on whether it was in standby on the previous shutdown)?