Requiem for the splicer or anatomy of the failure.

Submitted by kirill on Wed, 03/02/2016 - 23:39

At last, more then 2 years after I have invented the splicer I must declare this project as partially failed.

Despite the silence on the site and very serious stress in my life, there was some movement here, I made several units that where distributed in Switzerland -to friends, FabLab Zurich and Hochschule Luzern among the others- but I failed when I attempted to mass-produce copper pieces in-house. But let's analyze errors and lessons learned step by step:

MARKETING

My initial idea was quite simple: create product prototype, do crowd funding campaign, let 3rd party company produce some stuff, order parts, make something in-house and -at last- assemble the units + deliver them to customers. There was one small issue: MOQ for copper parts was 200 pieces and each of them was priced at ~28$. So, I targeted to have at least 100 pre-orders in order to take the risk and dump almost 6k into copper parts.

How do you achieve that? At that time (2014) I could not launch Kickstarter campaign since back then Swiss companies where not able to do so, Indiegogo on another hand sounded a bit too shallow and small. Plus there was a PayPal requirement and as someone who had experience with PayPal as a seller I truly *hate* the bastards with a passion and try to avoid giving them any business as much as I can. So I decided “hey, I have my own payment gateway, my servers to run sites on, let’s try my own pre-order system”.

Well, guess what happened…. In 2 years I got total of 23 pre-orders. Could that be improved? HELL YES! My site got most hits and pre-orders not when I published the product on Thingiverse but when it was (by some miracle) featured on different blogs and news sites. By the time this happened though I was already busy with some personal issues that blocked me from doing anything about splicer.

Lesson 1

Bloody obvious marketing, marketing, marketing…. Spread the word, contact all news sites and blogs you can find, tell them you have something interesting. Most of them actually look for things to post and they would take your idea and make entry with it.

Lesson 2

If you can: Go with large crowd funding site… apply lesson 1 on top.

SUPPLY CHAIN

Even if I would have followed the both lessons and succeeded in marketing, I think that my project would fail (or cause me great grief). Reason? I did not have proper supply chain in place…

You see, when you make a prototype things look very simple, you order some parts from AliExpress or some other Chinese site, you measure them, make your prototypes to fit together and think that this is the way it will work when you will produce things en-masse. *Good luck* to you. What will really happen is that next time you will order Ceramic Heaters, PID Controllers etc. -even from the same seller- you will get things that are slightly different:

You will get Ceramic Heaters that are not 4.9mm in diameter but +/- .1mm, so they do not fit into or too lose in the the holes you drilled in copper parts.
You will get deformed probes on PID controllers.
Those shiny 12V power plugs for the back of your device will have shaft of 8mm but thread will end at 6mm instead of 8mm as it was in previous batch, so you will not be able to fix them in 6mm plywood.
And… and… and…

Chinese suppliers (As well as some western ones) often lack consistency in their batches. I knew that back from the day when we run textile business but I thought it would be a bit better with electronics / hardware components. Fool…. Risk of bad consistency is a bit less within a single order but even there it can strike you.

Lesson 3

Find consistent and reliable partners for all components BEFORE you launch your Crowd Funding campaign. Yes, reliable partner will cost you (quite a bit) more but trust me, it is worth it! If component is cheap enough you can also take the risk and buy large ‘AliExpress’ batch in advance, but you risk ending up with lots of useless things if campaign fails.

IN HOUSE PRODUCTION

Oh boy…

When I initially made splicer, only facilities I had in house was a Laser Cutter and wood / aluminum router. I always wanted to have a real metalworking workshop so it came that in recent 2 years I have heavily upgraded, to real albeit light milling machine (EMCO FB3), lathe (Sieg SC10), grinder, saws, stock, tooling, measurement, sanding equipment etc. It is amazing how much money you can sink into that and even then, you miss proper tools all the time. I put about 15k $ into it all.

New Workshop

Luckily only some parts I had to order specifically for splicer (1.75mm convex cutter that set me back almost 300$).

As a total novice in metalworking I read books, watched with fascination videos of Abom79, MrPete222, Keith Rucker and Metal Tips and Tricks, trying to learn. I went through stock material and tooling like through butter (Peng, there goes your 35$ spiral tap), but I learned.
I produced several splicer copper components in house and it worked wonderfully, the final result was much better then prototype, I was very proud to hold those quite complex pieces in my hands and see how well they fit together. But there was one problem: It took me almost 3 hours for each splicer, so I had to optimize, not cut one piece at a time but do 5-6 of them, so I can spare myself entire re-alignment and tool changes.

THIS is where real problems start, you begin to realize how complex it is to align 6 copper blocks in perfect line, you make mistakes by mis-aligning work pieces or tooling and one mistake ruins not one but 6 pieces. You realize that your used milling machine has a problem in that Z-axis shifts by up to .1mm as you move table on X axis. You realize that your Chinese vice does not hold multiple pieces properly. And and and. But I went through it all, fought, learned and at last when I was left with enough stock material for 17 units I thought that I am ready. xoxoxo

I wanted this to be perfect, so before I started the “production” I brought my 3mm convex cutter to grinder company to ensure it will be razor sharp! I am sure experienced metal workers will see where I am going to here….. You see, when you re-grind your tools their diameter often changes… so when I got my cutter back it was not 3.00 but 2.75mm. And I forgot about it, it read R1.5 and I assumed it is R1.5. Fool!

My next mistake was not to control the cuts after each run. This was fatal, so after almost 10 hours of labor I ended up with total of 34 70% done blocks (main heating and cooling blocks), nicely cut but entirely useless since the slots where too bloody small for 3mm filament! And no, you cannot just re-fix blocks in vice and run larger cutter because it is impossible to get proper alignment again, especially with several blocks together. This was when I finally said F.CK IT and went to sleep. Next morning I thought to repeat attempt but then I remembered that I simply had no more stock material left. That sealed the faith of Splicer.

Tired Miller

Box of failures

Lesson 4

Always control and measure after every step of your work, double check your DRO before cutting, do visual control etc. Make control after each step essential part of your ‘routine’.

Lesson 5

Let *professionals* do mass production for you, especially with such unforgiving thing as metalworking. Having your own workshop is nice but keep it for prototyping only, DO NOT attempt to mass produce, save your soul! By the way, I am interested to hear how this (small in-house production) applies to electronics, is populating and re-flowing PCBs in house as hard and fail-prone as metal working?

EPILOGUE

While I failed to sale or produce meaningful numbers of the splicer, one should not forget that it was never intended to be a real, serious product, it was a pilot project to try out the waters and learn from it. As such it has fulfilled it’s role perfectly, I have learned a lot during this time and it will surely help me in future.

Thank you to everyone who liked the idea, I really appreciate your trust and encouragements and if life permits, hope to come back to Maker community with new, exciting and properly setup projects in close future!

Tags:

filament splicer