Active Topics
-
The N900 is now 15 years old! (9)
to Nokia N900 by Arno_11 - 9 hrs, 19 mins ago -
Sailfish OS for the Motorola Moto G7 Power (XT1955-5) - (ocean) (4)
to SailfishOS by Maemish - 5 days, 21 hrs ago - more...
Tzoni |
2010-05-15
, 19:27
|
|
Posts: 24 |
Thanked: 6 times |
Joined on May 2010
@ Turku, Finland
|
#3651
|
I overclocked my N900 again. I use the same stuff that I used last time but now it doesnt use 500-850 mhz but 250-850mhz. Is this intented or did I screw something up?
|
|
2010-05-15
, 19:34
|
|
Posts: 202 |
Thanked: 30 times |
Joined on Mar 2010
@ CZE
|
#3652
|
Originally Posted by Tzoni
If youre using Titans kernel-power (from repositories), try setting your preferred profile via
I overclocked my N900 again. I use the same stuff that I used last time but now it doesnt use 500-850 mhz but 250-850mhz. Is this intented or did I screw something up?
Code:
kernel-config load <profile_name>
Code:
kernel-config load xlv
and then setting the minimum and maximum frequencies via
Code:
kernel-config limits <minimum> <maximum>
Code:
kernel-config limits 250 850
The profile name can be AFAIK
Code:
defaul, lv, ulv, xlv, ideal, starving
__________________
"I'm a Nokia fan. Ericsson ruled, Sony Ericsson SUX!!!"
Titans kernel up-and-running, N900 PR1.2 @ XLV 250-900 (Overclocking Guide)
"I'm a Nokia fan. Ericsson ruled, Sony Ericsson SUX!!!"
Titans kernel up-and-running, N900 PR1.2 @ XLV 250-900 (Overclocking Guide)
Last edited by michalurban; 2010-05-15 at 19:39.
|
|
2010-05-15
, 19:38
|
|
Posts: 126 |
Thanked: 327 times |
Joined on Nov 2009
@ Finland
|
#3653
|
Originally Posted by titan
As a hazard guess I'd say its because either the battery or the CPU power regulation cant handle the load. So this is kind of identical to all N900 out there. Maybe a very good unit could do 1200mhz?
1150 is the limit as no got 1200 working.
note that 1150Mhz requires almost 2.5x the current of the stock 600Mhz!
__________________
My OC kernel collection
My OC kernel collection
 |
|
2010-05-15
, 19:51
|
|
Posts: 324 |
Thanked: 201 times |
Joined on Apr 2010
@ UK
|
#3654
|
ive just been doing a bit of reading and disscovered that if you set the freq govenor to conservative it will scale up through all available freqs as well as down.
ive just started testing it but can confirm that its now hitting every freq on its way up to 1Ghz, ill report back what impact if any this has on performanceand battery.
so far it is staying at lower freq's for much longer and only jumping up when needed.
Last edited by matts76; 2010-05-15 at 19:53.
ive just started testing it but can confirm that its now hitting every freq on its way up to 1Ghz, ill report back what impact if any this has on performanceand battery.
so far it is staying at lower freq's for much longer and only jumping up when needed.
__________________
My N900@1Ghz Rocks
I wanted a phone, but i ended up with an all encompasing obsession
My N900@1Ghz RocksI wanted a phone, but i ended up with an all encompasing obsession
Last edited by matts76; 2010-05-15 at 19:53.
| The Following 2 Users Say Thank You to matts76 For This Useful Post: | ||
|
|
2010-05-15
, 19:58
|
|
Posts: 324 |
Thanked: 201 times |
Joined on Apr 2010
@ UK
|
#3655
|
im currently using cpufrequi to set it but cant set up threshold and sample rate they both appear to be 0 which obviously cant be right, im on maemo25 at the mo but tomorrow ill upgrade to 27 and play with the settings.
__________________
My N900@1Ghz Rocks
I wanted a phone, but i ended up with an all encompasing obsession
My N900@1Ghz RocksI wanted a phone, but i ended up with an all encompasing obsession
|
|
2010-05-15
, 20:40
|
|
Posts: 24 |
Thanked: 4 times |
Joined on Mar 2010
|
#3656
|
Originally Posted by Gusse
Hey Gusse,
@Titan
Would it be possible to add also "avoid frequency" entry to kernel-config -file or is that possible to do by removing unwanted freqs fom FREQS list? Almost everything else is configurable from file.
Just leave out the frequency in the list of frequencies and it will avoid it. I was just wondering the same thing too haha.
|
|
2010-05-16
, 06:14
|
|
Posts: 543 |
Thanked: 151 times |
Joined on Feb 2010
@ Germany
|
#3657
|
I use Titans kernel currently and I really like it so far, frequencies till 1GHz are stable (I didn't tested more yet) even when I leave for example the PSX Emu a while running.
Still for normal usage I prefer so the ULV / XLV kernel, the ULV kernel seems to be in idle mode @ 250MHz better for the battery?
I also wonder what the limit of the CPU is when it comes to the temperature, I saw that my device gets 35°C~ CPU idle when it loads the battery. What values do you have, is 40°C~ the upper limit?
Still for normal usage I prefer so the ULV / XLV kernel, the ULV kernel seems to be in idle mode @ 250MHz better for the battery?
I also wonder what the limit of the CPU is when it comes to the temperature, I saw that my device gets 35°C~ CPU idle when it loads the battery. What values do you have, is 40°C~ the upper limit?
|
|
2010-05-16
, 07:30
|
|
Posts: 38 |
Thanked: 13 times |
Joined on Oct 2009
|
#3658
|
OMAP 3530 (a relative of our OMAP 3430) is designed to run at junction temperature of up to 90 degrees (the maximal temperature inside the chip).
The design is such, that if you keep the external surface of the chip at 70 degrees or lower, the maximum temperature will not exceed this limit.
have a look at page 124 here:
http://focus.ti.com/lit/ds/symlink/omap3530.pdf
I doubt if you exceed this 70 degrees.
There are four problems however
1. We don't have an accurate sensor to measure the heat sink temperature.
2. When the temperature is elevated, you need higher voltage to drive the chip, for me, ULV does not work when I am not in an air-conditioned room. When it's hot, you need more voltage (which generates more heat).
3. Smart reflex in the n900 is very buggy and crashes the device. If it were working, the device would have increased the voltage and limit the frequency by itself when it get's hot.
4. Nokia designed the cooling system to handle different thermal loads, it might not be equipped to handle what we put against it. I guess we are all aware of that.
What you can do is measure the battery temperature, it's a crude way of estimating the CPU heat sink temperature, but it's rather stable and simple to measure.
I assume there is up to 20 degree difference between the two.
So when it's above 45 degrees I change my device to the default kernel (limited at 600Mhz and normal voltage). I assume Nokia checked it @ high temperatures, I didn't.
When it's above 38 degrees, I run titan's ideal limited to 600Mhz to save battery power.
When it's below 38 degrees I run a custom extremely low voltage settings which hopefully increases my battery life.
This allows me to run icedove 24/7 and have a 8-12 hour battery life.
The design is such, that if you keep the external surface of the chip at 70 degrees or lower, the maximum temperature will not exceed this limit.
have a look at page 124 here:
http://focus.ti.com/lit/ds/symlink/omap3530.pdf
I doubt if you exceed this 70 degrees.
There are four problems however
1. We don't have an accurate sensor to measure the heat sink temperature.
2. When the temperature is elevated, you need higher voltage to drive the chip, for me, ULV does not work when I am not in an air-conditioned room. When it's hot, you need more voltage (which generates more heat).
3. Smart reflex in the n900 is very buggy and crashes the device. If it were working, the device would have increased the voltage and limit the frequency by itself when it get's hot.
4. Nokia designed the cooling system to handle different thermal loads, it might not be equipped to handle what we put against it. I guess we are all aware of that.
What you can do is measure the battery temperature, it's a crude way of estimating the CPU heat sink temperature, but it's rather stable and simple to measure.
I assume there is up to 20 degree difference between the two.
So when it's above 45 degrees I change my device to the default kernel (limited at 600Mhz and normal voltage). I assume Nokia checked it @ high temperatures, I didn't.
When it's above 38 degrees, I run titan's ideal limited to 600Mhz to save battery power.
When it's below 38 degrees I run a custom extremely low voltage settings which hopefully increases my battery life.
This allows me to run icedove 24/7 and have a 8-12 hour battery life.
Originally Posted by Crogge
I use Titans kernel currently and I really like it so far, frequencies till 1GHz are stable (I didn't tested more yet) even when I leave for example the PSX Emu a while running.
Still for normal usage I prefer so the ULV / XLV kernel, the ULV kernel seems to be in idle mode @ 250MHz better for the battery?
I also wonder what the limit of the CPU is when it comes to the temperature, I saw that my device gets 35°C~ CPU idle when it loads the battery. What values do you have, is 40°C~ the upper limit?
| The Following User Says Thank You to sdpkom For This Useful Post: | ||
|
|
2010-05-16
, 08:41
|
|
Posts: 1,751 |
Thanked: 844 times |
Joined on Feb 2010
@ Sweden
|
#3659
|
Originally Posted by sdpkom
The interesting stuff though would be to know what temp the battery could hold before it breaks. The chance is greater that the battery or other components breaks before the cpu does.
OMAP 3530 (a relative of our OMAP 3430) is designed to run at junction temperature of up to 90 degrees (the maximal temperature inside the chip).
The design is such, that if you keep the external surface of the chip at 70 degrees or lower, the maximum temperature will not exceed this limit.
have a look at page 124 here:
http://focus.ti.com/lit/ds/symlink/omap3530.pdf
I doubt if you exceed this 70 degrees.
There are four problems however
1. We don't have an accurate sensor to measure the heat sink temperature.
2. When the temperature is elevated, you need higher voltage to drive the chip, for me, ULV does not work when I am not in an air-conditioned room. When it's hot, you need more voltage (which generates more heat).
3. Smart reflex in the n900 is very buggy and crashes the device. If it were working, the device would have increased the voltage and limit the frequency by itself when it get's hot.
4. Nokia designed the cooling system to handle different thermal loads, it might not be equipped to handle what we put against it. I guess we are all aware of that.
What you can do is measure the battery temperature, it's a crude way of estimating the CPU heat sink temperature, but it's rather stable and simple to measure.
I assume there is up to 20 degree difference between the two.
So when it's above 45 degrees I change my device to the default kernel (limited at 600Mhz and normal voltage). I assume Nokia checked it @ high temperatures, I didn't.
When it's above 38 degrees, I run titan's ideal limited to 600Mhz to save battery power.
When it's below 38 degrees I run a custom extremely low voltage settings which hopefully increases my battery life.
This allows me to run icedove 24/7 and have a 8-12 hour battery life.
|
|
2010-05-16
, 10:26
|
|
Posts: 38 |
Thanked: 13 times |
Joined on Oct 2009
|
#3660
|
GP, one of the largest manufacturers of rechargeable batteries defines
0-60 degrees as working temperature
0-55 degrees is the range the sensor is designed for and accurate at.
90-98 degrees is the auto shutdown temperature.
It's safe unless you exceed 98 degrees.
Typically in electronic systems, if you supply enough voltage, and keep the temperature low enough, transistors will work.
I had my N900 work in a 44 degrees (outside) all of yesterday, it works fine, I'm sure the internal temperature was higher then this.
When it's more then 50 degrees, it's not comfortable in your pocket anymore.
0-60 degrees as working temperature
0-55 degrees is the range the sensor is designed for and accurate at.
90-98 degrees is the auto shutdown temperature.
It's safe unless you exceed 98 degrees.
Typically in electronic systems, if you supply enough voltage, and keep the temperature low enough, transistors will work.
I had my N900 work in a 44 degrees (outside) all of yesterday, it works fine, I'm sure the internal temperature was higher then this.
When it's more then 50 degrees, it's not comfortable in your pocket anymore.
Originally Posted by AlMehdi
The interesting stuff though would be to know what temp the battery could hold before it breaks. The chance is greater that the battery or other components breaks before the cpu does.
| The Following User Says Thank You to sdpkom For This Useful Post: | ||
 |
«
Previous Thread
|
Next Thread
»
|
All times are GMT. The time now is 03:00.