The A20-OlinuXino-Micro board (Rev.C tested) emits enough RF noise to seriously degrade FM radio reception.
This appears to be a problem with the 6-16V input power supply. To reproduce the problem, take a vanilla a20-OlinuXino-Micro and plug a power supply into this while listening to an FM radio station (no sdcard is necessary).
The reason that I think it is the power supply circuitry, is that the FM interference stops if I unplug the 6-16V supply, while leaving the battery plugged in (BATTERY-LIPO1400mAh), and debian is still up and running, but there is no FM interference.
I have experienced this problem with both of my boards (I only have two A20-OlinuXino-Micros). I have tested in 4 automobiles, and with a little $15.00 FM radio from RadioShack. I have tested with wall power, car 12v power, and a separate 12V battery, each plugged into the 6-16V. I have tested with the radio on the same power supply as the A20-OlinuXino and with them on separate supplies. The problem existed in all tests.
This may not be Olimex's problem. Maybe all of these boards (RaspberryPi, BeagleBone, etc.) do this, and we just need to find a way to wrap them in a shield. I do not know enough about this stuff to make any suggestions. I tried wrapping the whole thing in aluminum foil, but I did not notice any improvement.
I am at a loss. I am hoping to make an automobile product out of the A20-OlinuXino. Any thoughts? Community? Olimex?
Thank you
please use iron metal box (aluminum is not magnetic isolation) which to be connected to GND
Thank you for the response. Enclosing in an iron metal box is definitely not ideal and shouldn't be necessary. There are plenty of Cigarette lighter voltage regulators out there that do not interfere with FM radio and are only enclosed in plastic. Again, it is likely the board's voltage regulator, since the problem does not exist when powered from lipo, only when powered from 12V.
Not interfering with FM radio would likely be a requirement for people making car pc's, which could be a large market.
I don't know anything about EMI. Maybe someone who knows about this stuff could take a look at that circuit and make a recommendation. We could pool money to pay that person (I could throw in
$200 or $300 US). It is completely reasonable that early board revisions would have problems, we just need to find a fix. Rev C seems to be emitting far too much RFI/EMI. Thanks again
Here is a sound file of the problem. You can hear the radio station completely cutting out. I plug in and unplug a 12V wall wart, a lipo is plugged in and debian is running the whole time (neither the lipo nor debian are necessary to see the problem). The radio is battery powered sitting on my desk, and the station is FM. Sorry, the audio is quiet so you have to turn it up.
https://drive.google.com/file/d/0B9Oqle-M8CLFaUNocXJTMTUtLWs/edit?usp=sharing
Again, no problem with a generic 12v input, 5v output regulator, enclosed in plastic, powering some other device.
Not knowing what I am doing, I attached a 100uF capacitor across the 6-16V input and there was a little improvement. I know we can fix this.
I guess the noise is emitted from the inductance of the input DCDC you can shield just this inductance
I just tested on a spanking new A20-OlinuXino-micro Rev.E. There is definitely a problem with the dc-dc emitting a lot of emi. You can see the problem by powering a bare board (no sd card, no battery, no display, no peripherals) via the dc-dc jack and trying to listen to a nearby radio.
I just sprayed an entire $35 can of Super Shield nickel conductive coating inside a completely enclosed plastic shell with only the a20-olinuxino inside, and grounded the sprayed shell to the board's ground; only a couple of ohms resistance at different test points). This is an industrial coating made specifically for this problem. There was only minor improvement. Still unacceptable interference.
Olimex - Will you address this?
It is understandable that problems like this will arise, but this level of noise emission makes the board unusable in a product. Especially if anyone hopes to receive FCC approval for their product. We can put $1000 toward your engineer's time to fix this for the community.
Have you tried the Cubie-Boards? They use only a 5V supply and may not have that problem. ITeadStudio also makes an A20 with 5V. I suspect that Olimex will not redesign the board for just a few users having problems.
Try to put such a ferrite filter on power cable just near A20 plug: http://en.wikipedia.org/wiki/Ferrite_bead
+1 misterzu, I purchased a bunch of chokes and and a couple of them made a huge difference, unfortunately there is still some interference. I will keep experimenting.
In the mean time, I am going to power the olinuxino's through the otg using a cheap dc-dc 5v regulator, similar to Lurch's suggestion. No interference at all and no metal enclosure necessary. And the lipo battery still charges, so I can cleanly shutdown when power is turned off.
https://drive.google.com/file/d/0B9Oqle-M8CLFSlNvS1JUbVZMZ28/edit?usp=sharing
The choke on this dc-dc converter must be better than the one used on the olinuxino board. It has 330 written on it. Maybe Olimex could swap in that choke for the existing one? Thanks again.
I found the solution in Olinuxino A13 board. It´s simple and cheap.
Solder an 0.1uF capacitor in parallel to C126 (electrolytic capacitor at entrance of MP1482DS). That's all.
The problem is that the interference is transferred to the DC source, this capacitor acts like high frequency filter.
In the A20 schematics, the capacitor seems to be C202
Good luck.
Hey all !
woodward i'm experience exactly the same problem as yours. Did you find any good solution ?
Here is a screenshot of with A20 on/off (yellow is A20 on, blue is A20 off : BIG DIFFERENCE!)
(http://s8.postimg.org/f15kwfgi9/screenshot_v_Md_Gn.jpg) (http://postimg.org/image/f15kwfgi9/)
So far I tried the ferrite around the 12v cable (no improvement), I put a 0.1uf cap in parallel with C202 (no improvement).
I thinking of using a DC-DC 5v converter on USB 5v...My project aim to use ham radio software with RTL-SDR dongle and it's really annoying.
What do you recommend ?
Thanks
Florian
I tried to find a solution in using a 12v -> 5v DC regulator to avoid the noise from the power jack alim.
I used the USB_OTG 5v connector but it is not really clear in the doc how much power can goes through this connector. I'm using some peripherals and the whole project is drawing about 1.6A at 5v.
Would it work with this USB_OTG connector ?
Thanks
The USB_OTG is doing strange. I'm using an Ipod 5V/1A charger through the USB_OTG. The green led start bright and then switch off. I have read similar behavior on the forum but with no clear answer...
So far the USB_OTG is not working !!!
In my last hope I tried the LIPO_BAT charger. I have a LIPO battery fully charge with 4.2V. When I used it with my A20 Olinuxino everything is powered well. When I tried to use an external power supply set at 4.2V the card don't start...
Does anybody have any suggestion ?
5V 1A isn't enough.
John
@John, I tried several power supply. One based on a LM338K a 5 A adjustable voltage regulators witch I set on 5V, the other one, the Ipod power supply was tested with board only, so I think it can handle it!
I tried to have a look at UART0 to debug the startup but I can't see anything wrong, BOOTP was one time looping...
I can't understand why the 5V/5A power supply could not power the board correctly...The green light light a few second then stop, even the Ethernet port stopped.
What could it means ?
There are limits to how much power a USB device can draw from the bus. A USB device playing by the rules (like the Olimex boards) must refuse to break the limit no matter how powerful the power supply is. Because the board needs more power than it is allowed to use, the AXP209 circuit shuts down the board.
For more details, see this post:
https://www.olimex.com/forum/index.php?topic=4058.msg17333#msg17333
(For the LIME2, but it applies to the MICRO too.)
@jmyreen thank for the explanations.
I understood the AXP209 can't give much power through USB_OTG, you meant the AXP209 is also working when we use USB_OTG as an imput for power supply ?
I powered the card with no peripherals, screen, only one micro-sd card and it's still not working.
The doc says :
QuoteNote that when powering the board from the USB_OTG, the power provided might be insufficient
to also power a bigger LCD connected to the LCD_con. However, this power option is capable of
driving the board when using external display connected to the HDMI connector.
Just for reminder, LIPO_BAT and 6v-16v barrel are working fine, but the in board power supply is noisy, that's why i'm looking for an other way of powering the board.
Quote from: splite on March 11, 2015, 03:55:20 PM
I understood the AXP209 can't give much power through USB_OTG, you meant the AXP209 is also working when we use USB_OTG as an imput for power supply ?
Yes, the AXP209 has three separate power source inputs: VBUS (from USB OTG), ACIN (external "A/C") and BAT (battery). It automatically chooses where it draws power from, depending on what's available to it. It can even mix USB and battery power, drawing as much as it is allowed from USB and the rest from the battery.
The APX209 does a lot more, too: is intended to be a complete solution for a battery powered device like, for instance, a tablet. The circuit also contains several voltage regulators producing the voltages the CPU and DRAM need, it functions as a battery charger, and it protects against over/under voltages.
Quote
I powered the card with no peripherals, screen, only one micro-sd card and it's still not working.
Well, I suppose it should work off USB power. I don't know why it's not.
On the other hand, looking at the schematic, I also don't understand where +5V for the peripherals is coming from when the board is only powered from USB OTG. Maybe I missed something...
Thank you for the explanations about the AXP209. I understand a bit more how this chip works.
I'm thinking I damaged something but I don't understand what and where, every single chip looks fine on the board...
I planned to power all the peripherals (mostly on usb) with the USB hub.
I have been doing a little test:
- Powering through USB_OTG nothing much is happening (not enough power ??)
- Powering through USB_OTG plus serial debug plugged on UART0, I can see the board starting up, looking for an DHCP server with BOOTP, taking an IP address, then try to load a register then stop and start again...Don't know why. But what I could observer, USB_OTG + serial debug is going farther than USB_OTG only.
Does the USB_OTG needs more than 5V to power the card ?
Quote from: splite on March 12, 2015, 09:27:53 AM
Does the USB_OTG needs more than 5V to power the card ?
No, the USB OTG is designed to to use standard USB bus power, which is +5V.
What I really wonder is: is anybody actually using the OTG port to power the board? There seem to be serious drawbacks to using only this way to power the board. Everything on the board that needs +5V gets it either from U14, a step-down regulator that is connected to the power jack, or U16, a step-up regulator that is connected to the LiPo battery. If the board does not get power from either of these, +5V is not available. Yes, +5V comes in on the OTG port, but it is only fed to the AXP209, and the AXP209 does not output +5V.
What this means is that the board's SATA power port does not provide +5V, nor do the GPIO and LCD extension connectors. The VGA HSYNC and VSYNC signals are pulled up to +5V, which is not there. The same applies to the HSCL and HSDA signals on the HDMI connector. The USB host ports do not work.
When you also take into account the current limit on the OTG port discussed earlier, I wouldn't recommend using the OTG port to power the board. Of course, this is only how I see it, based on my reading of the circuit diagram. I would be glad to have somebody from Olimex prove me wrong.
Edit: Never mind, I see it now. U16 works as a step-up regulator providing +5V from the AXP209 output. (And U14 is not a regulator at all. What was I thinking...)
@jmyreen thank a lot for all this explanation. True ! It would be really nice to have some info from olimex as well!
So what I can understand so far, is either the USB_OTG or LIPO_BAT are definitely not the best way to power the board.
I wonder how can I manage then to reduce the radio interference while powering the board with 12V.
I would love to have some info from @woodward who started this topic !
Still looking for a better solution...
For testing why not put a decent 5V power supply into the OTG. You want a good supply, not a cheap nasty one as otherwise you're wasting your time. A USB hub is not a good choice because you are wanting deliberately to take far too much power from a USB.
This still may not do what you want - reading the schematic and datasheets is the way to go. These boards are for developers! They expect to and need to read such things :) I know it's hard work but there it is.
I do hope you are SURE about grounding, shielding, where the noise is going / getting in as you're going to have yet more grief if you're wrong.
John
Hello
i just test using spectrum analyzer and i see also noise.
(my setup isn't proff - antenna it is a few cm wire near to power supply in olinuxino micro)
see video:
https://www.dropbox.com/s/a8wy76natvobskl/20150810_141357.mp4?dl=0
i must do some more test to confirm this.
Best Regards
Robert
PCB rev is F
Freq on spectrum analyzer is 85 - 115 MHz
You cen see signals from FM local radio stations when olinuxino is OFF
UPDATE 11.08.2015
video:
https://www.dropbox.com/s/lp1aih7yvp1d2k7/20150811_122602%5B1%5D.mp4?dl=0
olinuxino is powered from 5 V ext power supply via 5V_PWR connector (near the serial ata)
I solder 5V_E pad but when i add voltage to 5V_PWR connector i obserwe that voltage pulisng from 5V to 0V and micro can't start (red diode also pulsing)
after this i desolder L19 ad R80 and now it working.
I can't see now any noise on spectrum analyzer after first tests.
work in progress :)
UPDATE 17.08.2015
I just test rev G of the board and is much better. (without board modification from +12V)