Amps
[Top] [All Lists]

Re: [Amps] designing and testing RF Bipolar Power bias circuit

To: dansawyer@earthlink.net
Subject: Re: [Amps] designing and testing RF Bipolar Power bias circuit
From: Manfred Mornhinweg <mmornhin@gmx.net>
Date: Mon, 01 Jun 2009 02:26:09 +0000
List-post: <amps@contesting.com">mailto:amps@contesting.com>
Hi Dan!

 > The plan is to proceed with a
 > circuit based on individual diodes epoxied to the transistor caps.

In that case it's good to use pretty large diodes, so they can take a 
large current while their voltage drop is still relatively low. I 
typically use 3 ampere rectifier diodes for biasing 30 watt amplifier 
stages, which need about a half ampere bias current at most. Trying to 
use 1 ampere diodes for that, the voltage drop always end up way too high.
You might extrapolate from this experience, according to the power level 
you are working at.

In some cases I have used a very low resistor (half ohm or less), to 
form a voltage divider with the other resistor connected between base 
and emitter, which usually is 10 ohms or so. This too helps in avoiding 
too high an idling current.

The problem is that when the amp is working hard, the diode junctions 
are always much cooler than the transistor junctions, and this tends to 
cause thermal runaway. That's why you usually need a biasing voltage 
rather on the lower side of the typical diode drop range.

The best approach is to sense the temperature with a diode, and then put 
this in a loop within a voltage regulator, to overcompensate for the 
temperature change the diode sees. This produces more stable idle 
current over the temperature range. Almost all comercially made amps use 
this method. They typically use an LM723 regulator, or some more modern 
implementation.

Manfred.

========================
Visit my hobby homepage!
http://ludens.cl
========================
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps

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