Battery Life compromise.


My plan wasn't to transmit heat to the back plate - but to have a gap in the back plate with exposed fins between two batteries - a trough full of fins so to speak.
Ah thanks for the clarification. That design would keep the heat away from hands (which could get uncomfortable with my design).
 
My  plan wasn't to transmit heat to the back plate - but to have a gap in the back plate with exposed fins between two batteries - a trough full of fins so to speak.
Dust is the least of our worries. We don't have active air flow channels in pandora 2 (unless they can somehow fit a fan inside the battery area), so we need to have a large area for conducting and disposing heat. 

A few fins will not be good enough if they are not actively cooled. The only way to go is transmit the heat to some big metal panel on the outside. 

We can either put the main pcb on the back of the screen so heats can get out easily, or we can put the batter beween the keyboard and the pcb, with the pcb at the bottom. 

However, more issues arises, as the first one will make the upper part super thick, and uncomfortable to hold, the second will make changing batteries super difficult. ( some sort of removable keyboard may help, like the ones from TI's calculators)

84comp.jpg
 
[…]


A few fins will not be good enough if they are not actively cooled. The only way to go is transmit the heat to some big metal panel on the outside. 


[…]
You know, a big metal panel won't cool better than a finned metal panel.


The main aim of fins in heat dissipation is to have a larger surface in contact with the air, it gives a better cooling even without fans  ;)

However, more issues arises, as the first one will make the upper part super thick, and uncomfortable to hold, the second will make changing batteries super difficult. ( some sort of removable keyboard may help, like the ones from TI's calculators)
Removable keyboard will only add more parts and cost. I don't think it's a good idea.
 
Last edited by a moderator:
Also, the motherboard is an integral part of the keyboard, so making it removable would need a second PCB
 
I'm probably insane for suggesting this - and the costs may prohibit it's consideration - but why not a Silicone-based Thermal Pad?  Not as an entire case...but maybe as a part of the case where the heat needs to be dissipated the most?  Look, I'm not no engineer, and I know very little about thermo-dynamics...but maybe?  I do know that when I put my hand in a fire it hurts.  Also, a cold beer can on the hand helps....
 
I'm probably insane for suggesting this - and the costs may prohibit it's consideration - but why not a Silicone-based Thermal Pad? Not as an entire case...but maybe as a part of the case where the heat needs to be dissipated the most? Look, I'm not no engineer, and I know very little about thermo-dynamics...but maybe? I do know that when I put my hand in a fire it hurts. Also, a cold beer can on the hand helps....
You can't make a case out of thermal grease. If you're talking about silicone rubber / silicone caulk, that does not have a high thermal conductivity.
 
Yeah, I know...it's not really rigid in itself, but maybe as part of a heat vent?  Look, I did preface that last post with "I'm probably insane"...

Obviously metal heat sinks are better....I just was wondering what people thought.


Edit 1 :  Okay, scratch that....the "pad" that I was looking at has film on both sides...duh.  Now I invite the tomatoes - throw em at me... (winces).
 
Last edited by a moderator:
It's just that I have this GK802 quad-core "mini-pc" stick...and I swear that I read somewhere that the back of the unit is used as a heat sink and was mentioned that it was silicone-based.  I'm holding it right now and the whole thing feels metallic, so maybe it was a marketing ploy, or exaggeration, or maybe the "chinese" sources just meant that it was a spot of paste/film touching the CPU and the back panel.      

I should have known that I hadn't stumbled upon something new if you all haven't heard about it.   ;)  

BTW, I'm wearing my safety goggles...I hate tomato-paste in my eyes...
 
If cooling is going to be a problem, then power consumption is too high.

The P2 should have lower SoC power consumption for low to medium loads. Only for high loads the power consumption is probably going to be significantly higher, but that is only a problem if those high loads are sustained for an extended period of time. I'd prefer a solution with a thermometer that automatically scales down the clock speed or even shuts down entire cores if things get too hot. Wasting lots of valuable space and weight on cooling just to be able to sustain a cpu-burning load for a longer time doesn't sound like a good idea to me.

In nearly all practical uses, the big loads will come in short bursts, which is not a problem at all. Unless you all want to use your P2 for bitcoin mining or similar multi-core number crunching.
 
^  I think that "we" (i.e. not me, but developers) could include those type of instructions into the next Super Zaxxon if it was deemed necessary, but it seems like this current soc is almost dummy-proof.  I can't overheat or kill it if I tried....well, I haven't tried that hard.  BTW, your system info program is invaluable.

I think that people are just hypothesizing here - what if the OP Team does move onto a more power-hungry cpu/gpu combination?  What power might it draw, what heat might it produce...some of us are still in the early stages of thinking along these lines (me).  I'm learning...and I've realized that reading posts upon occasion is far different from interacting with the same people.  I'm humbled by my own stupidity sometimes.  Also, I'm happy and hopeful...it means that I'm among some really great minds and a great community.
 
If cooling is going to be a problem, then power consumption is too high.
Agreed.  My suggestion of having a 'finned heat trough' on the bottom centers around the idea of needing some additional thermal release for the system.  It isn't about having a 4" tall 3 Kilo Copper fin stack with a 120mm fan on top.  Just a passive radiator and thermal distribution 'plate'.  The whole distribution plate & radiator could be made out of a single piece of 24ga aluminum sheet that has been folded 12 or so times in the middle to form fins.  That could get us 2-3x the thermal 'cap' that we'd have with a closed case design.

Yes, that DOES mean that the system would be able to burn through it's battery 2-3x faster when running under full load to full load.  So?  Battery life measurements should be done by task - not by 'how fast can we discharge it with everything maxed out'.  One measures practical expectation - the other measures minimum.

The biggest drawback to the 'dissipation plate with finned radiator in the middle' is likely that it would require a split two battery system - which isn't such a bad idea.  Just pick an 'as of design moment' very popular phone (Galaxy S4 maybe?) and fit it for two of it's batteries - one on each side.  Then use the knock off batteries that are sure to flood the market and be easy to source for years to come.

The finned plate isn't a 'must do'.  It is an alternative if the team finds an SoC that is perfect in every way - but it has the potential to create 'too much' heat under extended full load.
 
Why put the fins in the middle.

Metals are generally good enough thermal conductors that your could put the fins towards the side and have one large battery.

Having said that, I do quite like the idea of two batteries, as them you could swap them without loosing power, and you could use one of the bays for a hacking project but running some extra wires into it from the mainboard.

I'd also suggest getting batteries made specifically for the P2, even if they are the same as <popular brand smartphone> just because Chinese knockoffs are probably rather poor quality.
 
Why are people so afraid to lose power? If there's a functional hibernate, especially one you could get into by holding down some button, then does it matter if you have to power down? Dual batteries complicates the design in a lot of ways just for that one feature. It also means you have to buy and carry spares in pairs in order to get the same extra capacity.
 
^ Agreed.  I don't think that the P2 would need to have more than one battery.  "Hot swappable" batteries aren't really necessary if you're cognizant of what battery life you have left and and save any work before powering down.  If this were really to be an issue with prospective customers I guess you could design a P2 PCB with a capacitor that would keep a charge for 15-30 seconds (?) in order to swap batteries without losing power, but that seems a bit overkill - and would take up some precious board space...
 
 I guess you could design a P2 PCB with a capacitor
Or a button-cell.

I think most people would rather the extra board space were used for something other than complex power supply setups.
 
I guess that a button cell could serve dual purposes in that scenario - as a temporary power backup and to power a real-time clock.  Still, (like you said) it'd take up valuable space on what's already going to be a crowded PCB.

BTW, my Pandora's time jumps allover the place with the current configuration - not that I really need an accurate clock, mind you.   ;)
 
Why are people so afraid to lose power? If there's a functional hibernate, especially one you could get into by holding down some button, then does it matter if you have to power down? Dual batteries complicates the design in a lot of ways just for that one feature. It also means you have to buy and carry spares in pairs in order to get the same extra capacity.
I'm not concerned so much with loosing power.  

Using a readily available cell phone battery would shorten one element of design and be one less 'proprietary' item in the system.

Using two of them doesn't mean having to carry spares in 2s - or even charging them both at once.

Much like our software handles the charging of one battery now - just with two batteries.  The logic doesn't have to be difficult.

When unplugged:  Discharge A to 5%, check B - if B>5%, switch to to B, discharge to 5%, check A.  If both batteries <=5% hibernate.

When plugged in, charge and discharge from the same battery.  Check A, if A <=95% then Charge A to 100%, switch to B, check B <= 95% then Charge to 100%, if both >= 95% then don't charge any.  Check back in 20 minutes.

There would also need to be some logic for 'battery ejection' so that it switches to the other battery if one is to be removed.

You -could- in that scenario carry none, one, two, three or whatever spare batteries - they don't have to be used in pairs as they're individually charged/discharged.

A person could also de-weight their Pandora by one battery and go single-battery if carrying the extra 6oz is too much for them.

Yes, that adds a bit of software complexity over the single proprietary battery solution we have now - but where will we get a Pandora 1 battery 5 years from now?
 
What I said is that you have to carry spares in twos if you want the same extra capacity. That last part is important.

What you're describing adds substantial hardware complexity. You need the capability to switch in hardware which battery is supplying power and the ability to monitor both the one you're powering and the one you aren't without switching power to it. You don't automatically get these things in software.

There's no de-facto standard in the world of cell phone batteries, you're just trading in one proprietary solution for someone else's proprietary solution. Sure, if you pick the design in a really popular phone it'll be supported for a while and there will be after-market solutions. But there's no guarantee that you'll be able to get a convenient supply for years down the road. I doubt it's any easier to get batteries for a 10 year old phone than it will be to get a Pandora battery five years from now; in fact, I have a good feeling ED is more likely to support the latter than anyone the former.

There's also the issue that no cell phone battery in pairs is going to be ideally sized for the console like a custom one would be.
 
Last edited by a moderator:
where will we get a Pandora 1 battery 5 years from now?
 
ED's shop :p

I'd be much happier if there was a truly standardised battery size that made any sense.
 
Last edited by a moderator:
Back
Top