[sdiy] Resistor types

[harrybissell]

Well I'm gonna jump in here so maybe I can get in on the Macy's action :^P

Peter Grenader wrote:

> Shoot me department:
>
> I've found that the global use of 1%'ers, without probable cause due to the
> absolute need based on the specific circuit to be overkill.  OK, shoot me
> and cock your weapon again:

Bang !    I have a whole cabinet of 1/4W carbon films (and carbon comps) which
has
been largely unused since the mid 80's when metal film became viable.  With
little
exception today, there is no reason to specify a carbon comp except for

1) pulse power rating - carbon comp can really, REALLY suck up an overload.
Like if you have
a 1/2 watt power dissipation that consists of a low duty cycle 5 amp pulse !!!
Carbon COMP
can do it - metal and carbon film cannot.  That is the reason that bulk metal
and ceramic resistors
have recently been developed.

2) RF.  Low inductance (nuff said).  I had one circuit that failed with a metal
film (not a synth btw...)
because metal film (with in this case, a spiral wound element) got hot from RF
induction heating
and blew out. Carbon Comp would survive this easily.

other side:

Why NOT use a metal film at 200 / $4 ...   it can't be a cost issue.
You can always use a 1% in place of a 5%
The ratio match at 1% (even if you sub an odd value - like 110K for 100K) is so
good
that diff amps, summing networks, phase shift stages, filters etc... work so
much better
right from the git-go.

> While in audio applications I've heard issue that they keep noise down, I
> still think it's up to the individual circuit, with a careful analysis
> required to determine if it's beneficial before it's globally incorporated
> for 'save sake".   Lock and load again:

Ready... Aim...   ;^P

> I find it absolutely overkill in processing controllers where the main
> function of the module is variable, like processors and sequencers and the
> like.  Point that barrel again and say 'cheese':

hey... 50ppm temp stability ain't cheese.  Sure, you are going to adjust the
pot anyway...
but MOST of the resistance could be very stable, and only a small part variable
(a GOOD
pot is often 150ppm).

unless you are one of those lazy engineers who puts the pot from rail to
rail... and figures
the right value is in there somewhere  (he blushes and looks at the floor :^)

> The downsides are obvious (to me):  coming back two days later, or however
> long it takes for you to forget what in the heck was where on your
> vectorboard and trying to figure it out by reading 1% color codes or worse
> yet, rotating part bodies to attempt to read hidden alphanumeric values  is
> obnoxious.

OK.  Then use surface mount :^P
Then you not only can't read the value... you can't easily lift one end. So cut
the
run and patch it instead...

Seriouly.. Peter has a good point. There are two solutions

1) Bright light and a good magnifier. A neat little Bausch and Lomb "10x
Coddington" costs
$10 or so... and you can hold it siz inches from the board (really good depth
of field).
Good light is essential... I use a combo of incandescent and fluorescent to get
goos color
rendition...

...and I squint a lot :^P

2) Hey Peter... y'know those alphanumeric resistors can have the leads bent in
any
direction... when you put it in the "Speedy Lead Bender" (think they are about
$2 still...
the little red ones :^) ... you are allowed to bend it so the numbers SHOW when
installed

(i know... Murphy says the one you didn't notice is the one you need to check)

there is another solution (that sometimes does not work)... OHMmeter  If you
reverse the
leads you can often get a reading, sometimes not...  Then I lift one lead out
of the

3) Vector T42-1 terminal... wires go on the bottom, component on the top. Grab
with a mini
hook and unsolder - POP and you have one end free

> No use wasting more bullets on me, I'm pretty much swiss cheese at this
> point.

Perhaps we could mop you up with a solder sponge... and wring you out once more
???  :^P

> On the other side of the very side coin:  I  remember once not so long ago,
> someone on AH (i think), in an attempt to preach  the fantasmagorialness of
> a (name withheld, but it's a brand of modular) bragged about it's abundant
> use of trim pots, sighting that has an obvious marker of it's design
> superiority.

Hmmm.  A rev 2 prophet had 120 trim pots. The rev 3 had 60.  I re-did mine and
added
another twenty to calibrate the poly mod section. More thrim pots is bad... so
are too few.
The right number is "All the ones you need, are there"

> x0x0,
>
> Harry :p
>
> kidding...
>
> Magnus
>
> no!..
>
> Peter

no...

H^) harry    (back to you Peter :^)