[Amps] Soft start, how to determine if needed.

[Jim Thomson]

In a message dated 2/25/2010 1:03:57 A.M. Central Standard Time,  
alexeban at gmail.com writes:

Yes,  there is.
The most endangered part are the rectifiers. The specification  includes a
current value called " non repetitive peak current", usually  with a time
limit of about a cycle or so. Use that value, assume a dead  short at the
capacitors and calculate a value for the resistor, such that  the peak
current at the peak of the input waveform voltage is less than  this 
current.

### That rating is for 8.3 msecs.. that's  it.  Since it takes several cycles to
charge the caps up, that  single cycle surge rating goes DOWN.  The 6A10
has a 400A, single [NON repetitive] cycle surge rating, that drops down to
just 175A with 10 cycles,   and down to 70A  with 100 cycles. With no step
start, you will eventually fry the on off switch contacts, diodes,
and everything else.  So it is required, either step start, or a slow
ramp up, with a triac. 

### run a typ 3 kv  power supply, with a 50 uf filter cap, through Duncan's
PSUD, power supply modeler, [with soft start option shut OFF]  and run the
simulation. The peak current through the DIODES, on turn on, blows it right
off the chart!  [40A]   On the 240vac side,  that translates to fried  switch
contacts.  [366 A]  




Keep in mind that the transformer action changes the current  so that 1
ampere at 2500V translates to 10 amperes at 250V!

###  And peak pulses  every 8.3 msecs  will be an easy 5+ A
That translates to 100+  A  on the  240 vac  side every 8.3 msecs
which is bad enough.    Like Gerald says below, you will get a massive
Vdrop. 

##  With these big C filter supplies,  this is where you fellow's are losing
it all.  Since the AVERAGE load  is 20A   [using Gerald's example below]
the PEAK load... [every 8.3 msecs]   can be 3-6 x  the average...=60-120A

##  I contend, you have to size the wire from B+  supply, back to main 
200A panel,  to not only handle the 20A  AVERAGE current,  but it now
has to be sized even BIGGER still,  to handle the V drops, from the PEAK
current flow !   Sure, 12 ga  wire will handle 20A, but the V drop across 12 ga
wire with 60-120A flowing, is phenomenal, even with a real short run of 12 ga.

## My 'fix'  for all this was to install a 100A  sub panel [fed with 3 ga] ,on the
other side of the wall of my shack, abt 30'  away from main 200A panel.Then
real short individual  6-10' runs of  10 ga wire, to each of my 3 x L4B's. 

## Bottom line is, when you measure a.. say 4-10 vac  drop at your xfmr
primary, rest assured, the peak V drop is 2-6 x that amount... bad news.


At 240 VAC,  the resistor comes out around 1 ohm or less so that it can be
wired in  series and left there, even without a shorting relay. Use a wire
wound  resistor, such as Dale's and install on a heat sink.

##  try it, it will not work.  1 ohm is not high enough in value  for a 
step start.  But it's too high, to leave in there on a permanent basis. 

Jim   VE7RF



Alex     4Z5KS






Date: Sat, 27 Feb 2010 15:36:41 EST
From: TexasRF at aol.com
Subject: Re: [Amps] soft start, how to determine if needed?
 
Hi Alex, I didn't see anyone reply to your suggestion to add a one ohm  
resistor in the plate transformer primary so I will.
 
This is a really bad idea from voltage regulation viewpoint. It would be  
the same as having a one ohm source impedance from the AC line service.
 
As we have covered several times in recent weeks, the peak current in the  
secondary and primary will be five or more times the average current. This 
is  because all of the needed power is drawn from the line during and near 
the ac  peaks.
 
With a 20A average load, peaks can be 100A or more. The peak voltage drop  
across the resistor could be 100 volts in this case. The output voltage will 
 drop right along with the primary voltage as much as 2000v peak. If the  
transformer is normally 2500vac, peak voltage is 1.414 X 2500 or 3535v. 
losing  2000v of this leaves 1535v peak or 1085v rms. Not good!