[AMPS] HV PS topics

[John T. M. Lyles]

I designed HV plate power supplies for the Broadcast Electronics FM-1.5A,
FM-3.5A, FM-5A series transmitters in 1982-3. In the 3500 watt model
(4CX3500A final), a resonant choke input was used, followed with a cap, and
a second section of smooting/filtering for the higher freq components
(240,360, 480, 600, 720 Hz) that get through the resonant section. Both the
resonator choke and the second small choke were designed to operate with
the DC potential, but were installed in the neg lead. We did so, because of
the reasons Ian mentioned. Better to have occasional spikes of HV across
the frame as opposed to having a continual DC potential, attracting dust.
It has worked well for years in hundreds of units in the world.

>I believe so... a choke input supply still needs a glitch resistor, to
>handle the current surge out of the smoothing capacitor. If the choke
>and resistor are in the negative side, the choke would connect to the
>junction of the glitch resistor and the negative end of the capacitor.
>That point will spike towards several kV negative if there is a
>"crowbar" short from B+ to chassis, and if that point arcs to chassis
>the surge current will bypass the glitch resistor.

>Once again, it implies that the choke would need to be insulated to
>broadly the same standards as if it were in the B+ lead. The difference
>would be that in the positive lead the voltage stress to chassis is
>present all the time, while in the negative lead it only happens in rare
>emergencies.
>73 from Ian G3SEK

---------

Series R for current limiting

I would recommend that the wirewounds be avoided in this application if you
can afford to use a Globar style ceramic tube resistor of similar value to
what KM1H suggested. If there is significant current in an arc, I have had
the wirewounds rip apart, buckled the enamel coating. They are solenoidal.
And inductive, so that one has to think about critical damping of the
resonant ringing of the HV supply during a transient event like an arc.
Remember that the L and C are series resonant. The additional L of the
resistor may add a pole to this thing, I don't know if it has been modelled
or tested. Using pure noninductive R such as the old Carborundum AS series,
will work very well. They also can run at elevated temperature if you so
chose.

>While we're delving into glitch resistor design theory, how about some
>practical advice? Before replacing the plate choke, I'm going to install a
>glitch resistor in the LK550 I've been working on. KM1H suggests a 15-ohm
>25W enamel wirewound. I found what appears to be a suitable part from
>Newark, an Ohmite 02F2280 ceramic core wirewound resistor with a flame
>resistant vitreous enamel coating. It's one of those tubular types. There's
>a mounting bracket available that appears in the drawing to be a piece of
>metal that goes through the hollow tube of the resistor and has mounting
>tabs at both ends. It looks like the bracket will hold the resistor up off
>the chassis by at least 1/4", positioning the solder lugs straight up. It
>looks like a nice way to install the resistor, but is there any problem with
>using this type of bracket? Does it hold the resistor rigid so that it won't
>roll around? If it's not suitable, what alternative mounting methods are
>popular for this type of resistor when used for glitch protection?
>
>What's the typical result of a high-current failure in a resistor like this?
>Does it come apart in such a way that the lug soldered to +HV could come
>loose and contact the chassis or short to the chassis through the bracket?
>73, Dick, WC1M

I was concerned for the same reason about the Ohmite sticks when I used
them in bleeder applications. So I did not use the stock metal clips, but
resorted to using a #8 screw through the solder hole, and an insulating
standoff post. Used a solder lug under this screw for wiring. I was also
concerned with arcing from the core of the ceramic tube to the clip to
ground, using them at 5500 VDC on the FM-5A. Problem with the approach was
that the hole dimensions and tolerence are not accurate or adequately
specified for the alignment of the ceramic posts. Also, there is no solder,
so the connection is able to loosen with time and heat. A suitable
lockwasher is required, meaning one has to tighten the screw significantly
and risk breaking the lug loose from the resistor or cracking the ceramic
post to chassis. Had I done it all over today, I think I would stick with
the stock clips, bend them well so that they really grip, and don't worry
about rotation or arcovers.


John
K5PRO





--
FAQ on WWW:               http://www.contesting.com/ampfaq.html
Submissions:              amps@contesting.com
Administrative requests:  amps-REQUEST@contesting.com
Problems:                 owner-amps@contesting.com
Search:                   http://www.contesting.com/km9p/search.htm