Help needed in getting from concept to proof-of-concept

[oodavid]

Morning,

Quick intro - I'm a seasoned programmer specialising in making business-operating-systems for light industry.

I'm totally green WRT electronics, but I have a commercial idea borne of my freelance work so have been teaching myself about schematics, PCB design, prototyping, manufacture etc. While I've gained a lot in the past few weeks (and my god is it fun!), I'm keen to get some assistance to speed things along.

The first bit of help I need is to validate the concept and get some advice in estimating costs for prototyping, costs for a batch of 100 units, then 500 units etc.

My goal is to make:

• A head-mounted HD camera recording 1080p 30fps video + microphone audio to an SD Card (with wide-angle lens, much like an action-cam)
• An external power supply to keep the weight down and the operating time high, (6 - 10 hours) (the user will wear this on their utility belt)
• A smaller internal battery in case the external power cord is accidentally pulled (to give the operator time to re-attach or for the device to shutdown)
• Two LEDS - green for power & recording and a red LED for errors & warnings (ie: low battery, low memory etc.)
• 4 additional sensors that log their data to the same SD card, each to a different text file (the data should be time-stamped so it can be tallied up with the video during analysis)
• A black-and-white screen - to show live sensor readings and power (no need for video viewing)

There's a few obvious omissions such as internal clock, processor, ram, flash memory for OS etc. I've a more detailed document showing an idea of the form-factor and explanation of the 4 sensors, but I'm not quite sure where I stand regarding requirements for getting NDAs signed etc. but certainly the above should give a rough idea of the type of project I'm looking at.

I'm based in Newcastle, UK and am looking initially for some consulting advice about the project, then (if the idea floats) onto the next stage, prototyping!

I have no idea if a prototype will cost £1,000, £10,000 or £50,000 - whatever it costs I'm happy to have a serious crunch of the numbers.

EDIT - handy thread regarding video

[steve207a]

Hi first you must have a list of the parts you require next   you need to find the best place to source the parts you need at the best cost and you need to know if the source is reliable ...you may not be able to source the parts from one place and you may need many places to obtain the parts also the amount ordered will reflect in your price costs Can you source a ready made unit from china or part of the unit  this will keep your costs down and speed any prototype build .........a good free design software is Design Spark from RS components you just need to register with them totaly free this will help you as can do schematic /pcb  and 3d picture of pcb autoroute of pcb ect

and golden rule is price high you can always come down if needed
good luck... just ask my  advice is free

steve

[Ian.M]

Meanwhile in Shenzhen someone has probably already got a hackable battery powered SD card HDCAM on the market that already logs a serial data stream (GPS), so if you dont speak fluent Cantonese you will be at a severe disadvantage, and when you get it to market, if its popular, you may have cheap competitors within a few months.

[nctnico]

Meanwhile in Shenzhen someone has probably already got a hackable battery powered SD card HDCAM on the market that already logs a serial data stream (GPS), so if you dont speak fluent Cantonese you will be at a severe disadvantage, and when you get it to market, if its popular, you may have cheap competitors within a few months.

I doubt it. I have lots of work getting designs done which the Chinese can't... Just look at Siglent and Rigol: they can't get the last 5% right so their market is limited to the niche which isn't profitable to begin with.

[John_ITIC]

For reference; count on a cost of approximately 15K USD per month in ""senior" consulting cost and several months of development work to get a prototype. So, in short, your engineering cost will be closer to $50K USD rather than the other numbers you threw out there. You can likely pay less for a junior engineer/consultant but the cost will likely be the same due to less efficient work.

Essentially, you'll need a CMOS sensor, and an FPGA running some embedded CPU so you can capture the raw data, compress it (or else your SD card will  not be able to keep up with the data stream) and write it to SD-card using a FAT32 or exFAT file-system. Note that you will have big problems with getting that much data onto an SD-card since you will not be able to find a reasonable cheap driver for it. The UHS bus is patented and requires an insanely expensive royalty payment. Therefore, you will pay an arm and a leg to be able to use an SDXC card, bandwidth wise.

Overall, I think you can put together such a system but the cost will be very high so you will need to have a big market and sales organization to recoup the development cost. If you need a consultant to work with you, I'm available. PM me if interested.

Due to size and cost, I suspect some form of ASIC is used for similar high-volume products, like go-pro and the like. General-purpose, off-the-shelf parts such as FPGAs require quite a lot of power and take up lots of space so I'm not sure whether you can put everything together in a small enough form-factor to make it work. Research would be needed to flesh this out.

[m98]

No need for FPGAs or custom chips: http://www.ambarella.com/products/sports-cameras
But actually I wouldn't even want to get into that camera stuff, and just use something like the GoPro Hero and design an accessory for it, that uses the available accessory-bus.

[Koen]

Hello, is there more to it than concurrently recording video and sensor data ? Are the sensors highly specialized ?

[Smokey]

Kind of like these?

http://www.liquidimageco.com/collections/featured-products/products/model-368-torque-series

http://ivuecamera.com/

http://gecocam.com/the-product/

http://www.rollei.com/products/actioncams/actioncam-glasses/rollei-sunglasses-cam-200,p100264

[kmossman]

You are making it much too complicated for a POC, or even a first model.

First, get the Chinese equivalent of the GoPro - it is dirt cheap and feature rich.

Second create a sensor log system that logs the data.

Yes you have two SD cards.  Big deal. Getting it working first.  Marketing counts more than finesse.