[Milsurplus] Dynaverters

[Bob Camp]

Hi

Given what they had to do it's kind of amazing that the transistor 
power supplies worked as well as they did. Charging the high voltage 
capacitors pulled a pretty good surge. They generally had a simple 
feedback arrangement to keep them oscillating. This made the reliable 
(low parts count) but prone to odd failure modes.

Replacing the transistors (Ge to Si or Si to Ge) was a fairly common 
thing to try and more or less here's what usually happened:

1) The bias was off and the transistors were full on (germanium parts 
in a silicon circuit). The oscillator didn't kick in and you simply had 
a lot of DC through one of  the transistors and one side of the 
transformer.

2) The bias was off and the transistors were full off (Si in Ge). Not 
exactly a disaster. This assumes fixed bias, which often was not the 
case.

So you hear the stories and re-bias the gizmo

3) Germanium wants low VBE but high current (low beta). Silicon wants 
high VBE and lower current. Silicon feedback windings fry the germanium 
bases on turn off. Silicon parts do not always oscillate properly in 
the germanium circuit (not enough voltage).

In go a couple of protection diodes on the bases and maybe fiddle a 
capacitor.

4) The silicon parts in the germanium circuit run "faster" than the 
germanium's did. Often it would not just be faster and 50/50 duty 
cycle, but some odd "burp mode" of oscillation. The transformer does 
not like this very much and smokes out. Not an easy thing to diagnose 
since it's often load dependent.

Even with the right parts in the supplies these gizmos were not super 
reliable. Anything that stopped the self oscillation was likely to 
create issues. Cranking the vehicle's engine was a common way to kill 
one. If the starter motor kickback spikes didn't do it the low battery 
voltage probably would. They *should* keep oscillating down to roughly 
half supply with the right transistors in the circuit. With strange 
parts, who knows.

So you re-rig the feedback caps and damp out the winding with a few 
resistors and maybe a choke or two.

5) At low supply (12 volts) the germanium parts where generally more 
efficient than the silicon parts (at least the early ones). The heat 
sink may or may not be big enough to get the heat out of them. It's a 
race since the silicon parts are rated to higher temperature.

Of course if you put modern design silicon transistors in either of the 
old designs then you are going to much faster, much higher gain than 
the original designs. Strange oscillation waveforms .....

	Take Care

		Bob Camp
		KB8TQ



On Mar 26, 2005, at 5:17 PM, Hue Miller wrote:

>>> Depending on the era you will either have nice fat germanium power
>>> transistors or later on silicon parts.
>
>>> The transformer feedback
>>> oscillator was pretty similar in both cases, but replacing the
>>> germanium's with silicon parts generally resulted in disaster.
>>>                 Bob Camp
>
> Why disaster? I could see you might have to change a bias resistor or 
> 2,
> but???  -Hue Miller
> ______________________________________________________________
> Milsurplus mailing list
> Home: http://mailman.qth.net/mailman/listinfo/milsurplus
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:Milsurplus at mailman.qth.net
>