mbed Application Board Design 101 - The need for reverse-polarity protection!

Consider this scenario in the software world...

You try to log on the Apple App Store unsuccessfully. After three attempts, Apple notifies you that not only has your account been closed but all of your electronic purchases have been erased from the Apple iCloud....

My scenario... apply reverse polarity to the mbed Application Board.

Poof!

• mbed NXP LPC1768 destroyed
• mbed application board destroyed
• sparkfun RN-XV WiFly module destroyed

It got me thinking after the accident, how an experimental design board for labs, classroom and hobbyists could ever do designed so badly. I'm not a practicing board level designer but I do know poorly thought through designs when I see them, having brought up boards as a firmware engineer. When you look at the BOM for this board and you look at the exotic components such as the joystick, temp sensor and RGB LED, you have to wonder how much a simple diode, bridge rectifier combination or FET configuration would have added to the cost.

Well so much for my experimentation with the mbed platform and associated development environment. I really don't fancy spending another $124.95 at AdaFruit for the Xively Jumpstart Kit and $34.95 at sparkfun for the RN-XV WiFly module. Perhaps the designer of this board had really good reasons for omitting the most basic of reverse-polarity protection circuitry... I'd love to hear his/her reasoning!

livid!

Hey Graeme! Sorry to hear you destroyed your kit. I happen to have a spare I can trade you for. Let me know if you're interested. BTW - The parts you listed are all powered via USB so external power wasn't necessary unless you have another peripheral attached not mentioned above.

Considering there are a whole bunch of methods to power it, revere polarity protection isn't as trivial as you make it be. Also what I do remember from it, it might not always be 100% clear how the LPC1768 should be placed in the application board, however iirc doing it the wrong way isn't ideal, but shouldn't blow up everything.

So that makes me wonder out of curiosity, how did you power it? The two normal ways would be either USB, or via the power connector. Both are just standard pinouts, so to be honest shouldn't need reverse polarity protection. If I swap the power and ground of my USB charger and connect that to my phone I also have serious doubts if it is going to survive that.

If you for some reason decide to power it directly via the power pins of the LPC1768 and swapped two pins: It is really a special board if it survives that. I don't know of any of the other ARM boards which do have protection there, and also for example Arduino boards do not survive that. The rugged Arduino does, but that one is specifically made to manhandle. But default Arduino's do not survive this. Also as you found out, also the WiFly modules (or any other) do not have reverse voltage protection.

Of course it is really irritating for you. And it seems Sam has a very nice solution. However I also think the way you made your post is aggressive towards the designer(s). You act just like it is the standard to add that, while it clearly isn't. Not to mention there are a whole bunch of other options I can give you to blow up your board by miswiring. I really hope that post was made when you just blew it up and were kinda emotional due to that...

Erik - for the record, I didn't say it was trivial - it just wasn't thought through in the context of the application. And no, I didn't reverse the orientation of the LPC1768, so please... don't put me in that bucket. The board was powered with a 9VDC 1000mA out, center pin negative power adapter via a 2.1mm Jack - Micro-USB Plug Adapter Cable. The power was required as I was going to try different radios in the USB socket. "aggressive towards the designer(s)" - really? It's more frustration at the designer/s. Consider a University lecturer kitting a whole lab in a EE/Comp Sci teaching scenario and students grab the first power supply they can find... Poof!

Sam - thanks for your prompt response and offer.

I just checked, it is explicitly shown on the PCB that it is barrel negative. You made a mistake, I can understand you are frustrated after something like that. I can understand you wish it could have handled it. But I reread your post and still consider it agressive, where you blame the designer who follows standard practise of being incompetent (yes thats what it comes down to).

Erik - I never disputed my mistake, I dispute who would design something without considering this as a potential outcome powering the application board. I don't understand your last sentence.

He guys, just to cool things down, I don't see any aggression here, just useful discussions.

The versatile power arrangements of the mBed was a big area of concern for new users (like me) when the product was launched and the documentation in beta. Everyone wanted to understand what happened with USB and external power plugged in, but it is nicely designed and versatile and works.

My big worry has always been the reverse power issue and I hate barrel connectors, particularly those where the end can be reversed. I too have questioned in my mind why reverse polarity protection was not included.

On the mBed module it probably as simple as 'space', but if I have a development or application board and apply power to that surely reverse connection should be incorporated at the barrel connector on that? So I'm not trying to be aggressive, but practical here and if the reverse connection isn't catered for here (I haven't checked), I question the design and ask 'why'?

Also some of the students I have worked with from around the world have great difficult understanding the polarity of barrel connectors! It would seem they interpret statements like "The polarity of the barrel is negative" as the plug on the end of the wire is the barrel connector so with a negative polarity this means the pin is negative as the outside 'the barrel' is ground!!! I always measure mine for correct polarity and voltage (and unloaded volts often raises a few eyebrows!).

@Graeme, sorry you had this expensive issue, but it certainly renews our caution on this topic. Thanks

Kevin - thanks for your input. Hopefully this might speed-up a re-spin of the board or at least serve as an application note for experimenters, hobbyists, innovators and students alike.

I wouldn't blame the mbed guys for this. I would go one step back and question who was it who had the bright idea to design power pack connectors that are identical except that they have reversed polarity. I am always *EXTRA* careful before I connect anything up to one of those things.

Is there a good reason for this because I have never been able to understand it? Lucky that the same guys weren't involved in the design of 240v plugs and sockets.

To be fair, the vast majority of dev boards do not have reverse polarity protection. These type of boards are not designed for consumers, they are designed for engineers and developers who know what they are doing, or are learning :)

For a budding electronic or embedded developer, a good habit to learn is to check and double-check circuit connections, a bad habit to learn is to plug anything in and expect it to work. It's unfortunate if that first lesson proves costly, but it will save many $$$ later.

These boards have many connectors and exposed pins many of which would cause damage if connected to the wrong voltage - it's not practical to protect them all. I think it is a reasonable expectation. If it was a consumer item, or an industrial controller, you might expect every exposed connector to be protected up to +/- 48V.

The fact is customers demand cheaper boards, not boards with more protection, so designers don't have much incentive to add components many users will think are unnecessary.

The usual result of mis-wiring is a dead board, which is not really a costly option, compared to electrocution or fire, which is obviously a hazard with main supply type voltages, and where preventing mis-wiring by end users is mandatory, although clearly the electrician doing the wiring is expected to have the appropriate training and certification.