Date: Mon, 16 Apr 2012 11:45:15 -0400
From: "Carl" <km1h@jeremy.mv.com>
Subject: Re: [Amps] Filter Capacitors
carl sez...
Thats because PSUD completely ignores the basics.
%%% No it doesn’t. The damn thing is dead on accurate every time. Have you
even tried
using the latest version of it. It hasn’t let me down once.
> ## By playing around for hrs on end with psud, you can minimize the
> effect some what.
More stupidity.
If you want to use a choke input on a real PS without ringing then you have
to bleed it heavy and use a sane value of filter, 8-12uF is sufficient for
acceptable ripple. Most of the old ham TX with a choke used too much L and
not enough R in order to save weight on the power xfmr for desktop rigs. On
a class C rig the modulator in class AB or B was a steady load in addition
to a hefty bleeder current. Key the Class C amp PS supply relay and there
was no spike. Look at an old Collins KW-1, 30K, etc, they were clean. A
Johnson Desk KW also.
%%% Who’s being stupid. On something like a DX-100, and similar
TX, the push-pull modulator shared the SAME B+ supply as the RF deck.
On AM phone, as soon as you key the TX, you are sucking a TON of plate
current just for the RF portion. That’s in addition to the bleeder current.
Then when you talk, you have yet another load on the SAME B+ supply..
from the modulator tubes. On CW, it’s a different story. Modulator is
not in use, draws zero current. All you are left with is the bleeder load.
Those TX all ran Class C on CW.... which implies they draw ZERO idle current.
Now start keying away on CW. Now you are slamming a huge on-off-on-off load
on the B+ supply.
> ## In these 1960?s and 1970?s supplies, with the typ 8 x 200 uf caps in
> series, like what was
> used in a SB-220 or Drake L4B, ripple is typ 3%. Use a string of 600
> uf caps..and ripple is
> down to 1%.
Who cares, nobody can hear it.
%%%%% The primary function of the high C filter is to improve
dynamic regulation. The resulting low ripple is just a side effect...and
amounts to a free lunch. 25 uf in an SB-220 or AL-1500 is fubar...esp
when caps are so cheap these days.
> Ripple is just inverse to C used. Rich measures still sells his 560uf @
> 450 vdc
> caps. A few friends have used em in several projects. They work good
> too.
But not as replacements in PS with marginal transformers. It might be OK for
casual no coders but certainly for serious SSB, CW and digital....or AM.
%%% 560 uf caps in any SB-220 isn’t gonna blow up the plate xfmr any time
soon.
>
> ## These days, large value caps are everywhere. And no, they wont cook
> your plate xfmr either.
> I wired 8 x 2500uf caps in series for one of my L4B supplies, and it
> works great. That?s
> 312 uf in total. Ripple dropped from 3% down to just .24% Dynamic
> regulation is superb, it doesn?t
> budge.
And the transformer is cooking along with it. You now are pulling the
difference between what the original did with the sag and now with no sag,
there is no free lunch as the transformer is now working harder.
Put an IR gun on it.
%%% I did put the IR gun on the L4B plate xfmr. Heat rise is miniscule
at best..and that’s with 8 x 2500uf caps in there, wired outboard in
their own box. A string of 560 uf caps isnt gonna put a dent on it.
>
> ## BTW Roger, when you short out one of these large value lytics, it does
> NOT cause the cap to explode,
> why would it.
Thats more CB type talk. Shorting out a big cap CAN rupture it, its not a
given but when it does happen you may be in for a surprise. A cap should be
left to discharge thru its bleeder resistors until down to about 10% voltage
and the thrus a 10K o so resistor on the chicken stick. Hook the stick
directly from the HV to ground while working on the PS/amp to prevent a
partial recharge
%%% I always just let the supply bleed down! With a 100k 3w mof
across each lytic, it takes a while to bleed down. Now if ur in a real rush,
and have no patience, then simply wire a 10-50K resistor in series
with a vac relay, and wire that mess directly between the B+ and B-
Hit the switch, and you will suck it down real fast. Plan B is just to 1st
shut off the B+ supply, then hit the PTT, go for idle current, and that WILL
suck it down to zero even faster. I cant use Plan B anymore cuz the B+
supply is interlocked with the RF deck.
%%% Roger was implying that lytics just blow their brains out for no reason
at all, they don’t. Never seen that happen yet, even with real old caps.
The biggest problem with old lytics is they are being operated at way
too high a voltage. The EQ resistors are too low in value, and cooking the
caps.
The EQ resistors have lousy tolerances resulting in individual V drops across
each cap in the string that vary all over the map. Combine that effect with
operating the caps almost at their max V rating, and you have a recipe for a
disaster. Your leakage current on any cap will be way up, when run between
90-100% of its rating. That just compounds the problem..esp when 100 K
eq resistors are being used. Run the caps at 70-80% of their V ratings, and
leakage current falls through the floor.
> ## put a 50 ohm 50-225 watt glitch R in series with a FAST HV fuse....and
> you will never have a problem
> with any high C supply. 3 kv divided by 50 ohms = 60A of fault current.
> Its 60A of fault current
> regardless of whether you have 25 uf or 250uf. All that happens is
> the glitch R will LIMIT the fault
> current to a safe value. Meanwhile the fast HV fuse will INTERRUPT the
> fault current within 2 msec.
>
> ## You naysayers who keep saying High C filters cant or wont work, or
> are a bomb, have yet to try it.
> There are now several HV supplies built, using high C filters. And
> several of the 7-10 kv no load variety
> using 3900-5400 uf caps, 450 vdc type..and enough in series such that they
> are only operating at
> 70-80% of their 450 vdc rating.
And all the necessary peripheal parts needed for safety add to complexity
and cost and the end result is zero added value to the amp unless the xfmr
is overated to start with. A kit step start wont do the job either and when
the relay shorts as they do...especially in Ameritrons.....then the xfmr
goes next.
For those with commercial ham amps its dangerous and stupid.
For HB the builder better have a clue what he is doing....many dont and just
get carried along with the hype.
Carl
KM1H
%%% say what. A simple HV fuse inserted between plate xfmr sec
and FWB or doubler. A 2nd HV fuse inserted in series with the
glitch R completes the protection. 100% effective..and reliable.
Cost is cheap, and no added complexity, unless you think a couple of
fuses is somehow complex. Step start with high C filters is a non issue.
Heck, you can do it external if you like, and manually control it, or use
a timer. Use a real relay, and correct value step start resistor, and there is
no problem. If the plate xfmr has the correct size fast breaker or fuses, again
there is no issue.
%%% A string of 560 uf caps in any table top amp is not going to cause any
problems at all. What Hype are you talkin about. Nobody builds a 3-7 kv B+
supply these days with a total of 10-20uf in em. The dahl xfmrs handle
high C
filters just fine. Their Z is so low, the charging currents every 8.3 msec
wont
create additional heat. The duration of the 120 x pulses per second is lower
with high C filters vs low C filters. Low C filters draw less peak current
every
8.3 msecs, but the duration of each pulse is longer. It comes out a wash.
%% I asked dahl himself about using high C filters. His response was go for
it.
He told me I could use several hundred uf in total, and it wont put a dent on
his
plate xfmrs. He had already done it, using a helluva lot more C than I have
available !
later... Jim VE7RF
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps
|