Amps
[Top] [All Lists]

[AMPS] Glitch Protection

To: <amps@contesting.com>
Subject: [AMPS] Glitch Protection
From: 2@vc.net (2)
Date: Wed, 29 Aug 2001 14:10:12 -0700
>
>2 wrote:
>>
>>>
>>>I have a question for the group that I have seen discussed before. It
>>>regards glitch protection.
>>>
>>>It has already been pointed out that fusing the HV supply line provides
>>>little if any protection due to the way fuses blow when faced with a
>>>castatrophic event. What I am considering is inserting a normally closed
>>>vacuum relay in the plate voltage line and triggering it with an electronic
>>>over-current sensor monitoring plate current. I see that the Jennings RF-4
>>>series relays can open as quickly as 4 msec.
>>
>>//  Can the contacts break 200A at 4kV?
>>
>Quite so - the purpose of the glitch resistor is to limit the current to
>a maximum that will not damage the tube. The resistor (along with the
>series resistance and inductance of the smoothing cap) limits the peak
>current, even in the event of a crowbar short from B+ to chassis. 
>
>For tubes up to 1500W dissipation, Eimac suggests limiting the current
>to about 40A, which is why glitch resistors have a value of a few tens
>of ohms for the typical 2-3-4kV B+. Without the resistor, peak currents
>are much higher.
>
>>  
>>>Has Eimac ever done any studies delving into how fast an 8877/3CX1200/3-500
>>>transitions from stable to "full-on" parasitic oscillation? 
>>
>CONFUSION ALERT  - who says *all* these events are due to parasitic
>oscillations? What about gas arcs (there are a billion neutral gas atoms
>inside a typical so-called "vacuum" tube) or direct shorts from B+ to
>chassis? 
>
>Let's just say that big current surges sometimes *do* happen, and the
>amp needs protection against them.
>
>>//  Eimac has been somewhat silent on the subject of parasites.  I have 
>>never seen a Eimac Engineering Bulletin on such.   QST ex-staffer David 
>>A. Newkirk theorized that a parasitic could reach full throttle in 10 to 
>>100 micro-seconds.   
>
>Maybe so, but if the current surge is caused by a physical crowbar
>short, the rise time could be much faster. In either case, even a 'high
>speed' relay is waaay too slow.
>
>However, there *is* a place for a B+ over-current relay in glitch
>protection - that is to disconnect the mains supply to the B+
>transformer. Otherwise, whatever caused the glitch may just keep on
>going (been there, seen that... it seemed to take me forever to hit the
>switch). For this backup protection, a reaction time of a few
>milliseconds is fine - but ONLY if there is a glitch resistor to catch
>and limit the initial surge.
> 
>>Eimac was slow in recommending a glitch resistor be 
>>put in series with the anode supply.  
>>
>Slower than when? Eimac's Bulletin #17 on 'Fault Protection' (which I
>quoted above) was published in 1987.
>
//  I first heard about it in 1979.  

>>>One would have
>>>to take into accounts that Parasitic Oscillations can range in degree from
>>>mild to severe. Would such circuitry add any degree of protection?
>>>
>>//   My guess is probably not.  However, a resistor will do the job 
>>providing it can pass the short to ground test -- i.e., subsequently 
>>exhibit no physical damage or change in resistance.  The beauty of a 
>>glitch resistor is simplicity.  
>
>And also speed - the resistor has no 'reaction time' of its own, 

//  good point

>and actually helps slow down the rise-time of the current.
>
>>For instance, typical grid-current 
>>"protection" circuits for 8877s use a 1A to 2A transistor to interrupt 
>>the flow of grid-current.  However, during a parasite, grid-current goes 
>>out of sight.  An 8877 cathode is capable of >50A-peak emission.  This 
>>amount of current will C-E short a 2A transistor in short order.  Thus, 
>>there is essentially no grid protection during a major parasitic 
>>oscillation. 
>
>Please can you give an example or reference to such a deficient circuit?
>I've never seen one like you describe (but maybe I've led a sheltered
>life). Where exactly is this 2A transistor?
>
//  In the grid-current path -- for an 8877, between the cathode lead and 
gnd.  
>
cheerio,  Ian

-  R. L. Measures, 805.386.3734,AG6K, www.vcnet.com/measures.  
end


--
FAQ on WWW:               http://www.contesting.com/FAQ/amps
Submissions:              amps@contesting.com
Administrative requests:  amps-REQUEST@contesting.com
Problems:                 owner-amps@contesting.com


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