[Amps] Crowbar components

[R.Measures]

On Jul 25, 2004, at 1:31 PM, Ian White, G3SEK wrote:

> Having been away for a few days, please excuse my replying to several 
> points all together.
>
> Rich wrote:
>>>> Frank --  Why go to the trouble of interrupting the flow of electric
>>>> mains current by special means?  The stored charge in the filter-C 
>>>> is
>>>> what can produce enough peak current to cause bent filaments in 
>>>> 3-500Zs
>>>> or gold to sputter from the grid of 8877s, and this current is 
>>>> already
>>>> limited by the glitch resistor -- provided it is able to withstand 
>>>> the
>>>> Joules available from the filter-C without going kaput.
>>>>>
> Depends what you mean by "special means"...
>
> When the peak current through the tube has been limited, by the glitch 
> resistor or by a crowbar, you still have to switch off the mains.
>
> A fuse, a circuit breaker, a mechanical relay, a solid-state relay - 
> any of them will do, and in this context none of them is particularly 
> special.
>
> Frank replied:
>> Well, I see the crowbar as a very elegant solution - and I've never 
>> done one myself so far. So, technically speaking - I can't really say 
>> what the killer argument for a crowbar would be. I see it as a second 
>> protection in case the Glitch-R doesn't do his job or something else 
>> happens out of the blue...
>>
>> And, of course, building one and testing it is in itself a new one 
>> for me.
>>
>
> * That it reduces the surge current through the tube to almost zero! 
> At best, a glitch resistor can only reduce the peak surge to several 
> tens of amps.

In theory, an electronic crowbar reduces the surge current to almost 
zero.  However, electronic crowbars are not instantaneous, so a 
maximum-fault test measurement of peak-I with an oscilloscope and an 
appropriate current shunt is warranted.
>
>
>
>>> I see it as a second
>>> protection in case the Glitch-R doesn't do his job or something else
>>> happens out of the blue...
>
> I'm not getting through here...  the crowbar circuit is instead of the 
> glitch-R, not as well as.
>
>>
>> Hello, Frank --
>> To prove that a crowbar protects the tube adequately, one needs 
>> measure the peak fault current,  To do this, connect a 0.1-ohm 
>> precision resistor from the neg. HV to ground, and connect an 
>> oscilloscope across the resistor to measure the peak V.  When the 
>> pos. HV is shorted to ground with a large screwdriver, for a 2, 
>> 3-500Z amplifier, the peak V should be less than 20V, which = 
>> 200A-pk.   For oxide cathode tubes of similar output capability, the 
>> peak current should probably be under 100A.
>
> Tube manufacturers already told us how to test a crowbar circuit. 
> Eimac in Bulletin 17, and Seimens in their tube data, both say that if 
> you short the HV supply with a piece of very thin wire, the crowbar 
> circuit must shut it down without damaging the wire. For example, you 
> must be able to short a 3-4kV supply with a piece of 3-amp fuse wire, 
> without blowing that fuse!
>
> The important thing about this test is that it limits the maximum 
> ENERGY that the crowbar will allow through into an arcing tube.

I have never found an arc crater in a grid-filament shorted 3-500Z that 
I autopsied.  Sure, some kind of current surge must have produced the 
strong EMF which bent the hot filament helices enough for them to short 
against the grid, but it was seemingly not an arc.   If it had been an 
arc, the tube must have had gas, but I have never seen a gassy tube 
with a grid-filament short.

> Not current; energy.

Current produces the EMF which bends conductors and damages tubes.
>
> A scanned version of the original OE5JFL article is now at:
> www.ifwtech.co.uk/g3sek/misc/oe5jfl.pdf
> It is in German and English, and contains a few more details than 
> Hannes's web page.
>
>
>
> -- 
> 73 from Ian G3SEK         'In Practice' columnist for RadCom (RSGB)
> http://www.ifwtech.co.uk/g3sek
> _______________________________________________
> Amps mailing list
> Amps at contesting.com
> http://lists.contesting.com/mailman/listinfo/amps
>
>
Richard L. Measures, AG6K, 805.386.3734.  www.somis.org