Noisy Power Bus on Linino One

131 views
Skip to first unread message

B Bulghur

unread,
Sep 17, 2014, 11:41:02 AM9/17/14
to lin...@googlegroups.com
Hello,

I am prototyping a system with the Linino One (MircoDog).  My design incorporates a set of (low current) op amps to condition raw signals as input into the analogue inputs.  When I power via the Micro USB the power bus is quiet.  However, I'd like to use a 7805 as the regulated supply.  Using the regulated power supply as a source, noise results when the Linino boots.  In other words, it is quiet at first and then as it gets to work, noise results -- buzz, buzz, pause, buzz, pause, lower buzz, loud buzz, etc.  Clock noise is dirtying up the supply bus.
  • Scenario 1: Connect the +5 to VIN and used this regulated supply to power both the Linino and the op amps. 
  • Scenario 2: Connect the +5V to VIN and use the Linino +5V out to supply the
I have used all the necessary bypass, anti-ripple and other supply filtering throughout the circuit and changed supply and bypass caps and minimised the problem but not eliminated it.  This is on all Lininos and I can replicate the problem on various supplies and designs. 

Has anyone seen this problem?  Perhaps VIN is not the right way to inject the +5V into the Linino?  Any clues about how to think about this problem would be appreciated.

Thanks!

Bulghur, the Great Baltan

Master Pine

unread,
Sep 22, 2014, 10:48:20 PM9/22/14
to lin...@googlegroups.com
Hi Bulgur,

yes, VIN is the right way to inject the +5V into the Linino.

Power consumption of Linino is not constant, especially during boot time. If the power supply is not adequate, its output could be unstable and dirty.

What about your power supply ? What is it beside the 7805 regulator ?

Please, let we have a complete pictures in order to analyze your problem.

regards
Master Pine

B Bulghur

unread,
Sep 23, 2014, 4:25:21 PM9/23/14
to lin...@googlegroups.com
Thanks Master Pine, 

Attached is a schematic of the PS.  I am feeding it with a 9VDC 1Amp wall wart into the circuit.  I have tried a number of LF caps up to 1000uF on the input as well as HF of .1 and .01 on the input and output filtering in addition to some smaller caps to filter out the noise that is intermittent and in the 100+ kHz range.  I also tried attaching a 10nF to 1nF between the Linino +5V pin and GND with very little reduction in noise.  Finally, I connected a beefy bench supply and feed that into VIN.  On the bench supply, at start up the amperage draw fluctuated around 300mA and then settled into operation at 250mA.  The noise dirties up not only the +5V but the VIN. This noise disappears if I feed it through the USB and then tapping at the +5V -- which is typically loaded down to 4.4V.   The 7805 is the 1A version. 

Finally, I connected other regulators and wired them up to give 5.1V (to account for the voltage drop across the diode on the Linino VIN) -- no luck -- I got the same noise. 

I really appreciate your help on this...
 
Screen Shot 2014-09-23 at 13.10.11.png

Master Pine

unread,
Sep 26, 2014, 3:55:26 AM9/26/14
to lin...@googlegroups.com
Hi Bulghur,

from your schematic and you description it looks all fine.

The main difference between powering through USB or wall wart is the grounding of the system.

The USB GND of a PC is connected to power line earth, while wall wart isn't.

The same happens when you use a bench supply, because its GND is not connected to earth.

Try to investigate in this direction, doing an analysis of GND and earth paths and connecting GND to earth.

Do you have any update from your side ?

Let me know.


Master Pine

B Bulghur

unread,
Sep 26, 2014, 1:15:22 PM9/26/14
to lin...@googlegroups.com
Hello Master Pine.

Ah yes, GND.  So, that was my first investigations.  At this point, I have pulled the Linino out of the original circuit to test it with just the PS and the Linino and have fastidiously connected all the GNDs.  I then connected the ground bus to the GND on the scope and also tested ground continuity all the way back to the mains and they are all common, i.e. I can touch a ground on any part of the circuit, scope, PS, all the Linino GND pins and even the neutral to GND at the mains and they are connected with no more than 2 Ohms of resistance between points.  

With the Linino alone, I get several burst per second.  When I attach the DogUSB, it gets much noiser with what looks like data leaking into the supply rails. 

Finally, I did power this from a bench supply -- no ground issues there either.  

I am perplexed because looking at the Linino schematic we should not be getting this noise.  What do you think about the voltage drop across the zener at the Linino?   (I've included part of the Linino schematic.) Do you think that the diode protection is resulting in a voltage drop across the diode that serves a platform for the noise to not drain to ground?  

I've also included an example of the Linino only noise.  Scale is 100mV.

Again, I appreciate your help on this.  


Screen Shot 2014-09-26 at 10.05.55.png
photo.JPG

Master Pine

unread,
Sep 29, 2014, 5:34:05 PM9/29/14
to lin...@googlegroups.com
Hi Bulghur,

please, let me know the version of your Linino ONE.

Thanks.

Best regards
Master Pine

B Bulghur

unread,
Sep 29, 2014, 10:43:42 PM9/29/14
to lin...@googlegroups.com
Hello Master Pine, 

The version is v02 read from the bottom of the board, upper left hand corner.  

Thanks, 

Bulghur

Master Pine

unread,
Oct 8, 2014, 1:26:55 PM10/8/14
to lin...@googlegroups.com
Hi Bulghur,

the current consumption of the board is about 200 ~ 250 mA with a periodic peak > 1A due to packets transmission of Wi-Fi section.

The > 1A peak is the cause of the noise that you have measured.

To voltage drop is due to:
  • diode D2 (*)
  • copper section of power supply wires
(*) no more present in future versions of the board

To solve the problem I suggest you to bypass diode D2, has showed in the attached picture, and use robust wires to supply the board.
 
Best regards
Master Pine

On Friday, September 26, 2014 7:15:22 PM UTC+2, B Bulghur wrote:
IMG_0001.jpg

B Bulghur

unread,
Oct 15, 2014, 5:01:18 PM10/15/14
to lin...@googlegroups.com
Great!! Thanks for looking at this.  So basically, while it is transmitting, it is spiking over 1A.  This gives me some ideas on the PS.  I will bypass it and let you know how it is changing.  I did put .1uF caps at the pins of VIN and +5 to minimise noise.  

When are the new ones going to be available.  I think we just put an order for another five.  More jumpering... 

Thanks, 

Bulghur

Master Pine

unread,
Oct 15, 2014, 5:38:31 PM10/15/14
to lin...@googlegroups.com
Hi Bulghur,

from my last post we made some new and more precise measurements.

The base current is about 220 mA with peaks of 420 mA, duration 1.6 ms.

The peaks are recurring about every 100 ms.

See attached pictures.

Thank you for reporting the issue.


Master Pine
TEK00002.PNG
TEK00003.PNG

B Bulghur

unread,
Dec 11, 2014, 10:05:03 PM12/11/14
to lin...@googlegroups.com
Hello Master Pine, 

Are you removing the 'offending' :-) diode on your current production versions?  I am getting ready to produce a new set of prototypes (several) and I need to know if I need to allocated time for the assemblers to remove/bypass the diode.  

Thanks, 

Bulghur

Tony Jin

unread,
Dec 17, 2014, 10:55:58 PM12/17/14
to B Bulghur, lin...@googlegroups.com
Hi Bulghur , 

Here are the power consumption of Linino One and Linino One + Add-on , I provide them the power via the VIN pin using power supply  and I didn’t hear the noise as you mentioned. 
Due to the test result , I think the noise which isn’t from diode , can you check the inductor you’re using if it has enough rated current ?

Thank you.


Linino One 
Min:112.974mAdc
Avg:127.104mAdc
Max:377.393mAdc

Linino One + Add-on 
Min:281.195mAdc
Avg:298.396mAdc
Max:566.460mAdc



Tony Jin 金禾桓
 


dog hunter Inc. ( 多航有限公司 )
114 台北市內湖路一段91巷17號8F-4
8F-4, No.17, Ln. 91, Sec. 1, Neihu Rd., Neihu Dist., Taipei City 114, Taiwan (R.O.C.)
Tel: +886(2)87518533
Mobile +886-926-594-689

--
You received this message because you are subscribed to the Google Groups "Linino" group.
To unsubscribe from this group and stop receiving emails from it, send an email to linino+un...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Reply all
Reply to author
Forward
0 new messages