[time-nuts] Test gear with embedded PCs Re: Basic

[jimlux]

David C. Partridge wrote:
> Jim,
> 
> It might appear on the 2nd user market sooner, but the odds are you
> won't be able to either repair it or calibrate it as the manufacturer
> will have been the only supplier of either of these services, and no
> service manuals will exist.

But is this any different than existing test equipment? I agree that 
there will be some weird widget interface between the embedded PC and 
the hardware, and that might be challenging to reverse engineer and 
duplicate, but overall, I don't know that it's any different than doing 
it for 20 year old gear.

Different processes, but fundamentally the same kind of problem.

What would be a bigger problem is availability of device drivers and 
such, especially if the OS has some sort of inherent life limit built 
into it (e.g. a digital rights management feature like Windows Genuine 
Advantage.. can't connect to the server, and your scope stops working)

For the intended original market, having to connect to a server every 6 
months or year when it's in for cal isn't a big deal.  However, in the 
recycled market, 10 years later,....



> 
> If it is still in support, the mfr will calibrate/fix it for you if
> your pockets are deep enough (probably as much or more than you pay
> for it).  If (as is likely), it is out of support, then it will only
> be good for re-cycling or land-fill :-(

Yes..


> 
> Hmmmm does anyone but us old fogies see anything wrong with a
> business model where stuff can't be fixed and has a support lifetime
> of 5 years or so ?????
> 

I don't know that it's "can't be fixed" any more than any other old test 
equipment.  There's plenty of HP gear out there that has parts that 
cannot be obtained any more, and folks who are motivated find 
substitutes, etc.

It's certainly "uneconomic" to fix, in the sense that for someone who's 
using the equipment in their business, there comes a point where it's 
cheaper to buy/lease new gear rather than fix the old stuff. And, an 
equipment mfr can make a legitimate decision to not design for infinite 
repair life in exchange for lower original sales price.  Yes, this sort 
of shafts the hobby/tinkerer market, but it's the economic world we live in.

And not only the hobby market gets the problem.  At JPL we've got 
bunches of 8663 signal generators that are decades old, and for which 
there's no equivalent modern replacement that has all the "features" of 
the 8663.  (that is, the new E8663 doesn't work anything like the old 
8663 in terms of sweep behavior, phase modulation, or reference input 
handling)

But, because those 8663s were real workhorses, and because we have 
enough hangar queens to scavenge parts from, we kept them going for 
long, long after their intended life span, and never invested in finding 
a suitable new replacement (or, more properly, finding a new replacement 
and working around its idiosyncracies, like we did with the 8663).  Had 
we had a regular "replace every N years" strategy (where N is 5 or 7 or 
??) we wouldn't be lulled into complacency.

(note that a given space mission has a lifetime from "buy equipment" to 
"end of mission" on the order of 7-8 years.. for something like Cassini, 
it takes 7 years just to get to Saturn, after 3-4 years of development 
of the hardware)

It's really a "test setup design philosophy issue".. how much do you 
depend on idiosyncracies? Or can you design for a generic widget.

Even if you're working in the consumption of surplus, if you can "design 
for the generic widget", then you shouldn't care that there's a planned 
obsolescence thing going on.  IN theory, all that obsolescence should 
result in more surplus gear on the market at lower prices.

(assuming that the surplus market evolves...)