[Amps] crowbar TSPA

[Ian White GM3SEK]

John T. M. Lyles wrote:
>Mark
>
>That crowbar looks to be a good idea. Wolfhard Merz of DESY in
>Germany has built a light triggered series thyrister CB to protect
>klystrons at even higher voltage. His paper can also be found online
>at the 2003 PAC conference or the 2004 EPAC conference.
>
>All of the crowbars I deal with use either Ignitron or spark gap for
>the shunt element. I am in the midst of designing a new one right
>now, and was thinking of the same old mercury ignitron. For several
>kiloamps at 30 kV, its hard to beat, although silicon devices are
>coming closer as both the DESY and ASTEC projects have demonstrated
>to the world.
>
>I suppose you won't need more than one or two of the GTO or
>thyristors in series for your power supply? You can get IGBTs for ~5
>kV rating now, but would need a way to hold them latched until the
>capacitor is drained. Also expect to pay a lot of money, probably
>more than your amplifier costs.
>
>73
>John
>K5PRO
>
>>
>>Message: 3
>>Date: Thu, 20 Jul 2006 17:24:20 +0100
>>From: Mark Hill <g4fph at mjha.co.uk>
>>Subject: [Amps] Crowbar that HV!
>>To: Amps Reflector <amps at contesting.com>
>>Message-ID: <7.0.1.0.2.20060720170234.0380daa8 at mjha.co.uk>
>>Content-Type: text/plain; charset="us-ascii"; format=flowed
>>
>>Folks,
>>
>>Talking of protection.  I'm contemplating building an crowbar for the
>>HV supply in my 3CX3000A7 PA.
>>
>>The PSU filter capacitor is really too large (130 uF) and the amount
>>of energy it holds when charged to 5 kV is frankly scary.  Eimac's
>>AB17 seems unclear about how much of a bang this valve can take (4J;
>>50J; something in-between?).  I'm not sure whether my ear drums would
>>be OK under conditions of flash over either!
>>
>>I found an interesting article here:
>>http://www.astec.ac.uk/rf/rf_preprints.htm (a little way down the
>>page there is a PDF file entitled, 'A Solid State Crowbar for RF Tube
>>Protection').
>>
>>I like the way that the series thyristor combination appears to have
>>been driven, i.e. individual current transformer for each device with
>>a common primary turn through them all.  Has anyone tried this
>>technique?  Seems as if it might be cheaper and potentially more
>>reliable than a bunch of opto couplers and transistors etc.  Is
>>anyone on the line using a crowbar of any description?
>>
>>Regards.
>>

The opto-triggered crowbar by OE5JFL has been copied by several UHF 
amplifier builders with very valuable tubes to protect, and works well. 
Since that article was written, 1200V thyristors have become much less 
expensive, so the circuit is quite economical to build.
http://www.qsl.net/oe5jfl/flashover.htm

The cascaded opto-couplers are not as elegant as a single primary 
winding looped through a chain of  magnetic cores... but in that case 
you'd have to design your own pulse circuit to drive it. The advantage 
of the opto is that the complete working schematic is available.

Another point that applies to all thyristor chains is that, in real 
life, a chain of thyristors will never fire absolutely simultaneously. 
There will always be one that fires slightly ahead of the rest, and that 
will cause a sudden step-up in voltage across all the others. The 
"dV/dt" effect will then help the rest of the chain to fire reliably.



-- 
73 from Ian GM3SEK         'In Practice' columnist for RadCom (RSGB)
http://www.ifwtech.co.uk/g3sek