Mobius Super Capacitor.

[Isoprop]

@Isoprop any word on what's causing this? I have the standard battery and again yesterday, 10 mins into my journey, I looked up to see the a Mobius off. It has happened twice this week

Did it start properly? I use external (battery pack) power all the time and have NEVER had the camera turned off on me. However, I also have Power-off-Disconnect disabled.
How have you set the important params, i.e. from the GUI: Power On, Power On Delay, Power Off Standby, Power Off Disconnect, USB Data Port charging and Motion detect? These are the only parameters that should affect turning off the camera.
I can't rule out the possibility of a firmware glitch. It's all but impossible to check every possible combination so a precise description together with the configuration file is needed. If I can replicate your problem I can report it. If the camera only turns off sporadically, then it will be very difficult to find the cause. Since I don't use my camera as a dashcam I can only test it on the workbench.

 

[GJHS]

No, as with the others, it never turned on. It seems, as if sometimes it just doesn't turn on, especially if you stop at ACC before starting. I will send all info later when the camera is here.

 

[murcod]

I check mine is on (ie. operation of red flashing LED on the rear) every time and have never found it to have not turned on. My car does have a keyless push button start, so the Mobius goes straight from nil power to continuous power. However, I do sometimes start the engine from the "Accessories power" position (ie. Mobius already powered) with no problems.

My wife would never check the operation of hers.... I look when I'm in her car and have never seen it not operate. It did capture the close miss last week so was working fine that day.

 

[Gurkha9]

I guess that's why it's recommended to format the memory card from time to time no matter what cam you have.

 

[Isoprop]

Mine wasnt starting sometimes as I mentioned and got worse the last few days so it wouldnt start at all. I would get one red flash and thats it, but recently not even that. So went through all checks and variables.....

Turns out after clearing the memory card so the mobius wasnt having to recycle the footage (like it has on the 1.13 FW quite happily since it was released) it works fine now. So for some reason it couldnt manage on a full card even though it could before and no settings were changed. I will of course have to wait to fill the card to see if this happens again.

Sooner or later that will always happen in any device that uses FAT32. It's inevitable, I'm afraid. The FAT gets corrupted. However, if the FAT is intact, the camera should start normally if loop recording is enabled. That will be difficult to test....

 

[Isoprop]

External Power-On Auto Record

this problem with the camera stopping if power is applied and then stopped and applied again used to be quite common with dashcams (we had the same issue about 18 months ago), most have resolved it already, it's a firmware thing, they should be able to resolve it but will probably just need a bit of time with it

Since I couldn't replicate the start-up problem in my car, I've done some testing on the workbench.
What is most important is that the camera has power when the firmware has initialized. This is after approx. 5 seconds after applying power for the first time.
When power is applied, the firmware will start to load. It doesn't matter how many times you apply power during the first ~5 seconds, it's only the first time that counts.
Remember that the firmware is not running until it has done it's initializing stuff, like reading the parameters, checking the card, etc. When it's done it's stuff, it checks if there is external power. If there isn't, it will immediately turn off.
So, the magic word is 5 seconds. You must have stable power at the ~5 second mark.
I don't know how the developer is expected to resolve this. A small wait after the firmware is initialized will most likely influence the other start-up procedures, and I don't want to see it take even longer for the firmware to load!

 

[GJHS]

External Power-On Auto Record

Since I couldn't replicate the start-up problem in my car, I've done some testing on the workbench.
What is most important is that the camera has power when the firmware has initialized. This is after approx. 5 seconds after applying power for the first time.
When power is applied, the firmware will start to load. It doesn't matter how many times you apply power during the first ~5 seconds, it's only the first time that counts.
Remember that the firmware is not running until it has done it's initializing stuff, like reading the parameters, checking the card, etc. When it's done it's stuff, it checks if there is external power. If there isn't, it will immediately turn off.
So, the magic word is 5 seconds. You must have stable power at the ~5 second mark.
I don't know how the developer is expected to resolve this. A small wait after the firmware is initialized will most likely influence the other start-up procedures, and I don't want to see it take even longer for the firmware to load!

Now that I understand why and when it happens, for me it should be easy to avoid. In my situation it happens when I try to start the car without pushing in the clutch. The power turns on and then, because the car didn't start, I'm forced to turn the power off and on immediately. All within the first those few seconds. That would also make sense as to why sometimes when I'm sitting with the ACC on, listening to the radio for a while then I start the car there is no problem.
I really appreciate you taking the time to replicate the problem and I'm glad that you could find the problem. If your not getting paid for this, you should or free cameras for life. In every forum related to this camera, you are spending your free time helping to perfect it. I know I speak for all Mobius users with a big Thank you.

 

[Isoprop]

Now that I understand why and when it happens, for me it should be easy to avoid. In my situation it happens when I try to start the car without pushing in the clutch. The power turns on and then, because the car didn't start, I'm forced to turn the power off and on immediately. All within the first few seconds. That would also make sense as to why sometimes when I'm sitting with the ACC on, listening to the radio for a while then I start the car there is no problem.
I really appreciate you taking the time to replicate the problem and I'm glad that you could find the problem. If your not getting paid for this, you should or free cameras for life. In every forum related to this camera, you are spending your free time helping to perfect it. I know I speak for all Mobius users with a big Thank you.

Like Tom, I don't get paid! But I think most people know that already. We 'tinker' out of fascination in these tiny cameras and in the hope we can help others. It obviously helps tremendously that we have a direct relationship with the firmware developer. On the other hand the developer is continuously trying to tweak the firmware, so he deserves a big thank you too .

 

[Necromancer]

im not sure what happens.
after ~5 seconds, it checks for to see if there is power. if true -> it works.
if not -> camera shuts down.
if the camera shut down, and then power is applied, then it goes all over again, firmware starts to load, and after ~5 seconds it checks for power etc... so there is no problem there either.
so as far as i understand, the problem is only if the power is re-applied during the camera shutdown process.
is it not possible to add a check for power at the end of the shutdown? if true -> load firmware, if false, wait for power?

And i'll join GJHS with a big thank you for spending your valuable time helping us, it is very much appreciated.
and to the developer of course!

 

[Isoprop]

im not sure what happens.
after ~5 seconds, it checks for to see if there is power. if true -> it works.
if not -> camera shuts down.
if the camera shut down, and then power is applied, then it goes all over again, firmware starts to load, and after ~5 seconds it checks for power etc... so there is no problem there either.
so as far as i understand, the problem is only if the power is re-applied during the camera shutdown process.
is it not possible to add a check for power at the end of the shutdown? if true -> load firmware, if false, wait for power?

And i'll join GJHS with a big thank you for spending your valuable time helping us, it is very much appreciated.
and to the developer of course!

It's not as easy as that! While the camera's shutting down it's not possible to check for power, so there's a ~1 second gap there too. As soon as the firmware has properly shut down then everything is back to normal.
So basically there are two criteria:

1. There must be power after the firmware is initialized (~5 secs).
   
2. Power applied during the shutdown period (~1 sec) is ignored.

 

[Necromancer]

im not asking about checking for power during the shutdown.
check for power at the end of the shutdown process. add a check at the last "line" of the shutdown process, to see if there is power. if there is power, it means power was applied during the shutdown process. if so, re-load!
if the check returns false (no power), then complete the shutdown and return, waiting for power to wake up. i guess this way the time between this last check and standby would be a fraction of second. making this problem that much less likely to occur.

 

[Dashmellow]

I'm not really being 100% serious with this suggestion but it almost seems like the super capacitor needs a super capacitor. In other words, a brief back-up to the power going into the camera for the needed five second interval should the ignition switch cut out too soon.

 

[GJHS]

Thank you to the developer as well. I applaud him for listening to all potential users and adding features wherever possible that make this "action cam" great for many things and in many different situations.

 

[Isoprop]

im not asking about checking for power during the shutdown.
check for power at the end of the shutdown process. add a check at the last "line" of the shutdown process, to see if there is power. if there is power, it means power was applied during the shutdown process. if so, re-load!
if the check returns false (no power), then complete the shutdown and return, waiting for power to wake up. i guess this way the time between this last check and standby would be a fraction of second. making this problem that much less likely to occur.

However you do it in firmware it won't work! If you check in the last 'line', then the time between when the firmware was initialized and the 'last line' will still be 'dead' (~1sec.), so it basically boils down to the same gap of ~1sec. Even if you do the check more than once there will still be a possibility for a gap, actually many gaps.
There will always be a gap if the test is done by firmware alone. I don't know how external power is detected by the hardware, but I can imagine detecting to avoid any gaps is not possible with the existing hardware. Maybe all that's needed is an additional capacitor to keep the power interrupt high for a second or two, but I'm purely guessing here. Dashmellow, you may not be too far off here, but there's certainly no need for a supercap

 

[Dashmellow]

<snip> Maybe all that's needed is an additional capacitor to keep the power interrupt high for a second or two, but I'm purely guessing here. Dashmellow, you may not be too far off here, but there's certainly no need for a supercap

Thanks, Isoprop. I only have a rudimentary knowledge of electronic components but it just seems to be commonsense that there might be a simple technique for maintaining the initial power for a few seconds longer.

BTW, let me add my vote of thanks for your efforts.

 

[Necromancer]

Isoprop, i dont see the problem.
Just like you dont care if power is lost or regained during the firmware initiallization, this way it doen't matter if power is lost or regained between the time the firmware was initialized and the time the shutdown is re-checking for power one last time before going to standby.

Say the firmware is initialized, checks for power and doesnt find it -> shuts down. During the shutdown power was re applied, as the shutdown is almost done it probes for power again and finds it, and trigger the initialization process again. Whats the problem?
It doesnt matter what happens between the firmware initialization and the shutdown power check. It only matters what those checks find at the moment they are done.

If its not possible to simply check for power, for exaple if its just triggers when there is a change in voltage without knowing if power was restored or actually revoked, it just switches a flag from 0 to 1 for a brief moment (one iteration) and if thats not catched immediatly then its lost, in this case it would be a bit more complicated but not very mutch. Really depends on the algorithm thats used.
But....errr...thats not likely, since the firmware checks for power after initialization and is able to tell if power is connected or not, so i guess it is able to check at any given time if power is connected. So the first solution should work.

I dont understand why it wouldn't be possible to fix with the firmware. It seems to me you have all the tools needed.

 

[Isoprop]

Isoprop, i dont see the problem.
Just like you dont care if power is lost or regained during the firmware initiallization, this way it doen't matter if power is lost or regained between the time the firmware was initialized and the time the shutdown is re-checking for power one last time before going to standby.

Say the firmware is initialized, checks for power and doesnt find it -> shuts down. During the shutdown power was re applied, as the shutdown is almost done it probes for power again and finds it, and trigger the initialization process again. Whats the problem?
It doesnt matter what happens between the firmware initialization and the shutdown power check. It only matters what those checks find at the moment they are done.

If its not possible to simply check for power, for exaple if its just triggers when there is a change in voltage without knowing if power was restored or actually revoked, it just switches a flag from 0 to 1 for a brief moment (one iteration) and if thats not catched immediatly then its lost, in this case it would be a bit more complicated but not very mutch. Really depends on the algorithm thats used.
But....errr...thats not likely, since the firmware checks for power after initialization and is able to tell if power is connected or not, so i guess it is able to check at any given time if power is connected. So the first solution should work.

I dont understand why it wouldn't be possible to fix with the firmware. It seems to me you have all the tools needed.

Assuming the firmware takes one second to power down, this is the time within which you must check for voltage. Of course you could use all sorts of 'tricks' to flag the power condition, but there will always be a time gap where you can't do any proper checking, for example when the DSP runs thru it's routines. I also very much doubt if you can interrupt the shutdown procedure and just tell the firmware to restart. It's the same on your computer. Once you click on 'power down' you can't just press the power button and hope the computer will restart. You have to wait until your PC has turned off.
IMO, this is definitely something that has to be addressed at the hardware level.

 

[Necromancer]

but there will always be a time gap where you can't do any proper checking

You dont have to do proper checking all the.time, you can make a check after this "gap" and based on the result conclude what happened during the gap. Thats thats whole point.

 

[Isoprop]

You dont have to do proper checking all the.time, you can make a check after this "gap" and based on the result conclude what happened during the gap. Thats thats whole point.

After the 'gap' the firmware is no longer running or it's passed control to the DSP! That's why I keep saying it needs a hardware solution. A small capacitor to keep the interrupt pin pulled high, for example.
Anyway, for the current version it's important that the camera has stable power after max. 5 secs.

 

[Isoprop]

Anyway whilst you two are squabbling over programming the FW I want to report that both my Mobius (A and B, one being a super cap and the other with a battery) are starting to not start after ignition once more. They were after each had a card format and working well, but that only seemed to last two days.

Turn key (wait 5 seconds which I always did anyway as car goes through checks) mobius starts, then turn over engine and all stops again. I have to then pull the power to the mobius, which was working but as before works less and less until it wont turn on at all. Seems to be in some sort of degrading cycle. Both my Mobius are on the later FW 1.13

So again, been leaving unplugged until car started and them plugging in USB to car lighter, but that is now not starting the Mobius each time either. This has now been duplicated in the house with 1amp Mains USB charger but only after the symptoms have shown in the car itself.

Sounds like you could have a plug contact problem. If you have a spare USB socket, an LED and a ~300Ω resistor you could make yourself a very effective test lead which will show you exactly when you have power.
Do you have power-off disconnect set to 10 seconds or disabled?
I don't exactly understand how you duplicated the problem in the house. I think the ~5 sec. to ~6 sec. marks is the key. If you apply power during this shutdown interval the camera will not power up.
For the car starting, waiting 5 seconds is either too long or too short. Try waiting 3-4 seconds or 7-8 seconds.