[Top] [All Lists]

Re: [Amps] Series PC Power Supplies

Subject: Re: [Amps] Series PC Power Supplies
From: "Will Matney" <>
Date: Thu, 01 Jun 2006 23:31:28 -0400
List-post: <>

I think what he's getting at is this. The power supplies to be seriesed are 
calculated to provide a certain set operating voltage at a certain set current. 
Together these seriesed supplies total the amount needed. The load requires 
this set amount of voltage and current to operate properly. Ok, if one supply 
fails in the string, there is a voltage drop. This voltage drop creates a 
current rise from the remaining supplies feeding the same load. If this current 
rating is over the maximum rating they are designed for, which most likely it 
will be because we sized it to use the number being used, it can make the other 
two fail, IE current trips, etc. Worse yet, if there's not good over-current 
protection, then you start having pass transistors to fail, etc, and have a 
catastrophic failure of the complete system. Most regular power supplies that 
are used generally have just a regulated voltage system, and don't even have 
current limiting. The ones that do are generally the upper e
 nd supplies with adjustable current, or having current sensing which in turn 
controls the pass transistors. If they would use FET's as the pass device a 
person would be better off, but most still use transistors which can go into 
thermal run-away, etc.. Most now only use an IC to compare the output voltage 
and run a driver transistor which in turn drives the pass regulator(s). A few 
may have a current sensing circuit, but again that costs more and a lot of 
manufacturers don't use it. This being the regulated supplies like Astron, 
Pyramid, and several others without adjustable current. About the only safety 
precaution they have is over-voltage pretection by a crowbar circuit. Switching 
supplies on the other hand are different, but can burn out too. I am still 
leary of switching supplies due to all the parts involved just giving one an 
extra excuse to fail. Switching supplies, I think, should be limited to weight 
limited applications. The less parts one uses to do a job correc
 tly, the less chance it has to fail. The simplest type of supplies are the 
transformer-rectifier-pass regulator type. One can build a good regulated 
series pass circuit using two small transistors, a large pass transistor (up to 
10 amps), and a zener diode. The regulator IC's really only take the place of 
the sensing transistor/comparator and the zener diode unless it has a current 
function which a lot don't. However, when seriesing or paralelling several 
supplies, you still have to worry about the lead and lag or each regulator 
circuit to the other in time. This being eaches reaction time to the current 
change from the load. Then you get into trying to build an outside circuit to 
try and get each supply to react equally with the changing current. That is a 
hard thing to do which is discussed in detail in the one book I mentioned, and 
a few others. One would think that small amount of lead and lag wouldn't amount 
to much, but let's remember what happens with surge current in
  capacitors for only part of a 1/2 cycle. It would be almost totally 
impossible to build two or three exact matching power supplies with exact 
matching characteristics and reaction times. The most one can do to try and get 
by is use the diodes, but there are some other ways that are way more 
complicated and expensive to do. Then, your better off to buy one supply to do 
the whole thing anyhow.



*********** REPLY SEPARATOR  ***********

On 6/1/06 at 6:41 PM Jim Forsyth wrote:

>Please explain that more clearly, especially the "opposite sense" part.
>you suggesting that the current somehow flows in the opposite direction 
>through the supply in current limit?
>Perhaps you meant to say the current may charge the output capacitor to a 
>reverse voltage. That is where the diode could be helpful as I mentioned
>my previous email.
>Jim, AF6O
>> All is fine until that load current rises to the current limit of the 
>> 'weakest' supply. When this limit is reached, that power supply will
>> down or limit. The remaiing load current is presented to that supply in 
>> the opposite sense.
>Amps mailing list

Amps mailing list

<Prev in Thread] Current Thread [Next in Thread>