[Amps] [ham_amplifiers] The power of an oversized electrolytic capacitor / pragmatism rules

[Leigh Turner]

Hi Jim,

I agree with most of your comments here; just to clarify I was referring to
the context of Peter's older boat anchor amp equipment and the constraints
in power supply design and component access prevailing back in the tube amp
era heyday prime. In the intervening years large sized oil-filled caps went
out of fashion and HV electrolytic capacitors evolved and got physically
smaller, electrically larger in uF, much lower ESR, with modern electro'
caps becoming much cheaper, more reliable with far better MTBF. Of course
the key to electro' longevity is to always ensure they remain cool and away
from proximate heat sources like EQ resistors, exhaust air, etc.

That's why today single-phase, a hefty HV transformer with lots iron and
copper, FWB, and single filter capacitor banks are the norm in ham radio QRO
amplifiers up to several kW. Above that power level then 3-phase supply feed
and rectification has its advantages.

Where your and my opinions differ Jim is only the number of uF needed to do
the job satisfactorily in an SSB linear :-)

I would argue that circa 60 uF or so of reservoir capacitance is more than
enough to achieve good B+ load regulation and low ripple. 

I certainly agree that increasing C on a puny voltage doubler is beneficial;
but I would only ever advocate the use of a FWB configuration. The common
practice of ham radio equipment manufacturers skimping on xfmr iron and
copper wire is a false economy. Some folk may argue that it’s an acceptable
commercial trade-off in moderate duty cycle SSB amp service. Drop a decent
Dahl xfmr into an SB-220 style desktop amp and watch the CW key-down and PEP
output power capability rise.

Thanks for the kind offer of a bunch of 2500 uF 450V caps Jim; undertaking
that experiment that won't be necessary as I don't for one moment doubt what
you say about the efficacy of using a big reservoir; I just don't accept the
benefit case if the amplifier already has a well designed HV power supply.
The more C that you advocate is merely masking deficiencies elsewhere :-)

Now when I refurbished my old Kenwood TS530S boat anchor Txcr rig years ago
I beefed up the size of the two original electro' caps in the HV doubler, as
well as augment them with a separate parallel string across the B+ equalling
about 150 uF. The result was the achievable PEP voice peak output went up 15
percent and the measured already good IM3 and IM5 on a simple two-tone test
dropped down a couple of dB as a result of the improved HV regulation. This
is perhaps a nice result on paper; but nobody at the Rx end of a QSO would
be able to tell any difference whatsoever. Kenwood's pragmatism ruled.

That is the main thrust of my argument for "enough, but no more" C Jim. An
extra hundred watts or so Pout in a multi-kilowatt output amplifier are
neither here nor there.

Cheers,

Leigh
VK5KLT

________________________________________
From: ham_amplifiers at yahoogroups.com [mailto:ham_amplifiers at yahoogroups.com]
On Behalf Of jim
Sent: Monday, 21 May 2012 11:04 PM
To: ham_amplifiers at yahoogroups.com
Subject: [ham_amplifiers] Re: [Amps] The power of an oversized electrolytic
capacitor / pragmatism rules


--- In ham_amplifiers at yahoogroups.com, "Leigh Turner" <invertech at ...> wrote:
> 
> Yes Peter, you won't go astray if you take your clues from what the 
> leading commercial and professional vendors of the tube amplifier era did
> in their power supply designs; all based on pragmatic time proven sound 
> engineering principles.

### leigh, those bozos are still living in the stone age. Wake up, its
2012... not 1953. Just take a look at the Henry 8k amp power supply. They
use a 6000 vac sec + resonant choke + a pair of 16 uf oil caps. No load B+
is 6000 x 0.9 = 5400 vdc. The 5400 vdc sags down to 4800 vdc with just a
1250ma load on it! The correct size resonant choke from Dahl just happens to
be the same size as the plate xfmr.. and in fact use the same identical
core, and weighs damn near the same as the plate xfmr itself! Now instead of
X weight, we now have 2X for weight. It drops 600 vdc under a steady state
load. Now that's what I call fubar engineering. Its fubar cuz you need a
6000 vac sec..with its real small gauge sec winding. You also require a 300
W CCS bleeder load as well, just more fubar. 

## Ok, here is how I do it..for the same 5400 vdc NO load B+ supply. I use a
3818 vac sec xfmr, plus a FWB, and a string of 16 x 2500uf @ 450 vdc caps.
Each cap weighs just 14 oz. That's a total weight of 224 oz for the caps = 7
lbs. That's also a total of 156 uf. Ripple is zero. Both static and dynamic
regulation are superb, and second to none. A single 100k @ 3watt mof across
each cap. Total diss = 18 watts for the 16 x eq resistors... and NOT 300W.
In both cases I just quoted, we are using a single phase, 60 hz plate xfmr.
We are NOT using a 3 phase plate xfmr, with the typ 208-360-480-600 vac
input. 

> 
> There's no significant benefit (perceived or actual) to be gained in going
> much beyond 60 uF or so in a single-phase FWB capacitor reservoir filter 
> HV power supply used in today's mainstream QRO amplifiers.

### again, you have no clue what ur talking about, simply cuz you have never
done it. On these rinky dink HV supplies, that use a doubler, using large
amounts of C is a great benefit. On any doubler, the caps get topped up only
60 times per second, if using 60 hz AC commercial single phase power. In VK
land, the same caps will only get topped up 50 times per second. On any FWB
circuit, the caps get charged up 120 times per second on 60 hz power.... and
100 times per second on 50 hz power. 

## With any doubler, since the caps only get topped up half as much as a FWB
circuit... the caps in the doubler now have to hold their charge TWICE AS
LONG! To pull that off, you require DOUBLE the C right off the bat. 

> 
> Homebrewers with a parts junk box full of big electrolytics may wish to 
> err on the large size with a 120 uF or so; but doing so will take you way
> past the diminishing return point.

### say what. BRAND NEW large lytics are cheap. It gets better. As soon as
you buy 25 of em, the price drops like a rock. Ameritron, et all, could
easily take advantage of this, I do. 
BTW, it's all based on energy storage, not XXX uf. The caps in a 6kw amp
need to store quadruple the energy of the caps in a 1.5 kw amp, its that
simple...to get the same ripple under load. 1% or way less ripple is so
simple to achieve. Energy in joules = ... voltage squared x uf.... then
divide that result by 2 million. 1 joule = 1 watt second. 

> Having a stiff / non-sagging AC mains feed and a "big" weighty core + 
> copper HV transformer are more important factors in achieving good load 
> regulation in dynamic SSB linear service.

## Old news Leigh. Everybody in W6 land figured that one out back in 1968.
200A service panel + big ga cu wire, a short run, if possible, and a BIG
hypersil xfmr, like a 1 cubic foot 10 kva CCS pole pig, and you are almost
there. 

## BTW, I just got 9 more emails from folks who stuffed big lytics into
existing HV supplys..and they all said the results were nothing short of
spectacular. That included everything from 811 amps to the typ 1.5 kw amps.

## Exactly how many B+ supplies have you built. I have done 23 so far. The
results I get are repeatable. If you don't believe me, I'll gladly send you
a free caseload of 2300uf lytics, + mating 100k equ resistors, and you can
do the experiment yourself. 

later.... Jim VE7RF 

> 
> Leigh
> VK5KLT
> 
> 
> -----Original Message-----
> From: amps-bounces at ... [mailto:amps-bounces at ...] On
> Behalf Of Peter Voelpel
> Sent: Sunday, 20 May 2012 1:01 AM
> To: amps at ...
> Subject: Re: [Amps] The power of an oversized electrolytic capacitor /
> 
> I certainly did.
> 
> I looked up what the commercials did.
> 
> No large glitch resistor, just current sensor and fast cut off for the
> transformers.
> 
> All have choke filter inputs, so have mine. The largest filter capacitor
> is sitting in my 4CX3000A amp running single phase it is 12µF.
> 
> Those within my 3-phase P/S are between 2-9µF. 
> 
> 73
> Peter; DJ7WW
> 
> 
> -----Original Message-----
> From: amps-bounces at ... [mailto:amps-bounces at ...] On
> Behalf Of Jim Thomson
> 
> For you folks that go on and on and on abt diminishing returns, why don't
> you toss the arrl book in the trash..and find out for yourselves. 
> 
> 
> _______________________________________________
> Amps mailing list
> Amps at ...
> http://lists.contesting.com/mailman/listinfo/amps


>