Industrial design: parts count

[fryingpan]

Just as an exercise. I have a design of mine that I'm perfecting before having it built SMD. I am trying my best to minimise the number of different values (as well as parts counts strictly speaking), but when does it really matter? AFAICT, it's nothing worth worrying about if it's a resistor, you might worry about it with semiconductors, you should worry about it with capacitors (especially fancier types) and integrated chips.

[marcelomd]

Minimizing the number of different parts is a scale thing.

Better to buy 1 big reel of 10k resistor than 1 of 10k and 1 of 12k.

One less thing to worry about keeping stock, one less thing to order (and to play the availability/lead time game). Also one less part to put in the wrong place. And, super important: price goes down with the number of parts ordered.

For a few units per year (or ever), being assembled by a big provider, that has thousands of parts, it doesn't matter at all.

[R.G.]

The sum of your questions add up in my head to your not-so-subtly mining the forum for design guides and a manufacturing plan.

It's OK. Many, many, many people have done exactly this in this and other pedal forums. Somewhere in my archives I have copies of the AOL (look it up  ) messages from Mike Fuller trying to be subtle and hush hush over the diode clippers to use in his protos of what became the Fulldrive. It's been a long road, and there are lots of people who have tried it, with varying results. There's a lot of competition out there. Just sayin' ...

Here's a far more useful manufacturing engineering tip:
1 - make a very exact and comprehensive list of every last part, scrap of paper, packaging material, box, etc that will be delivered to the customer; include a price for each item, accurate to 10%, including shipping costs of the item
2 - estimate as accurately as you can, the cost of assembly labor
3 - estimate as accurately as you can, the cost of order processing, packaging for shipment, and shipping
4 - sort this list by price, descending

The items at the top of the sorted list are what to worry about. For the kinds of pedals made by the people who haunt this site, including those that have gone into business to sell pedals, the top few items will include the enclosure, customizing the enclosure with holes, paint, and graphics, etc., ready to bolt the electronics into, and the pots, switches, jacks, etc. That is, everything except the electronics. If you use a very fancy IC in there, it will sort to the top. If you use near-unobtainium parts like sorted germanium devices and the like they will sort to the top unless you have/find a secret stash.

You can, in general ignore the cost of the resistors, caps, transistors, diodes, and so on that the circuit designers sweat bullets over. The total cost of all of these in the sorted list will probably come out to under 10% of the total cost, probably less. You would not be too far off to assume that the populated PCB is free, exclusive of switches and jacks.

As to inventory cost and the cost of fewer part numbers, I tried for years to worry about this, and it never mattered. It does not matter because there are whole companies dedicated to supplying you with any quantity from 1 to thousands of any resistor, capacitor, transistor, etc you can think of, and they will warehouse them, manage stock, and charge a very reasonably low price to do so. Sure, if you can convert all resistors to 1K, 10K and 100K, go for it. It's a fun way to keep you from being bored in the doldrums of a long design. But it won't make any significant difference in the price of goods delivered. The whole supply chain revolution from the 80s and 90s was intended to make getting just the parts you want, at just the right time, at a good price be easy.

[stallik]

Sensible advice RG.

And the assumption that the pcb is almost free exclusive of jacks and switches might lead a designer away from mounting these directly on the board so that most of the inevitable repair returns can be accommodated with a simple board change.

Lowest cost, minimal labour, rapid turnaround.

Having said that, tc electronic, upon the failure of my ditto pedal during warranty found it better to just send me a whole brand new unit - their standard procedure at the time - so perhaps their labour costs were the highest of all?

[fryingpan]

The sum of your questions add up in my head to your not-so-subtly mining the forum for design guides and a manufacturing plan.

It's OK. Many, many, many people have done exactly this in this and other pedal forums. Somewhere in my archives I have copies of the AOL (look it up  ) messages from Mike Fuller trying to be subtle and hush hush over the diode clippers to use in his protos of what became the Fulldrive. It's been a long road, and there are lots of people who have tried it, with varying results. There's a lot of competition out there. Just sayin' ...

Here's a far more useful manufacturing engineering tip:
1 - make a very exact and comprehensive list of every last part, scrap of paper, packaging material, box, etc that will be delivered to the customer; include a price for each item, accurate to 10%, including shipping costs of the item
2 - estimate as accurately as you can, the cost of assembly labor
3 - estimate as accurately as you can, the cost of order processing, packaging for shipment, and shipping
4 - sort this list by price, descending

The items at the top of the sorted list are what to worry about. For the kinds of pedals made by the people who haunt this site, including those that have gone into business to sell pedals, the top few items will include the enclosure, customizing the enclosure with holes, paint, and graphics, etc., ready to bolt the electronics into, and the pots, switches, jacks, etc. That is, everything except the electronics. If you use a very fancy IC in there, it will sort to the top. If you use near-unobtainium parts like sorted germanium devices and the like they will sort to the top unless you have/find a secret stash.

You can, in general ignore the cost of the resistors, caps, transistors, diodes, and so on that the circuit designers sweat bullets over. The total cost of all of these in the sorted list will probably come out to under 10% of the total cost, probably less. You would not be too far off to assume that the populated PCB is free, exclusive of switches and jacks.

As to inventory cost and the cost of fewer part numbers, I tried for years to worry about this, and it never mattered. It does not matter because there are whole companies dedicated to supplying you with any quantity from 1 to thousands of any resistor, capacitor, transistor, etc you can think of, and they will warehouse them, manage stock, and charge a very reasonably low price to do so. Sure, if you can convert all resistors to 1K, 10K and 100K, go for it. It's a fun way to keep you from being bored in the doldrums of a long design. But it won't make any significant difference in the price of goods delivered. The whole supply chain revolution from the 80s and 90s was intended to make getting just the parts you want, at just the right time, at a good price be easy.

Hehehe! I have *no* intention whatsoever to commercialise my creations. Not on a large scale anyway. You see, I actually started studying electronics and telecommunications engineering about a decade ago. I dropped out (I wasn't in the right place, mentally) but I've always been fascinated by the stuff. I don't really have that much of a solid theoretical background (I might have been considered a "mature" student back then too, and my maths chops weren't so great back in high school either, and time had passed... - they have improved, I must say).

The idea is that I might, one day, in the future, go back to studying and start a second career but certainly not as an, erm, "entrepreneur" of sorts, and certainly not in the MI world. It's not that there is too much competition, it's more that all that could be done has been done, mostly (a little tweak here, another little tweak there, come up with some slightly different filtering or tone response and you have a new design). That applies to both analog *and* digital design (you see some new thingies coming out now and then, which mostly follow the general availability and cost of integrated chips at a certain price point...). To be honest, I do this stuff because I find it *fun*.

[marcelomd]

My experience here is from when I worked developing device drivers for networking equipment. Think Cisco switch/routers. These were BIG PCBs, with thousands of components each. Assembled in house. One reel of passives was good for about a dozen units.

Really cool stuff, different context.

[fryingpan]

I've always been fascinated by heavy-duty switches, handling god knows how many gigabits of data each second and routing them correctly, while also intercommunicating with other local switches to decide which of the switches gets to route which packet with minumal delay. All stuff that needs bespoke processors whose throughput dwarfs basically anything else.

[R.G.]

Well, you fooled me on that one. 

[fryingpan]

Oh, no, but I get why you said what you said. I don't even find it wrong to do so (like, yeah, unless you're just blindly copying stuff...) but the kind of products that I think would make sense to offer in the MI world really need a team of designers (not only electronics engineers, which I am not, but also firmware developers, application developers, DSP specialists, production managers etc.) and on top of that you would really need the bean counters, marketing staff, etc. (at *least* a 10+ staff endeavour). Pedal making is the lowest hanging fruit really, which is why I am even just thinking of designing some - it was all borne out of a personal need to design my own "FlyRig", a multiFX pedal complete with a tuner, compressor, drive, fuzz, chorus, preamp with a tone section, a cabsim, a DI... the tuner is the hardest part for me (it makes more sense to buy a ready-made one and just add it into a modular design), and while you might think that there is a plethora of FX out there that would be good candidates for a rehousing, the truth is that they are either expensive (too expensive, probably, for what they are) or don't meet my requirements.