Re: [sdiy] SEM: power supplies?

[jhaible at debitel.net]

>> I am supposed to be receiving an SEM front panel/pcb soon and need to
>> purchase a power supply.  From reading about the SEM it requires a +-18.5V
>> linear power supply.
>
>Hi everybody,
> So is this correct, +/-18.5V ? I have schematics that call for the more
>common +/- 15V, but I have seen reference to the mentioned 18.5V supply.
>Anyone with experience powering SEM's ?

The SEM circuit works on +/-15V with one exception (see below). It has an on-board
 +/- 15V regulator for this purpose. So you need a higher voltage to feed that
+/-15V onboard regulator. I guess you need +/-18V for that, and the extra  .5
 would be to ensure that it still works when one regulator is on the upper limit of
its tolerance range and the other one is on its lower limit. (Just a guess. But
it may be worth a try with a carefully adjusted standard +/-18V supply.

Now one point in the circuit needs a higher voltage than +15V: That's the Drain
of the VCO core's FET buffer. This must be between
 (15V + the_FETs_threshold_voltage) *minimum*, and it should not bee too high
in order not to damage the FET either. Another point is the heat dissipation
inside the SEM module. I can imagine that you could even run a SEM on an
unregulated supply, but as you can see from the above, it might be dangerous.
But I think a crude regulated supply built from cheap 7818 and 7918 chips
would be fine, as the *real* regulation job is done by the onboard regulator.
But these are just some general thoughts. Don't do anything except you know
exactly what you're doing. SEMs are quite valuable these days, so I wouldn't want
to take any risks.

JH.