[Amps] Quadra step start

[Jim Thomson]

Date: Thu, 27 Apr 2017 14:00:48 +0300
From: Victor Rosenthal 4X6GP <k2vco.vic at gmail.com>
To: amps at contesting.com
Subject: Re: [Amps] Quadra step start

<But if the step-start also controls the filament voltage it should be 
<longer than 100 ms. I have always adjusted mine to be about 1s in order 
<to give the filament time to heat up. I have no idea what the 'correct' 
<delay would be, but the 3-500Zs seem to be glowing after about a second.

73,
Victor, 4X6GP
Rehovot, Israel
Formerly K2VCO

##   Agreed.  1-2 secs for my L4B  works fine, as I step start the B+ and the fils at the 
same time. 

## Ian, points noted on the use of arcol metal finned resistors.  I used the 20 ohm, 250 W variety, which was
the biggest thing they had at the time.  The metal finned  arcols are easy to mount.   But they are used on a 
larger, HB B+ supply.   I used the duncans psud program for years now.  Limitations  with it though, as is it
always has a load on it, so u have to minimize the load to almost nothing to simulate start up, even then, with their
step start box ticked off, results are sometimes misleading. 

## charge up time depends upon amount of total C used.  With huge values of C used, more time is required. 
The problem with the resistor type start up is...the caps never charge up to the full B+ value.  Reason is, there
is always some magnetizing current flowing in the primary of the plate xfmr.  Very little flows with 0-140 vac,
but above that, the magnetizing current rises quickly.  This residual magnetizing current also flows through the 
step start resistor, creating a V drop.  So the pri of the plate xfmr  will never see the full 240 vac..esp with
higher values of resistance used for the step start R.  My small pole pig has 1.9A  of mag current flowing through
it in normal operation.  My big dahl xfmr has 1.5 A of mag current. 

##  So we end up with the B+ only rising 80-90% of full value. When step start R is shunted, then we get a big
secondary surge, as the B+  jumps to its max value.   That sec surge can be minimized by using smaller values 
of step start R...but then the R is dissipating more instantaneous heat.  Bigger wattage resistors can be used, or
more of them used in series, or parallel etc.   But when u get well below 10 ohms, the initial current surge is quite
astronomical.   We 1st noticed the problem of the sec surge, when the AC  HV fuse, located between sec
of plate xfmr..and  input of diode board, would sometimes blow open.   The 20 ohm step start R was too high
a value.  2 were used in parallel, and 10 ohms was tried...and worked, end of issues, no more blown HV fuse. 

##  I came up with another method of implementing step start.    I used a small 0-270 vac at 5A  variac.
Variac is brought up slowly over 1-3 secs, so no initial surge at all, none. Voltage brought up to a bit higher than the
normal 240 vac I have from the mains,  like 250 vac.   So B+ is a bit higher than normal.  Then a  simple  DPST 
relay  removes the variac output  from the primary of the xfmr, while a big contactor simultaneously   applies 
mains 240 vac to the plate xfmr primary.  Then normal 240 on the xfmr, and B+ drifts down a bit from its temp
slightly elevated value.   In normal operation, a 2nd DPST relay is used to apply 240 vac to the input of the variac. 
Both relay coils are 120 vac types, wired in parallel.   No secondary surge, since no resistor used.   

## A small cam on the variac shaft activates a microswitch, when variac is fully CCW...which drives another relay coil,
for a 4PDT relay, which latches to itself.    I call this a fail safe circuit. So IF the commercial AC power went off, then
came back on a while later, HV supply remains off, no B+.  IE: variac has to be cranked fully CCW 1st, so the cam will
activate, and circuitry  applies the various voltages, etc.   Same scheme also insures that IF I forget to turn the variac fully
CCW, when done for the night, that nothing bad happens the following night.   Sounds more complicated than it actually is. 
End result is a real smooth step start..right up to slightly higher than normal B+ value.  I crank it up a bit on the high side initially,
so when variac is removed from the circuit, and normal mains 240 applied, that takes  a miniscule amount of time, during which 
the B+ will start to drop.   Having it on the high side a bit ensures no sec surge at all.   This is the sort of thing you have to look out
for when using big xfmrs and big C filter caps, with low esr on both. 

Jim  VE7RF