I've built many BOM/costing tools for one of my clients who is in the pharmaceutical manufacturing industry and it is absolutely the most complicated part of any new product development. The BOM is an all encompassing contract that should cover not only the parts, specs, and design process but also the routings, requisition/sourcing schedule, equipment/change-parts, and internal lead-times. If you have something that is built over three phases, you have to schedule the material, equipment, and labor precisely to avoid delays or over-stocking. On top of this, you have to factor in multiple levels of internal/outsourced testing, warehousing, shipping/distribution, cost of capital, and account for overhead.
Every time I see a hardware Kickstarter campaign, I wonder if the designers are familiar with the typical supply-chain-order process. Putting logos on t-shirts and coffee mugs is very different from contract manufacturing custom hardware, even if you are outsourcing 100% of the manufacturing and distribution process.
Yea, there are more failed / really-late Kickstarters than you see. It's a shame, but manufacturing is still super difficult on some budget w/ no time (and often little experience).
A lot of the ODM's have engineering and consulting shops attached to them who would do all of this for you. There is also a huge market of people experienced in the industry looking to help people who have never previously been involved in manufacturing.
So unsurprisingly the solution in this scenario is to just outsource it, again.
Andrew Huang was one of the engineers behind the open source Chubby. He has considerable experience with outsourced manufacturing in china and I thought this post would be helpful for anyone doing outsourced hardware.
His "Made in China" category: http://www.bunniestudios.com/blog/?cat=7 is also a fantastic read. It is rare to see inside the factories that make all our gadgets so I am thankful he took the time to document so much as he did.
"Andrew Huang was one of the engineers behind the open source Chubby."
chumby, but close enough! Not the first time someone has called them Chubbys, sometimes intentionally... there was lots of snickering about the 'Chubby 8"' product.
Also, I've never heard anyone call bunnie Andrew, even in real life.
Sorry! I always thought the thing was chubby. The name made sense in my head as the mascot octopus is a bit gooey.
I was not sure if I could call him Bunnie since I thought that was a nickname. Is there a chance Bunnie is not pronounced "bunny"? I think there is a chance I may not be very good with reading words.
It is pronounced the same as bunny (as in the killer rabbit from Monty Python & The Holy Grail) and it is technically a nickname, but I've never heard anyone call him anything other than bunnie even in a work setting.
This information is pure gold to someone trying to manufacture their first products. It's the kind of thing that's not in books, companies learned the hard way, and engineers learned from on-the-job experience at said companies.
The post says that outsourcing is the only practical option, but does anyone have a ballpark of the scale where it becomes worth it? It seems like you need to have substantial knowledge to pull it off without wasting a lot of money on errors. I would guess for a run of, say, 10, you're better off sourcing things more expensively in the US/Canada through some boutique manufacturer, and for 1 million you're better off outsourcing, but not sure about in between.
It'd be nice if it could import from a variety of weird formats, and export to those same formats.
It'd be nice if the user could select a bunch of components to include, and then send those to a supplier for quoting.
Imagine a subcontract manufacturer - they might get some ICs from one supplier, the passives from another supplier, etc. Having easy to split off groups of components to send off to get quotes for is good.
Add some minor spreadsheeting to total the money and tinker with the qtys. (Eg, cost for a one off; a five off; a ten off. Or costs for 10,000 over the next six months at monthly drops of 1,500 with a big end drop.)
But be aware that there's a bunch of really unpleasant almost-niche software already out there. It's often a kludgy bolt on to accounting / pay software.
In 2005 I was getting parts lists printed out on a dot matrix printer and sent via fax machine.
We asked for the spreadsheet file. We got a Microsoft Works .WKS file.
People in this industry are not always computer literate. (A manger had a parallel port printer. He was unable to connect it to his computer, despite there being only one connector the cable could fit it, and only one way round that connector could fit. Other managers wanted drawing documents on computers rather than paper so they could add changes. Try explaining why you can't scan a paper drawing and turn it into an auto-CAD file (when you don't own autocad) and why a gif file just isn't the same.)
They've spent a lot of money on their software. Something like Sage Line 100 (or similar.) So, if you can write something to interface with that list of weird accounting software you're pretty good.
BOM generation is generally a module within CAD/EDA software. The BOM generator typically pulls information from the design files (eg. quantities, designators, component type, ...) and also has the ability to bring in information from an external database ( eg. suppliers, last price paid, ...)
The main value in a BOM generator is this ability to automatically collect information from the design files and, typically in-house, component databases, due to its integration. Any standalone web app would risk losing this valuable integration.
Note that the component databases are typically in-house, built up over the years, based on knowledge accumulated over many designs. Some engineers will even carry a personal database from job to job with them.
Rather than a BOM generator, I'd say a valuable product would be in a serious on-line database of electronic components. EDA manufacturers and component manufacturers have half baked efforts, but they all suffer from their being an attempt at customer lock-in rather than being a truly useful resource.
It'd be a huge job collating and maintaining such a database, but it could be quite lucrative, especially if it included a facility to link suppliers with customers. Think imdb, freedb or cddb, but for components instead of movies or songs. The entry for each component might include:
* Manufacturer's data
* Schematic symbols
* PCB footprints
* Parametric information
* A list of suppliers
* MOQ, lead time, price, ... for each supplier
* and so on...
You'd also need to be able to generate library data in various formats to suit each EDA/CAD tool, or provide an API through with EDA/CAD tools could issue queries on the fly.
It would also be a neat trick to integrate with some of the circuit layout website that are appearing, leaving open the possibility of building an EDA "design flow" by exchanging data between using a set of web sites.
Nothing against China or it's people but if we keep building it up while it buys all our debt, this will end very very badly.
Why can we not build up our own ability to manufacture, or Mexico etc. ?
We're making a staggering number of millionaires and billionaires in China.
Plus when you're done giving them all your tedious designs and specs for your parts, a few months later after they ship your items, clones will appear on ebay and similar markets. If you had manufacturing a little closer to home you might have a little more control over that.
Then there are the problems like capacitor death from really low quality parts. Ever wonder why many electronic parts don't last more than a few years anymore? Another issue that might be resolved by manufacturing closer to home and being able to monitor part quality during the entire production process.
Nothing against China or it's people but if we keep building it up while it buys all our debt, this will end very very badly.
It is really not clear to me how this ends badly for us. If China keeps buying our debt and we fail to pay back, who loses? It seems to me the buyer loses.
Of course, China will not keep borrowing stuff to us. If you just look at things like trains, I am pretty sure Chinese will mostly source them from western companies (siemens, bombardier, ansaldo, etc.).
Coming for the UK my worry is that you lose almost all manufacturing capability and shift to financial/services industries. Short term that probably works out really well financially, long term... possibly not so much.
In the UK, I find it a real effort to find anything made in the the UK on the high street (and few things even designed in the UK, ARM processors probably being the exception). Debt is one thing, the loss of manufacturing/and engineering capability I find more worrying.
> Nothing against China or it's people but if we keep building it up while it buys all our debt, this will end very very badly.
Short term profits vs long term vision. You've got it 100% right, this will end very very badly unless a reverse trend sets in. The question is which CEOs will have the balls to forego short-term profits? It appears Apple is moving in this direction but that may just be a PR feint to appeal to a 'made in America' feeling.
Every time I see a hardware Kickstarter campaign, I wonder if the designers are familiar with the typical supply-chain-order process. Putting logos on t-shirts and coffee mugs is very different from contract manufacturing custom hardware, even if you are outsourcing 100% of the manufacturing and distribution process.