What would you spend £1500 on?

Just the core hardware for Hermione?

  • CF Frame £240
  • 8 x T-motor P13x4.4 CF Props £145
  • 8 x T-motor U3 Motors £515
  • 8 x T-motor AIR 40A ESCs £215
  • 2 x Gens-Ace Tattu LiPo £180
  • Garmin LiDAR-Lite V3 £115
  • Scanse Sweep £270

Or 3 complete kits of these?

DJI Spark

DJI Spark

My advice, don’t try a DIY drone unless you are stupid as I am!

2 thoughts on “DIY vs DJI

  1. But let’s be honest here: you’ve built an octocopter with premium T-Motor components. Your vehicle has better redundancy in the face of a motor failure *and* substantially more lift capacity. *And* you’ve added sensors that are better than what the Spark is using. If you had instead built a quad, stuck with a single LIDAR sensor, and bought a cheaper CF frame, the price of your DIY unit would look a lot more like the price of the Spark. The things you are comparing aren’t even remotely equivalent.

    • Hi Lars, Fair points!

      Under the covers, the main point of this post was to defer people who want to DIY a quad as a cheaper way to do racing etc and instead point them at the DJI. On the other hand, if DIY is what’s wanted, then fair enough – DIY beats DJI all the way!

      For me, the whole point of this project was to push a Raspberry Pi as far as I could to make a fully autonomous python drone. So my aims are very skewed compared to most DIY droners!

      If you look back in time though, 5 years ago, I was using DJI frames, motors & props for my true quadcopter, and the price was much cheaper. Back then, I used the Invensense MPU6050 which cost me more than the Sparkfun breakout of the MPU9250. Let me know the better IMU sensors that cost less the £/$/€20 ; part of the reason for all the extra sensors is that I’m pushing way beyond the limits the Invensense ones can take (e.g. double integration of (accelerometer – gravity) to get distance is only accurate for a few seconds before accelerometer resolution and accuracy errors integrate twice to product big distance errors. Hence the LiDARs. I did try URFs, but none supported I2C to 400k baudrate required by the IMUs, The MPU9250 is still there, but the expensive sensors are backing it up for ‘longer term’ flights

      The reason for the Octo was simply if it could be done with my Raspberry Pi; the reason for my change to T-motor ESC / motors / props is I then I knew any crashes I see could only be blamed on my software. For the quads I used Chinese CF props costing a tenth of the T-motors. It’s the lower 4 motors on the X8 that rule these out – they just don’t fit!

      The big expensive frame is to make enough space for all the sensors and the Raspberry Pi. OK, and by that stage, looking cool mattered to me just as much as flying well!

      I think both our POV are correct: if you want to build a high quality, pretty, autonomous quad / octo to use, don’t. Just go for a DJI Spark or Mavic (I’ve got a Mavic).

      But if you just want to understand how Quads work through DIYing one, you could probably build the whole thing for the same cost as the Spark



Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.