Power problems with OlinuXino WIFI

[klotzmando]

I am having the worst luck getting started with the OlinuXino. First I had a solder bridge that was distorting the colors on VGA. Now I am having problems getting the board to stay programmed.
The board seems to only like to be powered with a 6 amp 12 Vdc supply.
A little background.
The only things I have connected to the board are:

• VGA monitor
• Logitech USB unified Nano adapter for keyboard and mouse in USB host 1
• PL2303HX USB to TTL serial cable connected to UEXT pins 2-GND, 3-TX, 4-RX

I use PuTTY at 115200 and can watch the console messages as the board boots.
The only way that I have been able to successfully power the board up is with a 6 amp 12 Vdc power supply connected to the barrel jack on the board.
If I use either of two different 2 amp 12 Vdc power supplies or the OTG connector with a powered USB 3.0 hub, the board begins to boot, but soon reports a corruption of NAND and stops booting.
Once the board fails to boot, it will not boot to start screen with the 6 amp supply anymore until it is reprogrammed.
I am not able to reprogram the NAND with any supply other than the 6 amp 12 Vdc supply. It sometimes takes several attempts to reprogram the board even with the 6 amp supply.
All three of the 12 Vdc supplies read 12.4 to 12.5 volts when connected to the board.
On last note: I took the board to work and hooked it up to a bench supply set to 12 Vdc and it behaves the same as the 2 amp supplies. The desk supply is on a standard calibration schedule as we are a test lab so there is a very low chance that it could be a bad supply.
Other than just use the 6 amp supply, is there anything I can do to make this board more reliable?

[gintaras_bar]

Have you measured how much current your board is drawing from power supply? If you can't operate with 2A power supply, it seems suspicious - something is bad with that board.

And have you checked board for hot spots? May be some component is damaged and you will get smoke released soon.

[Tele]

Have you measured how much current your board is drawing from power supply? If you can't operate with 2A power supply, it seems suspicious - something is bad with that board.

And have you checked board for hot spots? May be some component is damaged and you will get smoke released soon.

I agree, 72W seems a little bit too much. It has to appear as heat, somewhere.

[klotzmando]

No hot spots, max current I have seen with a meter is 430 ma, so it makes no sense at all that I need a 6 amp supply. All supplies use similar 2.5 mm barrel plugs. the only difference I can see other than power is that the 6 amp supply does not use a grounded AC cord.
What really confuses me is that with a calibrated desk supply capable of well over 8 amps the board does not work
I think this is just a bad board and will try to return it to supplier for an exchange if that is possible.
Fed up with this one and won't spend anymore time on it.

[jwischka]

Have you measured how much current your board is drawing from power supply? If you can't operate with 2A power supply, it seems suspicious - something is bad with that board.

And have you checked board for hot spots? May be some component is damaged and you will get smoke released soon.

I agree, 72W seems a little bit too much. It has to appear as heat, somewhere.

Not necessarily. A supply's maximum rating doesn't necessarily tell you about the stability of its voltage supply. When the board is booting and things are coming on, you can have something that looks like a short, until enough charging current has been supplied. If the supply isn't stiff enough to deal with this, your voltage can collapse across the bus, and trigger unexpected behavior.

Now, that being said, a 12V 2A supply should be sufficient. What kind of supply are you using? I'm powering my board with at 12V/2A supply from a hard drive, and haven't had any problems. My guess is that you don't have a logic analyzer that you could hook up to the power line, correct? It may be difficult to really debug what's happening without some extra equipment.

[klotzmando]

Thank you for your response.
Yes I do have a logic analyzer, but like most simple logic analyzers it only shows on and off states.
That means that the changes in the levels of voltage will only show as on and off states. It would be certainly something I would be willing to try.
Not sure how that would help.

• What would I expect to see?
• Where would I be hooking up?

By The Way:
I have hooked this board up to a lab bench power supply which has recently been calibrated and is capable of 8 amps. The board will not boot or flash using that supply either.
I thought I had mentioned that but that could be easily missed in all the other details.
This same supply is used to power a PC104 board and a Freescale X12 board on occasion, so the question of transients at start-up would be very low on my probability scale.

[jwischka]

I would check and see if you have a voltage depression - any transition state on the power bus to whatever "low" level you can set (as high as you can set it). In truth, some boards may fail to start properly if the voltage is even a bit low, though they may be fine with significantly higher voltages (we have some PC104 boards at work that are 12V nominal, run fine all day at 14, and won't boot at 11.5). As far as actually setting it up, I'd put the logic analyzer on +VCC and ground, and then see if you get a high->low transition with the "low" being as high as you can set it. Make sure, obv, that your logic analyzer can take 12V.

What would really be more useful, probably, would be a full scale data acquisition system (I'm thinking something like an NI CRIO system if you had one) where you could write a program to trigger a waveform capture for an arbitrary voltage disturbance.

Now, all of that said, if you've hooked it up to a lab grade supply, I'd agree that voltage transients are unlikely. Since you have that supply, do you have the ability to bump the voltage a bit (say 12.3-12.5V)? I wonder if running it a bit hot might help. Alternatively, since the board should also take a 5V input, what if you ran it at 5.5V, or similar?

One other idea would be to connect a capacitor in parallel with your board and the power supply. That should effectively filter out any transients, if they exist.

Thank you for your response.
Yes I do have a logic analyzer, but like most simple logic analyzers it only shows on and off states.
That means that the changes in the levels of voltage will only show as on and off states. It would be certainly something I would be willing to try.
Not sure how that would help.

• What would I expect to see?
• Where would I be hooking up?

By The Way:
I have hooked this board up to a lab bench power supply which has recently been calibrated and is capable of 8 amps. The board will not boot or flash using that supply either.
I thought I had mentioned that but that could be easily missed in all the other details.
This same supply is used to power a PC104 board and a Freescale X12 board on occasion, so the question of transients at start-up would be very low on my probability scale.

[klotzmando]

All very good suggestions, thank you. For now since I have only had the board for 2 weeks, I'm loathed to spend a great deal of time trying to make the thing work. I will try to return it for an exchange with the company from which I purchased it.
I have other problems with this board since I got it as can be seen in my other posts.
If that fails I will attempt these steps. Thanks.

[klotzmando]

Final update and disposition:

Ok The company I bought the board from says that Olimex requires that I deal with Olimex, Olimex says that the re-seller sets the policy not Olimex. Who cares, except I am always suspicious when companies start pointing fingers at each other.

Also Olimex requested that I send the board back to them and if they find that the problem is a manufacturing defect they will replace it. That's actually a great offer, except I am in the USA and they are in Bulgaria so its international shipping charges plus insurance, a trip through customs and more time. (I have already spent two weeks just trying to get this board to work)

Heck the board didn't cost that much to begin with, so I guess it goes in my scrap pile.

Hey thanks for the help anyway. Who knows, once I finish this project with a locally supported board I will dig it out and try to make it work.
"Goodbye and thanks for the fish".

[JohnS]

As you're in USA surely your contract to buy the board was between you & reseller so it is up to them to honour it.  (Nice of Olimex to go further but it's the reseller who has the problem, legally.)

Who is the reseller?  Others may wish to avoid them!

[klotzmando]

I don't think it is appropriate to name the re-seller. It was not their final inspection or product that failed.
The fact that they have been instructed to refer their customer to Olimex for resolution is not a breach of contract.
I realize that each persons experience is colored by their expectations and in this case I have been disappointed by the process. Now that I am aware of the process, I will simply choose to not allow it to continue to disappoint me.
My only goal was to let those who were willing to help me know what has happened and to thank them.

[JohnS]

Any such referral is at your option, surely, not the reseller's, but if you're Ok with it then hey it's up to you.

I wouldn't be willing to be treated like that as I'd find it much more convenient to deal with a seller in my own country and return anything likewise.  They can send it elsewhere (return to maker etc) as they choose.  They're likely to be far better at customs doc than me, too!

Having to return something is usually no comment on the seller's inspection process (if any).

John

[Lurch]

As John said, it's your call. I would definitely name a lousy dealer who doesn't uphold contracts. He gets a good slice of the profits for selling it - including the cost of returns and repairs.