Granitehead
Advanced Member
- Joined
- Oct 16, 2009
- Messages
- 3,011
Roughly 21mm in diameter.
GamesCom, Shoulderbuttons and Heating
- Thread starter EvilDragon
- Start date
Linux-SWAT
Forum Addict!
- Joined
- Feb 13, 2010
- Messages
- 9,194
I'm back in France and missed a lot of stuff on these boards. Is there a GamesCom thread with pictures where I can share mine ?
fantomid
Very Active Member
Welcome back in France.
No thread, at the moment, I think !!
No thread, at the moment, I think !!
fusion_power
Advanced Member
- Joined
- Jan 18, 2010
- Messages
- 11,418
I have large thumbs, but I find the Pandora D-pad perfect in size.. Everything is a tight fit on the Pyra, I doubt there is much room.. plus I doubt much can change in the overall design at this point.
ptitSeb
Serial Porter
Welcome back.
I think the closest thread to what you ask is there: http://boards.openpandora.org/topic/16839-some-gamescom-retro-area-pictures/
Last edited by a moderator:
Linux-SWAT
Forum Addict!
- Joined
- Feb 13, 2010
- Messages
- 9,194
Ok, thanks !
Not sure if this has been suggested, but I've created a rough mspaint job to show an idea I had for the shoulder buttons.
Basically, the R1 and R2 buttons pivot on the same axis, R1 is taller than R2 and R2 fits between R1, with the pushbottons underneath at different horizontal positions. The smaller one should be more pronounced.
Basically, the R1 and R2 buttons pivot on the same axis, R1 is taller than R2 and R2 fits between R1, with the pushbottons underneath at different horizontal positions. The smaller one should be more pronounced.
Grench
Forum Addict!
- Joined
- Oct 3, 2008
- Messages
- 6,629
OK. Thats downright cool. I can even imagine how the mechanics of it would work on the inside. VERY nice. Hey ED - check this one out!Not sure if this has been suggested, but I've created a rough mspaint job to show an idea I had for the shoulder buttons.
Basically, the R1 and R2 buttons pivot on the same axis, R1 is taller than R2 and R2 fits between R1, with the pushbottons underneath at different horizontal positions. The smaller one should be more pronounced.
fusion_power
Advanced Member
Would not work, I also thought about this design but you can't separate both buttons enough to not press the wrong button accidently, at least not to 100%. And remember, these buttons are still pretty small, which makes it even harder to press only one of them (comfortably).
Last edited by a moderator:
In addition to the OMAP5, there's another IC on the board that deserves overheating attention - the PMIC. I remember when we worked on a project back in Autumn 2012 with OMAP5432. They were engineering samples, soldered on Panda 5 boards. We had them running for a while with Linux, and all was well. Unfortunately several boards died unexpectedly, and in all cases the PMIC was fried. This happened when we ran QNX for no more than 30min on the OMAP5 board.
A member of our HW team checked the fried boards, but we didn't found any specific indication about why it was fried - some boards were running on laboratory power supplies (which limit the surge current on power-on), and some were running on simpler PSUs, so no specific correlation there. In the end we concluded that the PMIC was somehow overloaded with the poorer power/clock management from QNX.
The PMIC for OMAP5 was a complex beast - as far as I remember correctly, it was multi-channel DC-DC with more than 15 sequenced voltage rails, and the OMAP5 consumed about 13-15A on one of the rails (I'm citing by memory). So I think that PMIC should be definitely cooled properly to avoid nasty surprises.
And by the way - the Cortex-A15 core is an absolute high performer. If this thing is successfully integrated into Pyra, it will be a dream box for the community.
Kind regards,
picmaster
A member of our HW team checked the fried boards, but we didn't found any specific indication about why it was fried - some boards were running on laboratory power supplies (which limit the surge current on power-on), and some were running on simpler PSUs, so no specific correlation there. In the end we concluded that the PMIC was somehow overloaded with the poorer power/clock management from QNX.
The PMIC for OMAP5 was a complex beast - as far as I remember correctly, it was multi-channel DC-DC with more than 15 sequenced voltage rails, and the OMAP5 consumed about 13-15A on one of the rails (I'm citing by memory). So I think that PMIC should be definitely cooled properly to avoid nasty surprises.
And by the way - the Cortex-A15 core is an absolute high performer. If this thing is successfully integrated into Pyra, it will be a dream box for the community.
Kind regards,
picmaster
Last edited by a moderator:
Well, we ran the devboards 10 hours per day, using nearly all the CPU power (with multiple emulators and a desktop) for 6 days straight at the GamesCom, but didn't have issues with the PMIC or a lot of heat here.
BTW: Switching to kernel 3.18 (or 3.17? I forgot) seems to have dropped the OMAP5 temperature by 10 degrees in a normal use case.
BTW: Switching to kernel 3.18 (or 3.17? I forgot) seems to have dropped the OMAP5 temperature by 10 degrees in a normal use case.
D
dcosnet
Guest
awesome news
TeDaDeS
Forum Addict!
ED, the PMIC might still heat up a lot when charging the battery.
I'm not sure if you had the opportunity to have a battery attached, but if not you can always try again and see if the PMIC gets hot.
(Btw. always when I visit around mid-night (~24:00+) the site is slow, 56K modem style.)
I'm not sure if you had the opportunity to have a battery attached, but if not you can always try again and see if the PMIC gets hot.
(Btw. always when I visit around mid-night (~24:00+) the site is slow, 56K modem style.)
I wouldn't be surprised that the heat sink is doing very little to cool the thing especially if you're not using thermal paste, if you're not using thermal paste then what's the point of even using a heat sink? Also what thermal paste are you using?
I know that me personallyl the cheap crap that most companies use for their thermal solution compared with some as5 or ic-diamond reduces the temperature by ~15-20*C under load which would greatly help here since it'd allow more contact with the heat sink. Since you didn't say you're using thermal paste I imaigne that you're not.
I know that me personallyl the cheap crap that most companies use for their thermal solution compared with some as5 or ic-diamond reduces the temperature by ~15-20*C under load which would greatly help here since it'd allow more contact with the heat sink. Since you didn't say you're using thermal paste I imaigne that you're not.
- Joined
- Jan 18, 2010
- Messages
- 11,418
They're running it entirely without the Heat-sink, thermal paste is irrelevant.
Last edited by a moderator:
Once again here too, he said "using a heat sink doesn't seem to help at all." Thus there's no thermal paste on it, and 70*C is way too hot for an SOC to be especially a mobile one. that thing needs a heat sink with proper thermal paste. My laptop's quad core cpu barely gets above 75*C under full load, and my mobile devices never get above ~50*C even under full load. I don't have any omap5s lying around. I do have a couple of omap4s and an omap3 though.
- Joined
- Jan 18, 2010
- Messages
- 11,418
And once again he took it off.
Last edited by a moderator:
The devboard comes with a heatsink preinstalled. I ran my tests with it and then removed it. Don't ask me what the manufacturer uses as thermal paste...
Similar threads
