PSP Hardware hack and development -- External Battery Pack

[Kamilion]

Hey folks, I've been doing some poking around, and I found this.

http://www.maxim-ic.com/quick_view2.cfm/qv_pk/1831

It's a maxim buck boost regulator that outputs at 5 volts, ~1.3-1.5 amps and can be driven by voltages as low as 0.9 volts (0.7 volts not guaranteed, but it does allow the batteries to drop this low before cutting out)

My problem is this -- it says it's a 1 to 3 cell unit, and I want to make a parallel battery pack with 4 AA batteries or 4 D batteries. NiMH.

The AAs are energizer 2500mAh 1.2 volt ones, for use with the 15 minute charger.

Any electrical engineers around?

I'm wanting to know if I'll have to use a two cell pack instead of a 4 cell pack.

I'm pretty sure this puppy would be able to power the PSP with a fully charged battery, and possibly even be able to charge the battery (the AC adapter is 5.35V 2 amp output, although I havn't measured the current drain and voltage drop while the PSP is running so far.)

The big deal is, I need overcurrent protection to keep the IC relatively cool.

Anyone willing to give me a hand with this project?

I've got a copy of eagle and I can kick up some PCBs to mount this on, but I was intending to use the MAX1703EVKIT evaluation board to start with.

If this works out well with a long pack life, I may consider mass producing and wave soldering these bad boys. Would anyone be interested?

My other question is: Theoretically, how many batteries can I hook up in parallel with this chip without damaging it? I'm assuming from my limited electrical knowlege that I would be able to hook up a great many batteries to this, and the load would be distributed over all the cells, using less amperage each cell.

Basically, my idea is thus:

The board will be encased in a enclosure, and have a PSP power cable coming out one end, and on the bottom, widened 9V snap-style connectors.

Each pack would have the widened 9V snap-style connectors on the top and the bottom, and everything is hooked in parallel fashion.

The more runtime you want, the more cells you snap in. Snap the first battery pack to the buckboost enclosure, then snap another pack onto that and so on and so on (This isn't really nessicary if using D batteries, because one can obtain 7000mAh D-size NiMH batteries for some crazy-long runtimes.)

I've already seen pelican has a four AAA battery box that pretty much connects directly to the PSP's external power connector, although I have no idea if it's regulated in any way or not, since it appears to be able to use aklaline batteries, which output at 1.5V, for a total of 6 volts (if fully peaked, more like 6.4V) which SEEM to be directly connected to the PSP connector (although since I do not have one, I cannot confirm this... Does anyone have this? Comes with the starter kit and the power kit... Be nice if someone could pop the case open and see how it's wired...)

So there's pretty much all I've been able to figure out so far. Any comments, suggestions, electrical engineers?

(Note, I would have posted this in general, but it seems to be locked... *shrugs*)

[fireether]

I'm a computer engineer, which is 1/2 electrical engineer and 1/2 computer scientist.

My other question is: Theoretically, how many batteries can I hook up in parallel with this chip without damaging it? I'm assuming from my limited electrical knowlege that I would be able to hook up a great many batteries to this, and the load would be distributed over all the cells, using less amperage each cell.

As many as you want. Ohms law is E=IR, voltage equals current times resistance. In other words, view the PSP as a resistance, and it will draw a certain amount of current based on the current voltage. However, the voltage is not variable, while the current will vary (more current when UMD is in use then when its not).

Having more and more batteries in parallel increases the maximium current output those batteries can produce over the average lifespan. In other words, (as an example) lets say 1 AA battery lasts 1 hour at 1 amp/hour. Two AA batteries would last 2 hours at the same current, or 1 hour at 2 amp/hour. And so forth. So it doesnt matter how many batteries you have in parallel, they will never produce more current than the PSP will draw. However, there are ways for devices to draw alot of current and then have their resistance lower and draw even more current (laser diodes come to mind).

For the PSP, I would say that using a stepup converter may not be that efficient.. Power equals current times voltage.. Or Current*Voltage = Current*Voltage.. If you take 2v and convert it to 5v, there will be more current drawn at 2v than at 5v. So as the batteries die, more and more current will be drawn from the batteries as their voltage drop, which can basically fry the IC, and makes the batteries die even faster.

The AC adapter for the PSP is rated at 2A, but often adapters are rated beyond the worst case drain. There may be a maximum of 1.5A, but a rating of 2.0A keeps people's adapters from going up in smoke. The best bet is to get a ampmeter and to measure how much current the PSP takes in several modes: playing game, watching movie, idling, copying files via USB, sleep mode, etc. That would give you a good idea of what you will need. However, if the PSP draws more than 1A, I wouldnt use a chip rated at 1.5A - it'll get pretty hot and will require a heatsink.

Definately get a second opinion, but in my opinion, its not very efficient and theres some problems (heating, current) with what you've proposed. But your idea has some good points - if you can find a voltage doubler that supports higher current, I'd be happy to work with you on a circuit schematic.

[darkfader]

you mean U=I*R :) E=energy or something

[Ihsan]

E is sometimes used for voltage, as It can be used U or V.

[fireether]

you mean U=I*R :) E=energy or something

E is sometimes used for voltage, as It can be used U or V.

It depends on where you learn electronics. E stands for voltage, as in E=IR where I learnt it.. But can also use V=IR.. V=CR.. Because I = current.. Etc.

Really the same thing, as long you know what the law is (Ohms) and know what each part of it is (voltage = current times resistance).

I know that E = energy in E=MC^2, which is very different from above. :)
Theres also how many electrons in a metal conductor, etc.. But thats going off the point.

[alfaromeo272]

you mean U=I*R :) E=energy or something

It depends on where you learn electronics. E stands for voltage, as in E=IR where I learnt it.. But can also use V=IR.. V=CR.. Because I = current.. Etc.

Really the same thing, as long you know what the law is (Ohms) and know what each part of it is (voltage = current times resistance).

I know that E = energy in E=MC^2, which is very different from above. :)
Theres also how many electrons in a metal conductor, etc.. But thats going off the point.

Yeah guys, best action for all you is to sweep floors but not to develop or hack something.

This microchip has sample scheme in it's documentation how to use it properly. If you're not able to repeat it...shame on you.
And don't try to remember physics from your school it's lame.

For all trying to create something - RTFM before writing

[fireether]

you mean U=I*R :) E=energy or something

It depends on where you learn electronics. E stands for voltage, as in E=IR where I learnt it.. But can also use V=IR.. V=CR.. Because I = current.. Etc.

Really the same thing, as long you know what the law is (Ohms) and know what each part of it is (voltage = current times resistance).

I know that E = energy in E=MC^2, which is very different from above. :)
Theres also how many electrons in a metal conductor, etc.. But thats going off the point.

Yeah guys, best action for all you is to sweep floors but not to develop or hack something.

This microchip has sample scheme in it's documentation how to use it properly. If you're not able to repeat it...shame on you.
And don't try to remember physics from your school it's lame.

For all trying to create something - RTFM before writing

Alfa, your point about physics being? :) As a computer engineer, I must know ohm's law, and thats not only from physics, but also from electrical engineering.

Also, guys refers to more than one person, so I was wondering who you were refering to.

My point was that while its possible, if the PSP draws too much current, its not feasible to use this particular chip because if its limit of 1.5 A.

[alfaromeo272]

Alfa, your point about physics being? :) As a computer engineer, I must know ohm's law, and thats not only from physics, but also from electrical engineering.

Also, guys refers to more than one person, so I was wondering who you were refering to.

My point was that while its possible, if the PSP draws too much current, its not feasible to use this particular chip because if its limit of 1.5 A.

You should know that Ohm's law has several variations depending on circuits - inductive, capacitor or mixed. Resistance can be active or reactive and so one.

PSP cannot consume more than 1.5A cause accumulator is 1800mA and it works from it not less than 3 hours. So maximum consuming current is no more than 600mA (for 222MHz CPU for now, I don't know about 333MHz).

It's elementary.

[fireether]

You should know that Ohm's law has several variations depending on circuits - inductive, capacitor or mixed. Resistance can be active or reactive and so one..

From what I was taught, thats if calculus is used. In electronics, I was taught that its purely E=IR. Circuits can be capacitive, resistive, and I never said that the PSP isnt one, the other, or a mix of the two.

The way I was taught to analyze a circuit was to take the circuit diagram, and then use KVL or KCL and find the voltage and current of each part of the network.. Which is a huge pain in the ass for very large circuits, like the PSP.

The point is, I never said that the PSP was totally resistive or capactive.. And the best way to measure the amount of current the PSP consumes is to hook up an ampmeter to it for a long period of time and get a number of measurements (20, 30, or more) and use those.

PSP cannot consume more than 1.5A cause accumulator is 1800mA and it works from it not less than 3 hours. So maximum consuming current is no more than 600mA (for 222MHz CPU for now, I don't know about 333MHz).

It's elementary.

So did you measure the current and get the average current?

All I was saying before is that IF the current in the PSP goes to 1.5A or more for a length of time, that can potentionally overheat the microchip.

Ive used voltage regulators before (1.5A max) and those have a metal tab for a heatsink - and I've gotten them reallly hot on ~1A before. So using that experience and applying it to the PSP, I'm concerned that if the current is very close to the cut off current for any length of time - it can blow out the voltage doubler. Plus, if this is the case, I dont want somebody to build 20 units, and then they will blow out minutes - its a waste of time and money. If they build one unit and test it, and if it doesnt work, that should be all the answer they need.

Last but not least, most datasheets for IC's have examples in them, if not all. Theres tons of circuit information out there, and I agree that the person building the doubler could've researched more. But I was just adding in my 2bits.

I'm not going to reply to this post anymore, because its not related to development, and I rather not contribute to a post & bump it up by result of replying to it if its not related. :) No offense to anybody.

[Kamilion]

I've done a lot of research on this, and you both raise good points. I tried getting my crappy little multimeter inline as a DC Ammeter, but no matter what I did, I couldn't get the PSP to turn on while it was in-circuit.
But chalk that up to $5 equipment from harbor freight.

alfaromeo, you forget that the 5V charge input is not running as the same as the lithium ion 3.6V 1800mAh pack inside the PSP.
Through the 5V charge port, you're not only running the charger for the battery, but the PSP unit as well, at the same time, causing a higher current load.


"For all trying to create something - RTFM before writing"
Okay, where's the manpage for this? Do I have to emerge it first? :P
Electronics and circuit theory are complex subjects, there is no single manual for it that I've found. If I had a camera, I'd show you the vast array of books behind me here in the office.


Anyway, change of plans. I had 4 NiMHs hooked up serially to the charge port. (Energizer 15 minute Rechargable NiMH, 2500mAh each, 1.2V nominal)

Here's what I found out.
When the batteries are at a full charge, they put out about 5.2 volts. Under no load, the PSP's AC adapter puts out 5.35 volts.
When the pack is freshly charged, and checked at 5.1v under load on the PSP, both the PSP is powered and the internal battery charges for about 30m-1H max. At that point, the voltage has dropped to 5.05v under load.
Now the PSP's power light is green, and the battery is not charging, even though it is not full.
At this point testing was being run by Wipeout Pure running it's browser, connected to PSPIRC at LCD backlight brightness 4 (AC mode) with the speakers at full blast, communicating every couple seconds over WiFi with Wifi power save mode off, display dim off, and auto-sleep off.

When wipeout pure was quit, the light started flickering between green and orange. There it sat in the browser until about 3AM. (The test was started under load from wipeout at 6:25:01PM when it switched from external power to 'in use', indicating the voltage fell too low to continue running from external power.) At this point, the internal battery was at 90% and the external pack's voltage had dipped below 4.82 volts until the PSP switched to internal, at which point the voltage rose to 4.87V I shut off the PSP and left the battery pack plugged in, and the charge light orange. When I woke up, about 8 hours later, the light was off and the charge on the battery was 93%.

So, just as these batteries get into nominal voltage range, it can't use them anymore. All the 5 volt regulators I've looked at require a higher input voltage (like 10 volts).

Also, to get the cable I needed to plug into the PSP, I bought a pelican car adapter, popped open the casing and desoldered the cable, and soldered a 9V snap lead instead. Inside, the board says: 1.5A max limit.
When the PSP runs on this from a 12V gel cell battery, it refuses to charge as well unless it's all the way off.

So, some kind of regulator is nessicary. I've revised my plans, intending to use the MAX1703 with 4 NiMHs serially at 5.25V Max 4.8V nominal input voltage (Slightly under it's 5.5 volt voltage limit.)

Your point on heat is well taken, I'll likely mount the unit in a aluminum enclosure with a beltclip/penclip unit for heat dissapation.
From reading the specs closely, the MAX1703 can peak at 3.3 amps and run continuously at 1.5A.

With the input voltage closely resembling the output voltage, it should be pretty efficiant.

I'll probably protoboard it so I can measure voltage and amperage through the whole circuit.

Which reminds me, I should pick up some thermocouples so I can measure the heat as well.

Thanks for the information, sorry I'm a little late in getting back, but I was on greyhound.

Edit:

Here's a little background on why I want to do this.
Eventually, we'll have applications and other software running on the PSP besides games. Sometimes, you want some extended runtimes.
Such as on an extended plane or bus trip or for instance, camping. I intend to pick up the GPS addon as soon as it's avilable.

The other plus of this, is much much faster recharging.
Four AAs can fit in the energizer 15 minute charger, and recharge fully in about 20 minutes. That's a hell of a lot faster than the PSP takes to fully recharge it's cells (2 hours and ~10 minutes from 0% to 100% for me, while fully off)

So, I don't have any of the software or skill to hack at the PSP, next best thing is hacking at the hardware instead until someone gets a foothold in the software.

Anyway, one last thing. If nessicary, I can use two MAX1703s to provide a higher output current. Plus it can be set up to use it's uncomitted comparator to provide a low battery signal.

[fireether]

You have a great idea :)

Definately prototype it first.

You could also measure the current into the AC adaptor (be careful) - but keep in mind that you will end up with higher current than the PSP is taking - due to some slight energy loss across the transformer, loss due to heat, etc. Its best to try and measure the current into the PSP if possible.

1.5A continous and 3.3A max would work, as long as it doesnt go above 1.5A while the battery is charging..

Another idea for extended time is instead of plugging it into the normal 5V socket, why dont you make a fake battery that plugs into where the normal battery does, and then that can be attached to your circuit? The PSP battery is 3.6V at 1800 mAh. This way if the normal PSP battery is a little low, you wont lose alot of external battery power by charging the internal battery, plus energy loss through the battery charger heating up a bit (if at all), etc..

Its more complicated that way than your original way, but its something to think about.

I look forward to any updates on this :)

[ChaosKnight]

Or just get a second battery...

[Kamilion]

Well, that's a bigger problem. The PSP uses a smartbattery, that either talks about it's tempature, or perhaps is more like my laptop, which has a whole slew of data about it's battery.

GiMPAQ BAT0 # ls --color -l -a
total 0
dr-xr-xr-x 2 root root 0 May 4 12:49 .
dr-xr-xr-x 3 root root 0 May 4 12:49 ..
-rw-r--r-- 1 root root 0 May 4 12:50 alarm
-r--r--r-- 1 root root 0 May 4 12:50 info
-r--r--r-- 1 root root 0 May 4 12:50 state
GiMPAQ BAT0 # cat alarm
alarm: unsupported
GiMPAQ BAT0 # cat state
present: yes
capacity state: ok
charging state: charged
present rate: 0 mA
remaining capacity: 2003 mAh
present voltage: 16650 mV
GiMPAQ BAT0 # cat info
present: yes
design capacity: 3093 mAh
last full capacity: 2003 mAh
battery technology: rechargeable
design voltage: 65533 mV
design capacity warning: 300 mAh
design capacity low: 0 mAh
capacity granularity 1: 100 mAh
capacity granularity 2: 100 mAh
model number: 3200
serial number: 00026022
battery type: 4C 69 00 00 00
OEM info: COMPAQ
GiMPAQ BAT0 # uname -a
Linux GiMPAQ 2.6.10-gentoo-r4 #1 Fri Jan 14 15:16:22 PST 2005 i586 AMD-K6(tm) 3D processor AuthenticAMD GNU/Linux
GiMPAQ BAT0 #

A battery that just provided voltage probably wouldn't work. This thing is deisnged and built to run and charge lithium ion batteries. It's got a power controller IC that likely talks to the battery about it's state to provide the OS with info such as % and Time remaining. It probably at least passes on the voltage and approxamate charge of the battery.

But the thing is, it's almost promised to be some kind of a serial interface, most likely a 1 wire interface. It doesn't have to be fast.

Sony has some 2250mAh internal batteries planned. The PSP would have to know the difference between them in order to still display the correct % and time remaining. Once they release those, I'll buy one of them to replace my current internal battery, and keep the original charged, and continue to use my external battery system. With the higher power batteries, chances are charge times would only increase (perhaps 3 hours for 2250mAh, give or take 20 minutes) and keeping the battery topped up from a couple AAs is invaluable on long trips. AAs are easy to replace.

But remember, sleep mode doesn't hold if you take the battery out.

So swapping batteries isn't much of an option unless you just saved at the 2% remaining mark a minute ago

The other nice thing about my experiment is this.

The PSP ran with the display at AC mode from 6:25PM to 3:00AM. That's almost 8 hours.

I'll have to do another test with it sitting in the browser with the display on low, and see how long it dies. And once again, playing mp3s with the display off.

8 hours on the highest brightness it has, with the speakers blaring at max, talking wifi to an access point nearby with power save mode off, auto screen dim off, auto sleep mode off, connected to PSPIRC with plenty of traffic passing.

Yeah. 4 AA batteries. Then I still had a full internal battery, plus I can just swap out AAs and extend runtimes pretty much as long as I want. I could charge 16 sets of AAs in the time it took to discharge one set. That's a little expensive, especally when all you need is 8 AAs and a 15 minute recharger. $30 for the battery charger, comes with 4 2000mAh.
I bought a set of 8 2500 mAhs for $25. (this was before I got my PSP)

The 2500s now run my PSP, the 2000s run my CD-based mp3 player, and / or a borrowed GBA (which has been returned to it's owner since I got the PSP :D )

I'm rather poor. (From this post it doesn't seem like it, eh?) However, due to family things, the rest of my family is scattered over on the east coast.
Due to me being poor, and having clocked just over 9400 miles on greyhound in the first 4 months of 2005 alone, I'll be pulling another 300 to get back home from where I am now in california. They pay for the tickets, I get on the bus and go. It's terribly boring, and rather uncomfortable (I swear they put those damn seats ONE INCH too close to eachother, just so it's completely impossible to get any kind of sleep on the damn bus) so now I have a PSP to entertain me on the go. My next purchase will likely be a 1GB memory stick duo, as soon as they hit $110 or $100 on pricegrabber. They're down to $113 right now. Then I run some Bleach episodes through pspvideo9 and entertain myself until someone figures out how to interface a larger data storage medium. (with a PGA or small ASIC)

On that topic, I also wanted to raise some information.

I have an idea for some hardware.

Basically, one of the hacks is to use a full size memory stick pro on a PSP.
Why not take this one step further? it's proven that it can work with memory sticks, and that everything is encrypted -- but it's not magicgate (AFAIK, it's only being used for ATRAC3 files) so a PC with a memory stick reader could easally access everything. Why not turn this around?
Would it be possible to emulate a memory stick interface, and pipe it to another storage medium?

I need all the hardware and software docs on the memory stick's hardware, software, any kind of data on the internals, the protocols, ETC.

Heck, I've seen IDE interfaces connected to commodore 64s.
... In fact, one might even be able to make a combination unit.

The microcontroller pretends to be a memory stick by emulating the protocols the memory stick itself returns to the host system.

However, using it's other IO pins, it's talking to a compactflash card. (Why compactflash? There's CF-to-almost-anything adapters out there. Also, compactflash can operate in IDE mode, perfect for easy access with a microcontroller.) In the casing, there's a compactflash slot.
(Also, since the microcontroller is handling the filesystem access, IO access could be achived by an EXT port on the device. Inject data or monitor while it's in transit with great accuracy. Analyse what's going on at startup. Watch the filesystem APIs work from the other end. Who knows? it could be looking for something that could be exploited.)

I know this kinda crap's quite possible, I just need one of those memory stick pro to memory stick duo adapters for the PCB memory stick that gets jammed in the PSP, a logging logic analyser, and the memory stick out of my crappy sony cybershot 2MP.