[Amps] filament supply options discussion, follow up

[R.Measures]

On Apr 24, 2004, at 12:01 PM, Steve Jackson wrote:

> Lots of good feedback so far.  My responses, in no
> particular order ...
>
> 1.  Measuring grid current:  thought of that, and felt
> that I could use some kind of low-drop shunt to
> directly measure at the bottle ... would have great RF
> characteristics and also super accurate DC specs, not
> hard to fabricate, either
>
> 2.  Peril of low fil resistance at cold power-on:
> actually, I leave some juice running on the fils at
> all times.  The fils are run at a very low current
> when "off" which allows the minimum load necessary to
> keep the power supply happy, and never allows the fils
> to go to totally "ice" state.  There are two relays:
> one is an SPST that shorts out a series resistor when
> powering up the amp.  Besides, the power supply
> current-limiter is easily adjusted to flatten out the
> peak, anyway.  Much easier to do this with a DC supply
> having all the bells and whistles than to have to
> design either a ramp-start xfmr primary gizmo or a
> nicer (like the L7 fil xfmr) secondary, which is
> perfect in that it naturally limits inrush; if it was
> only available anymore.  The second relay is a big
> DPDT 100A teclco-surplus latching contactor that
> allows me to reverse the polarity of the fil voltage
> every 100 hours (I'll have a Hobbs meter on the PS
> deck).
>
> 3.  Very easy to adjust the switcher to make anywhere
> from 4.5V to 6.5 v with its trimpot or remote control,
> plus it has a remote sense line that can accomodate
> the small I2R loss of the 8 AWG and Anderson PowerPole
> jumper I am planning.  Actual fil voltage will be
> metered in my amp, with a dedicated meter.  I'd have
> to buy a $$ Variac to try the same stunt with a
> conventional design.

=F0  A rheostat is more suitable for adjusting filament voltage since a=20=

10% adjustment range is enough and a Variac provides a 100% adjustment.
>
> 4.  The 100W supply will loaf at the 75W fil
> requirement of a 3-500z, and if I upgrade to the
> 3CX1500D7, I can simply put two of these in parallel
> as they are designed to do that (diode-sum the
> outputs).  $2 for a 100W fil power supply makes the
> whole thing seem cheaper.
>
> I am glad I haven't heard any showstoppers yet (not
> even from Rich!).  And, the savings here are adding
> up, so far, well into a three-figure number, not
> counting the hernia I am not going to have.
>
> Now, if I only had a real metal shop setup to build
> the cabinets ...

=F0  There is no need for a sheet metal shop if one uses the =
construction=20
technique used in the Ugly Amplifiers on my Web site.
http://www.somis.org/Building_ugly_amplifiers.html
PEM =A8 self clinching nuts and machine screws are used to fasten=20
precision-cut 0.125" 6061-T6 Al plates together to form a box.  One=20
trip to a metal dealer to have the plates accurately cut,  -  such as=20
Industrial Metal Supplies in L.A. - and the rest can be done in one's=20
garage with a vise and basic hand tools.  (The last time I did this,=20
the cutting charge was under $2 per slice)  Another advantage of this=20
construction method is that - in the Extremely Unlikely event of a=20
meltdown - there's quick access to any area.   Yet another advantage is=20=

that the resulting box is the right size, so there's no wasted space. =20=

However, for those who build RF amplifiers to impress un-RF-savvy=20
visitors, this construction technique obviously won't do.

cheerz, Steve

Richard L. Measures, AG6K, 805.386.3734.  www.somis.org