[Amps] PSU design question

[Ian White GM3SEK]

Bill, W6WRT wrote:
>ORIGINAL MESSAGE:
>
>On Sat, 7 Nov 2009 17:20:04 +0000, Ian White GM3SEK <gm3sek at ifwtech.co.uk>
>wrote:
>
>>
>>This resistor needs to be there, but it's only ever intended as a
>>backup; its value doesn't need to be a low enough to allow the HV supply
>>to deliver normal current.
>
>REPLY:
>
>Sorry, but I have to disagree with this design philosophy. If you are going to
>add another part as a "backup", add another large diode. A 100 ohm resistor of
>modest size could be blown open by a massive arc, thereby leaving the B-
>floating, whereas a diode will almost certainly develop a dead short, far
>greater protection. And big diodes are cheaper than big resistors anyway.
>
>73, Bill W6WRT

I could be more persuaded by your earlier message which took the 
opposite view, that the resistor is not needed at all.

But we are all in danger of focusing on one small part of the circuit 
without looking at the overall design philosophy for arc protection.  I 
use the normal well-established techniques, many of them learned over 
the years from this group.

1.  Surge limiting resistor in the B+ line [Eimac Bulletin 17]. This is 
vital, to set a maximum value for the current that could flow in the 
event of a crowbar short, tube arc etc. The resistor gives truly 
instantaneous current limiting, for the first few milliseconds until the 
HV supply shuts down in step 3.

2. Diode protection for the meters. Full scale on the grid and anode 
meters is just below the conduction threshold of the diodes (negligible 
current flows below 0.5V) so any overload is well clamped and won't 
damage the meter. [I believe this one came from Rich Measures.]

3. A fast anode current sensor which controls a relay in the HV 
transformer primary. [Various sources, going back to at least the 
1930s.]

Of course it's possible to go further than this, but in practice this 
level of protection is good enough for most kinds of tubes. In my own 
test-bed amplifier (GS35b, 3.5kV), a crowbar short inside the anode 
compartment shuts down the HV supply with a quiet click, and most times 
it doesn't even blow a mains fuse. Press a reset button and the amp is 
back on-air.

Seen in that wider context, an additional shunt resistor on the B-minus 
line is a very small detail. Most integrated desktop amplifiers don't 
have it and don't need it, because the B-minus line is permanently 
connected to the protection diodes.

With a separate HV PSU, my first line of protection is a multi-pole 
connector that prevents the mains relay from being energized unless the 
PSU also has its ground connection and B-minus return.



-- 

73 from Ian GM3SEK
http://www.ifwtech.co.uk/g3sek