Couple of hopefully simple questions

Magnus Danielson e93_mda at drum.it.kth.se
Sun Oct 27 20:42:51 CET 1996


>      I tend to use 24 gauge solid wire on my panels, and 22-24 stranded for 
>      PCB-to-panel and power distribution. The solid wire on the panel can 
>      be bent into neat paths like plumbing and it stays put. I run all my 
>      voltage divider pots in a daisy-chain, and just apply power to the 
>      whole front panel at some easy access point, usually near one edge.
>      
>      28-gauge is a little too thin for my tastes. Also, I try to use 100K 
>      pots for dividers, especially if there's 30 volts across them, to keep 
>      current draw to a minimum and to minimize impedance effects on the 
>      power busses. If you mess with one pot and it affects another separate 
>      setting, you probably have a power distribution impedance problem (not 
>      true zero ohms) or a mis-wire somewhere.
>      
>      10K is as low as I'll go for a divider that's across -15 to +15. That 
>      calculates out to 30/10k = 3 milliamps, and E*I = 90 milliwatts. The 
>      pot element will probably be running warm at this power. So beware of 
>      this effect as well. My motto is to always calculate out your power 
>      dissipation anytime there's voltages over 5 volts involved. Keep 
>      things in the microamps if possible, except for LEDs and stuff.

Also, useing high resistance pots will also lower the problem of haveing the
power daisy chained from a current point of view, however, daisychaining has 
the
drawback that if a power line to one of the pots close to the power feed breaks
the whole line behind will suffer where as if they have individual feeds then
will the impact of that error be smaller. However, it's even better to avoid
most of the faults that would cause the wire to break.

For instance, a wire which is soldered into a place must be kept mechanically
tigth so that no mechanical jiggeling of the wire occur during use and repair.
This is otherwise a great way to have cabels wear out in advance.
A soldering point should never be exposed to excess physical power when 
possible. The unsoldered component and wire should not keep things mechnical
stable. For smaller components (ICs, normal resitors, smaller caps and 
transistors) you can let the soldering hold their weigth but for larger 
components must the physical weigth be carried by propper support. The PCB 
shall
never be allowed to carry any major weigths either without propper support.
A good backplane PCB will have each contact screwed tighly to a mechanically
stable iron plate and the PCB will just carry itself, the standard PC 
motherboard is a good example on how you NOT should to backplane things.
Take a peek into a PDP11 or something to see a better way to do it.

PCBs migth develope microcracks which is quite hard to find from being exposed
to excess physical power.

Some solderings migth also get loose cause of excess physical power.

>      For power busses I usually go way overkill. Many times in a 3' wide 
>      modular my transformer/bridge/cap raw D.C. section could melt a coat 
>      hanger. So with this manly-man attitude I usually run 14-18 ga. 
>      stranded all over the place, using twisted 2/3/4 wire bunches (twisted 
>      with a drill motor) and solder lugs (remember, fuses are our friends).

True. Haveing good margins pays of in the long run. Sometimes you really have
use for it. I have seen many cases where what must have been seen as a smart
reduction (and costsaveing) turned out to be a really ugly problem.
Do not underestimate current, voltage, power etc. Haveing a big margin can
save your equipment during "disasters". Input/output protection can also be
wise...

Magnus




More information about the Synth-diy mailing list