3D Robotics

The paradox of DIY electronics warranties

3689388049?profile=originalOne of the trickiest parts of selling electronics kits, as we do, is how to handle customer support and returns. I'd estimate that 90% of the problems that users have with DIY Drones gear is due to user error: soldering mistakes, not following the manual properly, issues with their own PC/Mac, or simply misunderstandings and confusion about how to use the products (no doubt sometimes due to documentation that could be improved).


But perhaps 10% of user problems (and only a tiny fraction of all orders, of course) really are due to defective boards, which somehow made it out of our factory with a problem. Although every board is tested and demonstrated working before being shipped, there are some things that our test code doesn't check, such as what happens when that particular board is vibrated or subjected to cold temperatures. And there are other products that we simply resell and don't test, which could have come from their makers faulty (a recent batch of FTDI cables from China had the wires connected in the wrong order, for example).


So how to handle returns? The normal answer for companies that sell finished goods doesn't work for DIY kit sellers, since it's hard to tell whether the problem was due to a manufacturing error or user error. So most such retailers have a clause in their terms of service like this one, from MakerBot (which they discuss here), that pretty much says that the moment you power it on or touch it with a soldering iron, the manufacture's responsibility ends:

Returning Purchased Items. MakerBot Industries accepts returns for any unaffected item within 14 calendar days from the date of receipt of merchandise."Unaffected" means a device has never been assembled, powered up, programmed, or otherwise changed. MakerBot Industries cannot accept returns on purchased items that have had electrical power applied to them, or been otherwise programmed, changed, or affected.

The paradox, of course, is how can you know it's defective before you power it on? In practice, most of these companies will accept returns of products that have been powered on, but typically not if they've been soldered. We wrestled with this and decided to be more explicitly generous in our terms, while still limiting our liability:

Defective products manufactured in house will be replaced within 10 days of receipt, with the same item. All goods are checked before they are sent out. In the unlikely event that goods are faulty, they must be returned, to the DIY Drones store at the address listed on your order invoice. The DIY Drones store will not accept goods that have clearly been used more than the amount needed to evaluate their functionality. If the goods are found to be in working condition, and the lack of functionality is a result of lack of knowledge or assembly error on the customers part, no refund will be made, but the goods will be returned to the user at their expense

So this is a judgement call. Most of the time, when people ship us back "defective" boards, it's clear that they were simply soldered wrong or not defective at all (the user had plugged in their RC gear wrong or loaded the wrong code). In cases of soldering errors, if it's easy to fix we might do that, but if it's really been bungled, we may just send it back. We try to error on the side of being as helpful as possible, but there are limits to what we can do.


We've invested a lot in customer and technical support over the past few months and now have full-time employees doing both.  As the MakerBot post linked above points out, this is uncharted territory for open source hardware companies like ours, and we're all trying to find the path that is right for the community but also allows for sustainable businesses. It will evolve over time. But I wanted to let our customers know how we're approaching this, and I welcome suggestions on how to do it even better.

E-mail me when people leave their comments –

You need to be a member of diydrones to add comments!

Join diydrones


  • 3D Robotics

    Donald, the lock bit is never displayed in the GCS (although it will be in the next version of the APM software, which supports MAVlink). That doesn't mean you had a defective unit.

  • I'm on my second MediaTek GPS. The first had the longitude coding error. Although it did send data to the ArduMega board at the proper rate, the "Lock" bit was never set in the data in spite of a solidly lit LED on the device itself. After numerous failed attempts to reflash the unit, I returned it for one that I thought would surely work.  It didn't. This one doesn't communicate with the AP board in spite of having the proper 5 V at the MediaTek board and 3.3 V at pin 1 of the device. Any advice on this would be appreciated, otherwise I may exchange it a uBlox.
  • Thanks Johann, I'm getting that bug addressed, it's in the nmea library, I've posted in the forums if you'd like to know more, this probably isn't the right place to get into it.
  • Hi Johann,

    I don't think you read my comment correctly, please read again :)
  • I think the key Chris, is what the hardware is advertised to do, vs what it does. I do agree open sources is typically for people who want to help build a community project, and it's great for that. But if someone sells for instance, autopilot hardware (to keep things in context) and that hardware is said to support features x, y, and z through the use of software provided with it, and that software really only supports features x and y, then the provider of the product is liable for making feature z work, even if the software is open source, and even if the software isn't theirs. If they can't do so, then they must alter the advertising, and compensate the customer, this is a situation where a refund would be completely valid.

    One way around this is to offer (and advertise) the hardware as simply hardware, a combination of components, and not as an auto pilot. Or, as I believe diydrones to do, to advertise their feature set accurately, and make it known that some features may not be available until some time in the future.

    This is my understanding of consumer laws in America, maybe I'm wrong, and someone correct me if I am,
  • 3D Robotics

    Jon, that's a fair point, but things get ambigous in the "atoms/bits" divide. Open source hardware companies sell  atoms--electronics, by and large--and they can only be guaranteed to physically work. The features, however, are enabled by the software, and that is free and changing (ideally improving) all the time.


    For example, if you bought a MakerBot a year ago, as I did, it would only work if you knew how to hand-tweak incredibly obscure code--and almost nobody did. So most people couldn't get it to work. But if you simply let it sit for a year and then downloaded the code again, it was much improved and now works easily for most people. So the promise with open source hardware is that the hardware will work and the software will improve--and given the way open source projects work, it's almost impossible to guarantee more than that.


    Open source is typically for people who want to be part of a community development project, for better or worse. Along with a price that's often 1/10th that of closed source comes the assumption that people are going to participate, give back and be tolerant. If they're just coming for the cheap price, they'll typically be confused and put off by all the updates and change that comes with the territory.

  • I'll correct myself slightly and include that the products need to be exchanged for with money. Obviously FREE open source products are an exception. But products which include open source software or hardware, that are sold, still need to abide by all the same consumer laws as any other product.
  • You're right Johann, I do believe there is a misunderstanding about what open source means. It seems some companies assume it's a way of not being liable for the way their products are being advertised. (not accusing diyd). This is not the case. Nothing, not open source, nor diy, can wave the liability of an advertised product not doing what it's advertised to do. Now if it doesn't do what it's suppose to due to a some kind of user error, thats a different story, but you can't use "Open source" as "We're not liable for our product not doing what we said it would".


    Again I'm not accusing diydrones of anything, I think their products have been advertised quite accurately.


  • I honestly dont think DIYD could do much more. The key to this is end user expectation. Its important that buyers understand they are buying components that form a radically evolving system and NOT a plug and play solution. Perhaps a warning when purchasing? Or how about a DIYD board that teaches all the basics (soldering, loading code, powering the board correctly, maybe interfacing to an RC system) make it as cheap as poosible and offer a refund of the baord's cost when upgrading? Better to bust a 5 dollar board than a 60 dollar board!
  • I think WARNING placed in the appropriate places would help a lot ,more pics of port settings, baud rates ...and opamps on the power grounds , make the electronic part as plug and play as possible...and how about a active x program that would set the ports to the right ports and settings or keep software set on default settings...I put my aurduheli together today only to realize the pitch was reversed ,try load new code avr dude problem, took it all apart to find i was using wrong port : (  back together waiting for help on the reversed pitch... :)
This reply was deleted.