http://www.armadeus.org/wiki/api.php?action=feedcontributions&user=JeanLouisN&feedformat=atom ArmadeusWiki - User contributions [en] 2024-03-29T09:37:27Z User contributions MediaWiki 1.26.3 http://www.armadeus.org/wiki/index.php?title=Linux_Dynamic_Frequency_Scaling&diff=10285 Linux Dynamic Frequency Scaling 2011-11-26T17:20:36Z <p>JeanLouisN: removed modprobe commands, added current mesurement</p> <hr /> <div>{{Under_Construction}}<br /> <br /> ==Introduction==<br /> Dynamic Frequency Scaling aka DFS is only available on [[APF27]] and [[APF51]] based systems.<br /> <br /> You need to make a kernel with the CPUfreq drivers built inside :<br /> &lt;pre class=&quot;host&quot;&gt;<br /> $ make linux26-menuconfig<br /> &lt;/pre&gt;<br /> ==Linux configuration==<br /> &lt;pre class=&quot;config&quot;&gt;<br /> CPU Power Management ---&gt;<br /> [*] CPU Frequency scaling<br /> [*] Enable CPUfreq debugging<br /> &lt;*&gt; CPU frequency translation statistics<br /> [ ] CPU frequency translation statistics details<br /> Default CPUFreq governor (performance) ---&gt;<br /> -*- 'performance' governor<br /> &lt;*&gt; 'powersave' governor<br /> &lt;*&gt; 'userspace' governor for userspace frequency scaling<br /> &lt;*&gt; 'ondemand' cpufreq policy governor<br /> &lt;*&gt; 'conservative' cpufreq governor<br /> &lt;*&gt; CPUfreq driver for i.MX CPUs or &lt;*&gt; i.MX27 frequency driver<br /> &lt;/pre&gt;<br /> <br /> frequency scaling handling is done in ''arch/arm/plat-mxc/cpufreq.c'' or ''arch/arm/mach-mx2/cpufreq_imx27.c''<br /> <br /> ==To know whether the driver is working==<br /> &lt;pre class=&quot;apf&quot;&gt;<br /> # dmesg | grep freq<br /> i.MXC CPU frequency driver<br /> &lt;/pre&gt;<br /> <br /> ==Usage (sysfs)==<br /> * all is here: <br /> &lt;pre class=&quot;apf&quot;&gt;<br /> # cd /sys/devices/system/cpu/cpu0/cpufreq/<br /> # ls<br /> &lt;/pre&gt;<br /> * identify the current governor:<br /> &lt;pre class=&quot;apf&quot;&gt;<br /> # cat scaling_governor<br /> performance<br /> &lt;/pre&gt;<br /> * list available governors:<br /> &lt;pre class=&quot;apf&quot;&gt;<br /> # cat scaling_available_governors<br /> conservative ondemand userspace powersave performance<br /> &lt;/pre&gt;<br /> * change governor:<br /> &lt;pre class=&quot;apf&quot;&gt;<br /> # echo powersave &gt; scaling_governor<br /> &lt;/pre&gt;<br /> * change frequency manually (here 133MHz). '''Requires userspace governor !''' :<br /> &lt;pre class=&quot;apf&quot;&gt;<br /> # echo 133000 &gt; /sys/devices/system/cpu/cpu0/cpufreq/scaling_setspeed<br /> &lt;/pre&gt;<br /> * To know current CPU frequency (here 160MHz):<br /> &lt;pre class=&quot;apf&quot;&gt;<br /> # cat cpuinfo_cur_freq <br /> 160000<br /> &lt;/pre&gt;<br /> * To know how much time CPU has spent in its different state (here 160MHz and 800MHz):<br /> &lt;pre class=&quot;apf&quot;&gt;<br /> # cat stats/time_in_state <br /> 160000 243595<br /> 800000 585<br /> &lt;/pre&gt;<br /> ==Effect on current consumption==<br /> When there is little activity, the CPU frequency has nearly no visible effect on the current consumed.<br /> <br /> This is caused by the &quot;doze mode&quot; : when all processes are sleeping, the kernel turns off the CPU clock until the next IO or timer event.<br /> <br /> When there is a 100% CPU activity, the effect of the CPU frequency can be observed with a ammeter. <br /> <br /> The % CPU activity per process can be checked with the '''top''' command, which is supported by Busybox (but it may be necessary to configure and rebuild busybox to have it)<br /> {{Note| If you make current measurements using a development board like APF27dev or APF51dev, take care of making all the measurement with the '''same supply voltage'''.<br /> Unlike linear regulators, the switching regulator of the board reduces the current consumed when the input voltage is increased (to keep the power constant)}}<br /> <br /> * To Be Continued...<br /> <br /> ==Links==<br /> * http://lxr.linux.no/linux+v2.6.37/Documentation/cpu-freq/user-guide.txt<br /> <br /> [[Category:Power Management]]</div> JeanLouisN http://www.armadeus.org/wiki/index.php?title=Audio_analog_connection&diff=9945 Audio analog connection 2011-07-31T17:07:49Z <p>JeanLouisN: </p> <hr /> <div>== Precautions for using analog audio with APF27_Dev ==<br /> <br /> Connecting a simple stereo headphone and a separate microphone to the '''APF27_Dev''' requires no special precaution.<br /> <br /> But there are some more complex configurations which may cause the audio output to refuse functioning with no visible cause, which are described below. <br /> {{Note| The problems described here do not occur with the '''APF51_Dev''', which has no audio virtual ground.}}<br /> == Electrical background ==<br /> The output of an audio amplifier powered by a single positive voltage presents normally a quiescent DC voltage of approximately half the power supply.<br /> <br /> This DC voltage should never be applied to a loudspeaker, so the classical wiring involves a series capacitor to block this DC voltage (two capacitors for a stereo amplifier).<br /> <br /> The capacitance value has to be large enough to preserve the lowest limit of the audio bandwidth (example : 220 uF for a 16 Ohm loudspeaker).<br /> <br /> Another solution, avoiding the need for large capacitors, consists of creating a '''virtual ground''' for the speaker, the virtual ground being a DC voltage source with the same value as the quiescent voltage of the amplifier output.<br /> <br /> In this context, the APF27_Dev board is designed to use the virtual ground available on the TS2101 audio chip.<br /> <br /> The consequence is that the ground terminal of the audio output jack is connected to a 1.5V DC source (TS2101 VGND pin) instead of the board real ground.<br /> <br /> {{Warning| It is not permitted to connect this virtual ground to the real ground, this would short-circuit the 1.5V source and cause the TS2101 to enter the &quot;short-circuit protection mode&quot; and stop audio output operation (see P. 28 of the TS2101 data sheet).}}<br /> <br /> == Ground loop via mains earth ==<br /> <br /> Here is an example of virtual ground short-circuit caused by the mains earth connection :<br /> * the audio output of the APF27_dev is connected to a home stereo amplifier or any other audio equipement having a mains plug with earth connection<br /> * the APF27_Dev has its ground connected to some equipement having a mains plug with earth connection, for example :<br /> ** the development PC, through the RS232 cable<br /> ** the power supply (AC adaptor) itself<br /> <br /> ==== suggested workarounds : ====<br /> * unplug the RS232 cable<br /> * use a power supply with no earth connection<br /> <br /> {{Note| the Ethernet cable causes no ground loop, because it is isolated by transformers at both ends.}}<br /> <br /> == Ground loop via the microphone jack ==<br /> <br /> The audio input connector (microphone jack) of the APF27_Dev has its ground connected to the board's real ground.<br /> <br /> Connecting together the grounds of both audio jacks causes a virtual ground short-circuit.<br /> <br /> Examples :<br /> * connecting the audio output and input to the same audio equipement, like an audio analog mixer<br /> * connecting the audio output and input to a headset with microphone having common ground<br /> <br /> ==== suggested workaround : ====<br /> * when both jacks are used, connect only the ground of the input jack, leaving the ground of the output jack unconnected<br /> * then insert 2 capacitors in series with the audio outputs<br /> [[Category:Audio]]</div> JeanLouisN http://www.armadeus.org/wiki/index.php?title=Audio_analog_connection&diff=9944 Audio analog connection 2011-07-31T17:02:06Z <p>JeanLouisN: </p> <hr /> <div>== Precautions for using analog audio with APF27_Dev ==<br /> <br /> Connecting a simple stereo headphone and a separate microphone to the '''APF27_Dev''' requires no special precaution.<br /> <br /> But there are some more complex configurations which may cause the audio output to refuse functioning with no visible cause, which are described below. <br /> {{Note| The problems described here do not occur with the '''APF51_Dev''', which has no audio virtual ground.}}<br /> == Electrical background ==<br /> The output of an audio amplifier powered by a single positive voltage presents normally a quiescent DC voltage of approximately half the power supply.<br /> <br /> This DC voltage should never be applied to a loudspeaker, so the classical wiring involves a series capacitor to block this DC voltage (two capacitors for a stereo amplifier).<br /> <br /> The capacitance value has to be large enough to preserve the lowest limit of the audio bandwidth (example : 220 uF for a 16 Ohm loudspeaker).<br /> <br /> Another solution, avoiding the need for large capacitors, consists of creating a '''virtual ground''' for the speaker, the virtual ground being a DC voltage source with the same value as the quiescent voltage of the amplifier output.<br /> <br /> In this context, the APF27_Dev board is designed to use the virtual ground available on the TS2101 audio chip.<br /> <br /> The consequence is that the ground terminal of the audio output jack is connected to a 1.5V DC source (TS2101 VGND pin) instead of the board real ground.<br /> <br /> {{Warning| It is not permitted to connect this virtual ground to the real ground, this would short-circuit the 1.5V source and cause the TS2101 to enter the &quot;short-circuit protection mode&quot; and stop audio output operation (see P. 28 of the TS2101 data sheet).}}<br /> <br /> == Ground loop via mains earth ==<br /> <br /> Here is an example of virtual ground short-circuit caused by the mains earth connection :<br /> * the audio output of the APF27_dev is connected to a home stereo amplifier or any other audio equipement having a mains plug with earth connection<br /> * the APF27_Dev has its ground connected to some equipement having a mains plug with earth connection, for example :<br /> ** the development PC, through the RS232 cable<br /> ** the power supply (AC adaptor) itself<br /> <br /> ==== suggested workarounds : ====<br /> * unplug the RS232 cable<br /> * use a power supply with no earth connection<br /> <br /> {{Note| the Ethernet cable causes no ground loop, because it is isolated by transformers at both ends.}}<br /> <br /> == Ground loop via the microphone jack ==<br /> <br /> The audio input connector (microphone jack) of the APF27_Dev has its ground connected to the board's real ground.<br /> <br /> Connecting together the grounds of both audio jacks causes a virtual ground short-circuit.<br /> <br /> Examples :<br /> * connecting the audio output and input to the same audio equipement, like an audio analog mixer<br /> * connecting the audio output and input to a headset with microphone with common ground<br /> <br /> ==== suggested workaround : ====<br /> * when both jacks are used, connect only the ground of the input jack, leaving the ground of the output jack unconnected<br /> * then insert 2 capacitors in series with the audio outputs<br /> [[Category:Audio]]</div> JeanLouisN http://www.armadeus.org/wiki/index.php?title=Audio_analog_connection&diff=9943 Audio analog connection 2011-07-31T16:53:34Z <p>JeanLouisN: Precautions for using analog audio with APF27_Dev, avoidin short-circuiting the virtual ground</p> <hr /> <div>== Precautions for using analog audio with APF27_Dev ==<br /> <br /> Connecting a simple stereo headphone and a separate microphone to the '''APF27_Dev''' requires no special precaution.<br /> <br /> But there are some more complex configurations which may cause the audio output to refuse functioning with no visible cause, which are described below. <br /> {{Note| The problems described here do not occur with the '''APF51_Dev''', which has no audio virtual ground.}}<br /> == Electrical background ==<br /> The output of an audio amplifier powered by a single positive voltage presents normally a quiescent DC voltage of approximately half the power supply.<br /> <br /> This DC voltage should never be applied to a loudspeaker, so the classical wiring involves a series capacitor to block this DC voltage (two capacitors for a stereo amplifier).<br /> <br /> The capacitance value has to be large enough to preserve the lowest limit of the audio bandwidth (example : 220 uF for a 16 Ohm loudspeaker).<br /> <br /> Another solution, avoiding the need for large capacitors, consists of creating a '''virtual ground''' for the speaker, the virtual ground being a DC voltage source with the same value as the quiescent voltage of the amplifier output.<br /> <br /> In this context, the APF27_Dev board is designed to use the virtual ground available on the TS2101 audio chip.<br /> <br /> The consequence is that the ground terminal of the audio output jack is connected to a 1.5V DC source (TS2101 VGND pin) instead of the board real ground.<br /> <br /> {{Warning| It is not permitted to connect this virtual ground to the real ground, this would short-circuit the 1.5V source and cause the TS2101 to enter the &quot;short-circuit protection mode&quot; and stop audio output operation (see P. 28 of the TS2101 data sheet).}}<br /> <br /> == Ground loop via mains earth ==<br /> <br /> Here is an example of virtual ground short-circuit caused by the mains earth connection :<br /> * the audio output of the APF27_dev is connected to a home stereo amplifier or any other audio equipement having a mains plug with earth connection<br /> * the APF27_Dev has its ground connected to some equipement having a mains plug with earth connection, for example :<br /> ** the development PC, through the RS232 cable<br /> ** the power supply (AC adaptor) itself<br /> <br /> ==== suggested workarounds : ====<br /> * unplug the RS232 cable<br /> * use a power supply with no earth connection<br /> <br /> {{Note| the Ethernet cable causes no ground loop, because it is isolated by transformers at both ends.}}<br /> <br /> == Ground loop via the microphone jack ==<br /> <br /> The audio input connector (microphone jack) of the APF27_Dev has its ground connected to the board's real ground.<br /> <br /> Connecting together the grounds of both audio jacks causes a virtual ground short-circuit.<br /> <br /> Examples :<br /> * connecting the audio output and input to the same audio equipement, like an audio analog mixer<br /> * connecting the audio output and input to a headset with microphone with common ground<br /> <br /> ==== suggested workaround : ====<br /> * when both jacks are used, connect only the ground of the input jack, leaving the ground of the output jack unconnected<br /> * then insert 2 capacitors in series with the audio outputs</div> JeanLouisN http://www.armadeus.org/wiki/index.php?title=Audio_analog_connection&diff=9942 Audio analog connection 2011-07-31T16:45:31Z <p>JeanLouisN: </p> <hr /> <div>Connecting a simple stereo headphone and a separate microphone to the '''APF27_Dev''' requires no special precaution.<br /> <br /> But there are some more complex configurations which may cause the audio output to refuse functioning with no visible cause, which are described below. <br /> {{Note| The problems described here do not occur with the '''APF51_Dev''', which has no audio virtual ground.}}<br /> == Electrical background ==<br /> The output of an audio amplifier powered by a single positive voltage presents normally a quiescent DC voltage of approximately half the power supply.<br /> <br /> This DC voltage should never be applied to a loudspeaker, so the classical wiring involves a series capacitor to block this DC voltage (two capacitors for a stereo amplifier).<br /> <br /> The capacitance value has to be large enough to preserve the lowest limit of the audio bandwidth (example : 220 uF for a 16 Ohm loudspeaker).<br /> <br /> Another solution, avoiding the need for large capacitors, consists of creating a '''virtual ground''' for the speaker, the virtual ground being a DC voltage source with the same value as the quiescent voltage of the amplifier output.<br /> <br /> In this context, the APF27_Dev board is designed to use the virtual ground available on the TS2101 audio chip.<br /> <br /> The consequence is that the ground terminal of the audio output jack is connected to a 1.5V DC source (TS2101 VGND pin) instead of the board real ground.<br /> <br /> {{Warning| It is not permitted to connect this virtual ground to the real ground, this would short-circuit the 1.5V source and cause the TS2101 to enter the &quot;short-circuit protection mode&quot; and stop audio output operation (see P. 28 of the TS2101 data sheet).}}<br /> <br /> == Ground loop via mains earth ==<br /> <br /> Here is an example of virtual ground short-circuit caused by the mains earth connection :<br /> * the audio output of the APF27_dev is connected to a home stereo amplifier or any other audio equipement having a mains plug with earth connection<br /> * the APF27_Dev has its ground connected to some equipement having a mains plug with earth connection, for example :<br /> ** the development PC, through the RS232 cable<br /> ** the power supply (AC adaptor) itself<br /> <br /> ==== suggested workarounds : ====<br /> * unplug the RS232 cable<br /> * use a power supply with no earth connection<br /> <br /> {{Note| the Ethernet cable causes no ground loop, because it is isolated by transformers at both ends.}}<br /> <br /> == Ground loop via the microphone jack ==<br /> <br /> The audio input connector (microphone jack) of the APF27_Dev has its ground connected to the board's real ground.<br /> <br /> Connecting together the grounds of both audio jacks causes a virtual ground short-circuit.<br /> <br /> Examples :<br /> * connecting the audio output and input to the same audio equipement, like an audio analog mixer<br /> * connecting the audio output and input to a headset with microphone with common ground<br /> <br /> ==== suggested workaround : ====<br /> * when both jacks are used, connect only the ground of the input jack, leaving the ground of the output jack unconnected<br /> * then insert 2 capacitors in series with the audio outputs</div> JeanLouisN http://www.armadeus.org/wiki/index.php?title=Audio_analog_connection&diff=9941 Audio analog connection 2011-07-31T16:38:56Z <p>JeanLouisN: New page: Connecting a simple stereo headphone and a seprate microphone to the '''APF27_Dev''' requires no special precaution. But there are some more complex configurations which may cause the aud...</p> <hr /> <div>Connecting a simple stereo headphone and a seprate microphone to the '''APF27_Dev''' requires no special precaution.<br /> <br /> But there are some more complex configurations which may cause the audio output to refuse functioning with no visible cause, which are described below. <br /> {{Note| The problems described here do not occur with the '''APF51_Dev''', which has no audio virtual ground.}}<br /> == Electrical background ==<br /> The output of an audio amplifier powered by a single positive voltage source presents normally a quiescent DC voltage of approximately half the power supply.<br /> <br /> This DC voltage should never be applied to a loudspeaker, so the classical wiring involves a series capacitor to block this DC voltage (two capacitors for a stereo amplifier).<br /> <br /> The capacitance value has to be large enough to preserve the lowest limit of the audio bandwidth (example : 220 uF for a 16 Ohm loudspeaker).<br /> <br /> Another solution, avoiding the need of large capacitors, consists of creating a '''virtual ground''' for the speaker, the virtual ground is a DC voltage source with the same value as the quiescent voltage of the amplifier output.<br /> <br /> The APF27_Dev board is designed to use the virtual ground available on the TS2101 audio chip.<br /> <br /> The consequence is that the ground terminal of the audio output jack is connected to a 1.5V DC source (TS2101 VGND pin) instead of the board real ground.<br /> <br /> {{Warning| It is not permitted to connect this virtual ground to the real ground, this would short-circuit the 1.5V source and cause the TS2101 to enter the &quot;short-circuit protection mode&quot; and stop audio output operation (see P. 28 of the TS2101 data sheet).}}<br /> <br /> == Ground loop via mains earth ==<br /> <br /> Here is an example of virtual ground short-circuit caused by the mains earth connection :<br /> * the audio output of the APF27_dev is connected to a home stereo amplifier or any audio other audio equipement having a mains plug with earth connection<br /> * the APF27_Dev has its ground connected to some equipement having a mains plug with earth connection, for example :<br /> ** the development PC, through the RS232 cable<br /> ** the power supply (AC adaptor) itself<br /> <br /> ==== suggested workarounds : ====<br /> * unplug the RS232 cable<br /> * use a power supply with no earth connection<br /> <br /> {{Note| the Ethernet cable causes no ground loop, because it is isolated by transformers at both ends.}}<br /> <br /> == Ground loop via the microphone jack ==<br /> <br /> The audio input connector (microphone jack) of the APF27_Dev has its ground connected to the board's real ground.<br /> <br /> Connecting together the grounds of both audio jacks causes a virtual ground short-circuit.<br /> <br /> Examples :<br /> * connecting the audio output and input to the same audio equipement, like an audio analog mixer<br /> * connecting the audio output and input to a headset with microphone with common ground<br /> <br /> ==== suggested workaround : ====<br /> * when both jacks are used, connect only the ground of the input jack, leaving the ground of the output jack unconnected<br /> * then insert 2 capacitors in series with the audio outputs</div> JeanLouisN http://www.armadeus.org/wiki/index.php?title=FAQ&diff=9733 FAQ 2011-06-07T13:42:07Z <p>JeanLouisN: </p> <hr /> <div>==General Questions==<br /> <br /> ===What is the &quot;Armadeus Project&quot; intended for ?===<br /> The goal of this project is to allow everyone to easily develop embedded Open Source systems.<br /> When all the embedded systems in the world run open source software, life will be better ;-)<br /> <br /> ===What does &quot;Armadeus&quot; mean ?===<br /> &quot;Armadeus&quot; is the contraction of &quot;ARM&quot; and &quot;Amadeus&quot;. Indeed we chose ARM architecture for its simplicity and efficiency and our boards are as gifted to run embedded systems as Mozart was for composing music. Moreover the Armadeus Project was born in 2006, year of the 250th Mozart's Birthday.<br /> <br /> ===How do I get help ?===<br /> First take a look at the whole Wiki. If you can't find help contact us directly: [[Help:Contents]]. If you are located in France, we have members in Besançon, Montbéliard &amp; Mulhouse who can provide help. For Swiss members, we can provide help in Basel, Lausanne &amp; Zürich.<br /> <br /> === Why are there http://armadeus.org and http://armadeus.com ? Is there a company backing/hiding behind the association? ===<br /> <br /> &lt;http://armadeus.com&gt; is the webpage dedicated to Armadeus Systems. &lt;br&gt; <br /> The company develops and produces the main boards which are then available at a discount price to the association members (non commercial usage only!). <br /> <br /> Historically the Armadeus Project (association) was founded before Armadeus Systems (company). &lt;br&gt;<br /> After one year of existence we realized that it was quite impossible to achieve our goals without investing more time and more cash. That's why the company was founded.<br /> <br /> The association was started for our own pleasure and also to help universities and enthusiasts develop Linux embedded systems. &lt;br&gt;<br /> It is a non profit association and works mostly on founders' free time.<br /> <br /> The developments (hardware/firmware/software) performed within the association are under GPL/LGPL licence. &lt;br&gt;<br /> The hardware funded by the association can not be used in a commercial product.<br /> <br /> ==Hardware Questions==<br /> ===Are the Armadeus boards RoHS compliant ?===<br /> Yes<br /> <br /> ==Software Questions==<br /> ===Where are the sources ?===<br /> We have a project on SourceForge: http://sourceforge.net/projects/armadeus/ with a [[GIT]] repository.<br /> <br /> ===What are the performances of your board ?===<br /> [[Performances | Wiki page talking about board performances]]<br /> <br /> ===How do I get help/support===<br /> [[Help:Contents]]<br /> <br /> ===U-Boot seems ok but prints: *** Warning - bad CRC, using default environment - Is there something wrong?=== <br /> Everything is ok. It is an U-Boot feature that happens if the environment partition is empty.&lt;br&gt;<br /> Please read the U-Boot FAQ on this issue: http://www.denx.de/wiki/view/DULG/WarningBadCRCUsingDefaultEnvironment<br /> <br /> {{LanguageBar|FAQ|FAQ|FAQ}}</div> JeanLouisN http://www.armadeus.org/wiki/index.php?title=User:JeanLouisN&diff=9676 User:JeanLouisN 2011-05-17T06:19:58Z <p>JeanLouisN: New page: Jean-Louis Noullet, professionnel senior de l'électronique embarquée == Experience en Linux/ARM == * '''AT91RM9200/SSV''' : un ARM9 de Atmel, sur board SSV DNP9200 - simple et fiable. R...</p> <hr /> <div>Jean-Louis Noullet,<br /> professionnel senior de l'électronique embarquée<br /> == Experience en Linux/ARM ==<br /> * '''AT91RM9200/SSV''' : un ARM9 de Atmel, sur board SSV DNP9200 - simple et fiable. Résultat : un produit industriel en phase d'exploitation (pilotage de machine-outil avec interface web/AJAX)<br /> * '''OMAPL138/Hawkboard''' : double core ARM/DSP de TI, sur board Hawkboard. Projet abandonné après beaucoup d'efforts : instabilité irrémediable du hardware<br /> * '''iMX27/Armadeus APF27''' : en phase de découverte...</div> JeanLouisN