Actually kitting is probably much more expensive than the PCB. Been there done that.

Bill posted this in a recent thread. I didn't want to hijack it, but I am very interested in any real world experiences that anyone has had in kitting.

ETA: I mean experiences in designing and distributing kits, not buying and assembling kits.

 

Here are some kit making tips that can be found on the web.

http://www.ladyada.net/library/kits/kitmaking.html

This is the same person that does www.adafruit.com. They do plenty of kits...

 

Before I started to design my own stuff I built a number of Heathkits. I could make some guesses about problems you might encounter.

One is that you need to have all the components stocked in some depth, or be prepared to change the instructions and parts list to reflect any changes.

If you plan on selling to persons with zero experience, then you need to be concerned about trying to teach soldering technique and proper component placement. I was in a Heathkit store in San Diego when a customer came in to find out why his TV did not work. Despite the instructions, he had inserted the components from the foil side of the PCB's. That's hard on transistors.

The soldering thing is pretty vital, too. Having a tool kit available might help, so you can try to keep the kit buyer from using a soldering gun or some old 100 watt iron with acid core solder. You have to address the question of easily used solder (tin-lead with the rosin in the core) vice the hard to use lead free stuff.

The assembly booklet cannot have too many illustrations. Every part needs to be shown at life size. All color codes must be used per each resistor - 47K needs to be also yel-blu-orn every time.

 

Here are some kit making tips that can be found on the web.

http://www.ladyada.net/library/kits/kitmaking.html

This is the same person that does www.adafruit.com. They do plenty of kits...

Wow! Great links to a source I knew nothing about. Thanks.

 

I've been designing and selling my kits for a few years now. I'm now putting together a fairly complex home automatic & HVAC controller. Avoiding non through hole parts is getting difficult so I've had to give in. A 100pin TQFP is easy for some a nightmare for others but I'm working on a simple bolt clampdown to aid in assembly. If it's too much trouble I'll drop it as a kit and sell assembled only. I also plan to layout the PCB with pads for both the SMD & through hole part. I could also consider Schmartboard to do the PCBs as they have a wonderful method for easy surface mount soldering of SMD parts both large and small. It would add about $10 to each PCB though.

 

I have never tried to sell a kit, but I have built a fair number from Knight to Heathkit to RC receivers and transmitters to devices that were for a niche market (electrochemical detector for HPLC). I just built an ESR meter from AnaTek (http://www.anatekcorp.com/). It was an enjoyable experience for a snowy day, but that was all. It was not to save money by any extent. I would have bought the kit even in the absence of a price difference. It does seem to work well.

In the days of point to point wiring, kitting a product might actually have saved the manufacturer money who could pass some of the savings along to the consumer. Today, I think the kit manufacturer is selling fun, not end-user savings. I would guess that a nice kit may cost almost the same as the completed product to produce. If there is a difference, one phone call to customer service or one trouble shooting by tech service probably erases that difference. Fortunately, I did not have to call AnaTek for any help.

This opinion is completely out of school for me, but I suspect that with complete cost accounting, kits should probably sell for more than the finished product to have the same profit margin. If that is even close to being accurate, then to be a successful kit manufacturer, one should focus on why people buy the product, not its utility.

I don't mean to put Bill (blueroomelectronics) on the spot. I built one of his kits. It was well done. I would like Bill's opinion on my guess about profit margin, if he feels comfortable doing that. BTW, the kit I built from Bill (Inchworm) works great and is an artistic masterpiece compared to the AnaTek design.

John

 

I've been designing and selling my kits for a few years now. I'm now putting together a fairly complex home automatic & HVAC controller. Avoiding non through hole parts is getting difficult so I've had to give in. A 100pin TQFP is easy for some a nightmare for others but I'm working on a simple bolt clampdown to aid in assembly. If it's too much trouble I'll drop it as a kit and sell assembled only. I also plan to layout the PCB with pads for both the SMD & through hole part. I could also consider Schmartboard to do the PCBs as they have a wonderful method for easy surface mount soldering of SMD parts both large and small. It would add about $10 to each PCB though.

Some things at that level, you can order the board with the fine pitched/BGA items placed, leaving all the through-hole and larger SMD for later assembly. Saves a little bit of money, and you don't pay their part prices for smaller lots.

 

It is funny, but when I say kitting (although I was actually referring to kits in the original thread) I generally refer to that phase where you gather all the parts and put them in a bag to be assembled later. Even with a well sorted stock it is a very labor intensive process, almost as much hassle as the actual assembly.

Then there is step where I document the dimensions, using a scale (a precision ruler) and a set of good calipers. I like to draw what I'm building before I build it, and it is critical for laying out a PCB. All in all, if I had to do it for a living it would be expensive for the people paying for it. Since it is for me, I am merely kvetching.

I have also helped other people out occasionally by doing some of the kitting phase. When I buy parts I never buy one, especially if they are cheap.

 

Some things at that level, you can order the board with the fine pitched/BGA items placed, leaving all the through-hole and larger SMD for later assembly. Saves a little bit of money, and you don't pay their part prices for smaller lots.

That is an option, and one I've considered. Not sure it would translate into much of a savings over fully assembled. I'm trying to target the device with a plastic enclosure for under $200. A DIY non enclosure version will be even cheaper. Similar but I/O on front and back to the Global Cache device below.

PS when doing kits that need an enclosure picking one is a real chore. I'll either go with PacTec or Hammond.

 

I've been designing and selling my kits for a few years now. I'm now putting together a fairly complex home automatic & HVAC controller. Avoiding non through hole parts is getting difficult so I've had to give in. A 100pin TQFP is easy for some a nightmare for others but I'm working on a simple bolt clampdown to aid in assembly. If it's too much trouble I'll drop it as a kit and sell assembled only. I also plan to layout the PCB with pads for both the SMD & through hole part. I could also consider Schmartboard to do the PCBs as they have a wonderful method for easy surface mount soldering of SMD parts both large and small. It would add about $10 to each PCB though.

Why not do a hybrid design? Sell the board with all SMT components soldered, the user completes the through hole things. Or, have the 100TQFP chip on a little module which can be inserted into a through-hole socket, like the Explorer 16 board from Microchip does.