First 100 Units Checklist: https://predictabledesigns.com/must-do-first-100-units-pdf-yt
Reliability Testing Tool: https://predictabledesigns.com/reliability-tester-tool-signup-yt

Here some things I learned working in a start-up.
- Find potential customers and build some "minimal viable products" so a device that shows the core idea of your product. Support them and listen to their feedback. this will tell you a lot about if the product you want to make is what the market wants. This will give you very valuable feedback and improvement points.
- Pick your battles. Prioritize what you need to improve change etc. first.
some manual rework on your product needs to be fixed, but if it is only a small time penalty in production, you may need to spend your time elsewhere. Maybe it needs to wait until you go to bigger numbers.
- Be careful with tight tolerance and high specs. This one can give you so mant head aches. Tight tolerances make your parts very expensive as well as the assembly. instead look at what you actually need.
High specs. sound nice, you think it will help sell your product, but as soon as you claimed a spec. customers will not be happy if you lower it. Never take the best result in the lab as your goal. you may not be able to get there all the time.
Leave some headroom, and once you got your process under control, you can up the specs bit by bit.
- Put the brake on new features! you start to make a product and you have so many ideas. Ambition is high to make all these nice thins happen.
But in the mean time you do not take the time to properly finish all the other things. They sort of work, and the last steps are not nearly as exciting as all the things you want to add.
As a result the customer is left with a badly working product with lots of small issues that ruin the usability, or at least make it annoying to use.
- Adding to the previous point, do some updates where you only fix things every 3-5 design changes. This is specially true for soft and firmware. Looking at you Microsoft, who has not been able to completely transfer to a new UI since Win8 (hello control panel, my old friend🙄)

I just needed this video as im between prototype and first batch ... thanks!

Valuable content.  Thank you!

“Thank you for your efforts and the valuable information you provide. Could you please make a dedicated video explaining the necessary checks to perform on products before selling them

Great data.  Thx.

Good advice.
We had to trace cut and jumper wire (something like 1 cut, 2 wires) manually on 4000 units once... 😕
Admittedly, it was for a problem that only showed up in the field... backpower issue during low-power/shutdown mode, drained the vehicle battery between engine starts overnight/weekend... this was in Canada in the winter... Rrrrrr...rrrr.rrrr...click click click 😢
I'd add "pilot 10 units in the field first" and probably "in summer AND winter".
Oh, and perhaps most critically:
NEVER let MBA types rush the process. They will NEVER take the blame, they'll throw oyu under the bus at any opportunity.

Thank you

love you, man :D

Very good information thanks.

Thank you for the video

I was a software guy and was mesmerized  by electronics and finally  get back my EE degree. Making electronics  is not as comfortable  as ordinary software which allows failures😢

Any thoughts on resin printed enclosures before scaling up?
They have much better surface finish than FDM 3D printing

Another route for a product enclosure is to simply use one of the major electronics enclosure manufacturers. They already have a massive selection of enclosures, and many will CNC machine a stock enclosure for you. You can then combine your custom machined enclosure with a decal from a print shop to make a professional product that doesn't require an injection mold at all.

Make video about low power battery powered devices. 3.3v from battery: dc-dc vs ldo vs direct power. Battery monitoring: two mosfets and divider vs fuel gauge. Charging and power sharing: cheap charger vs pmic.

I remember having a floating reset pin on our first few thousand PCBs and more than half were already delivered before the quirks of it happening on the field were noticed. We simply didn't have any strong enough ESD events happening in our assembly and the EMI tests were done with firmware that would not log such resets either

My friend John , you have mentioned ( customer ) in many places in your product development path or cycle. Such customer characteristics couldn't be found easily all around the globe . Most cases customers need for final last revision of a product without any defect in design or production and no one would ever try his luck to wait you for finalizing your product... You are just dreaming for such customers , mate ! ....

Small numbers such as 10 tend to be built by hand, whereas 100 might be auto assembled. Do you think the auto assembly needs a pilot ?  The small runs are very expensive as they take a full setup but hardly produce any product. But 100 faulty products can be expensive too ..

'teel the money make the rules: in a startup (long long time ago) we went from 1 working EPROM MSP430 strapwired to the board to the 500 OTP MSP430 units then to the 1500 units! All of those 2001 pre-assembled PCBs had a defect (one jfet instead of a npn). My role was to streamline and extend to multitask hand production is such way no errors would go undetected before the product was put in the box. When the cash flow runs dry, frugal innovation is the best ally... Learn the qualities and defects of the humans involved, create and deploy a fast perfectly reproductible remanufacturing process and stick to it => over the 2000 boards, only 7 were discarded for defective PCB.
That jig is sci-fi, hardwood with precise pogo pins going into 1.5mm holes I drilled myself, twice, in case one fails. It provided programming, accuracy adjustment for sensors and full board functionality testing.
Abuse: smash with full force the product on hard cement to simulate an angry customer, no visible scratches should appear while maintaining the accuracy.

There's also the pitfall that you have your first prototype or 100 production enclosures made by a different manufacturer, or a different process, or a different material, than the next 10,000 ones. That 10,000 might end up in the bin because of tiny manufacturing tolerances.

Sometimes, you can design enclosures which can tolerate some deviation. But if you target a specific form factor, and you end up with 10,000 enclosures where the PCB rail is 0.2 mm (8 mil) too tight for the PCB to slide in, you are in real pain.

I like electronics. I hate enclosures.