Jan said:
Sorry Joerg, my English, but I mean sort of Die Hard (probably not right either),
let's say an extremely difficult way to do something.
I _KNOW_ about large production runs and the cost of calibration.
It all depends on what you do, in the case of Tek scopes (I have worked there)
calibrating a scope (from the book) involved somebody doing a _lot_
of adjustments.
I have done that, and I have done QC for them too, and repaired the stuff.
One could argue that bending a wire to the deflection plates would
require re-calibration I have seen them fine tune that way, mm maybe I did it
myself some time ;-) (100MHz analog 1Gs digital sampling scopes, forgot the number).
You can hardly say Tek made (makes) 'rubber products' or 'rubber designs'.
Their equipment is the reference for a lot of things!
Oh yeah, I've got some Tek stuff here in the lab as well. Good products.
All I wanted to say is that if a great company uses pots that isn't an
endorsement for me. It may still be wrong. But you can make even a
sub-optimal design reliable by chosing expensive parts such as $3
trimpots. That just won't fly in most of my designs.
Yes you can make accurate stuff at lower frequencies with opamps and <1% resistors,
but then there is maybe that little bit offset you want to get rid of.
It all adds up, try a Monte Carlo, all those 1% resistors, offsets, RF rollofs..
So then to get <X% accuracy at some point or other, you need trimmers.
Even more so in radio when tuning circuits, tuning Xtals, balancing multipliers, etc.
Analog is not only LF opamps, you know that.
Most of my stuff is north of 10MHz, some is above 100MHz. No pots. But
that can only be achieved if you think that way already when doing the
initial block diagram. An example was a fast digitizer card the size of
a large pizza that needed numerous ADCs ganged. The original one had
about 20 trim-pots and was a nightmare, even in production. Service
hated it. I replaced it with a design that was fully auto-cal. Number of
pots: Zero. Service loved that one.
Not always possible or even desirable.
Mostly possible, always desirable. At least for my clients ;-)
Yes for 1 million mp3 players you are right, not for the sort of things small
specialised companies maybe sell at the most 100 of, and calibrate in house.
Also not for the big ones who have some standard to go by, and have the environment
and qualified personnel like Tek.
Depends on the standard. Mine is different
This is true, I remember a group going to do preventive maintenance on a couple
of million $ pieces of video equipment, and re-adjusting all trim pots.
However there were very few problems related to trimmers, these things are reliable.
Replace if kaput.
There are bad quality ones too.
For lab equipment, maybe. For medical, mostly not. Latest after a
freighter pilot had to "nail it to the runway" because of side winds or
a lengthy truck ride across an unpaved road the truth comes out. Or
pieces of hardware, sometimes.
You can use DAC as digital poti for signals too, I have done that.
Yep. And for fast stuff those dual-gate FETs plus cheap DACs come in
really handy. But you need to servo those because they are drifty.
Sure, add a blue lamp and you get white light you can read by ;-)
BTW I just heard that per December 11 (as of today) the Dutch analog TV
stations are now completely switched off.
I dunno about the situation in Belgium, but analog is now officially
dead here.
You can now buy a digital settop box but I just looked and it said: 'Sold out'.
http://www.kpn.com/kpn/show/id=1456428
You speak Dutch?
It's been over 20 years now. I still understand it ok but speaking, I
don't know. A few years ago it came back after two pintjes in a pub in
Rotterdam. No idea whether that that would work again. Might need four
glasses this time ;-)
Sometimes I listen to Radio Nederland Wereldomroep, mostly because the
German station Deutsche Welle has managed to blow it in terms of access.
I can still follow the news in Dutch.
