comment on SteppIR was Re: [TowerTalk] New antenna system

[Jim Lux]

At 11:00 AM 11/9/2004 -0800, Steve Katz wrote:
>Hello Jim,
>
>I'll try to steer away from PCs and back towards antennas to keep the thread
>in line with the reflector intent.
>
>The Low Voltage Directive compliance testing for consumer electronic
>products doesn't cost much; depends who you use as NRTL, how busy they are,
>how much prep work you, as the equipment manufacturer, have already done and
>other variables.  But I have it done all the time, usually for about $3K.
>If FM sells 200 SteppIRs a year for an average of  $700 each, compliance
>cost is 2.1% of sales.  Kind of high, but not atypical for a small company.
>If they (or anybody) amortize an annual compliance budget of $10K over
>products sold, the customers will likely pay it all without knowing they're
>paying it.  Part of business.
>
>I agree with you about good engineering practice.  Design for compliance,
>whether you get certified or not, and you're usually better off.  Frankly,
>the "cost of designing for compliance" is usually "nothing."  It's a matter
>of knowing what to do.
>
>I pointed out the product safety stuff only to re-focus what people should
>have been annoyed about, instead of simply accepting: If a customer shorts
>the user interface cable between two parts of an appliance that are normally
>tethered in service -- whether out of stupidity, negligence or accident --
>it shouldn't require that parts be replaced.  It's nice to have great
>Customer Service and replace stuff that breaks, but it's cheaper and better
>to make it simply not break...especially when the cost of doing so is
>approximately nothing.
>
>WB2WIK/6

I fully agree.  I think the overall idea is that people wind up "making a 
product" without realizing that they are "making a product" with all the 
overhead that this should entail. A few kilobucks is a very reasonable 
amount, although I suspect that it would be a big single shot of NRE for a 
lot of products.  I think a lot of ham products from small mfrs have the 
majority of their NRE covered by sweat equity (board layout, mechanical 
design, etc.).  If you look at all the products, for instance, advertised 
in the back of QST, a lot of them have low capital dollars invested in the 
first unit. There aren't many high tooling cost or high startup cost 
manufacturing steps like  "pick and place" or injection molding.

If you were a would-be equipment designer selling into the ham market, for 
the vast majority of products, you could start small with a limited cash 
investment.  $100 to get the first batch of PC boards made. A few $100 for 
parts. A few hundred $ to get some sheet metal bent or front panel labels 
made.  Those expenses could be spread out over several months.  That's sort 
of a different matter for NRTL compliance testing at several kilobucks at a 
shot (and I'll bet the first time through, for a novice, would be 
substantially more expensive, either in time or in money).  I used to work 
for a small company that designed and sold highly specialized products as 
an additional line connected with their primary service business in a 
fairly narrow market, and that's exactly how they got started.  Eventually, 
of course, their products became the primary revenue generator, and, as 
their products moved into wider markets, they had to learn the whole 
compliance thing. (I think it was getting the CE mark for sales into Europe 
that was the first hurdle.)

Sure, if you're a business, and you have a decent business model, then you 
could get a loan for the capital you need.  Or you could finance your 
business on your credit card, like entry-level film makers do.  Of course, 
that requires a certain level of commitment and faith in that "first leap 
off the dock into the water". However, I think a lot of ham equipment 
sellers don't think of themselves as being in a business (or, at least, 
they don't start out that way!), and so, sort of wind up as businessmen by 
accident.

As far as designing for reliability and immunity to screwups goes... well, 
that often presumes that you set out to design the widget as a product from 
the get-go. In the "business model" I described above, that never really 
occurred.  Instead, you see more of an "evolving product", where successive 
versions of the widget are sold with improvements to fix problems 
discovered in the previous versions.  In the software world, it's the 
"don't spend so much time designing, get it out the door, and we'll fix it 
in the next release, if we're lucky enough to sell any of them" model.

Now that I think about it, the basic engineering practice you describe is 
what differentiates a "product designer/engineer" from a "circuit 
designer/engineer".