From: Winfield Hill
Subject: Re: A protection problem
Date: 8 Jan 2003 04:41:04 -0800
Organization: Rowland Institute
X-Newsreader: Direct Read News 4.20
> Winfield Hill wrote:
>> Ah, you need a self-resetting feature. OK, this will do the trick.
>> High-voltage back-to-back depletion-mode FETs. What we have here
>> are two FETs that are both normally fully ON, and hence nearly out
>> of the circuit, but which instantly go into current limiting if the
>> voltage across R3 reaches about 1.5 volts (150mA). The protective
>> circuit is inside the feedback loop to eliminate any voltage-drop
>> problems. The point marked "x" is a pair of protection diodes to the
>> two supply rails. "y" is two Schottky diodes to the supply rails.
>>. ,-- R1 --+-- R2a --+-- R2b -------------------, Vo = G * Vprog,
>>. | | | | G = (1 + R2/R1)
>>. gnd | __ x R3 |
>>. '--|- \________ s____ 10_____s __|____ Vo 0 - 25V
>>. ,--|+_/ | |__| |__ __| |__| 100mA
>>. | y --, \/ ,--
>>. prog ---' |_____/\_____| circuit withstands
>>. 400V load "fault"
>> The depletion-mode MOSFETs are Supertex DN2540N5, which have an
>> I_DSS of 150mA min, and are rated for 400V.
Data sheet at http://www.supertex.com/pdf/datasheets/DN2540.pdf
> Nice circuit but the FET is going to be doing about 90W and in TO220
> only rated at 15 :( I suppose ... maybe they would last long enough
> open an isolation relay.
Excellent thinking, nospam. Your suggestion combines Frank Bemelman's
slow-acting but highly-effective mechanical relay with my fast-acting
electronic limiter. Together they can protect Ken's fragile precision
25V 100mA output driver stage.
At 150mA and 400V the current-limiting FETs will be dissipating 60W.
Their thermal mass can easily handle this for 25 to 50ms while the
Ooops! The DN2540 is being discontinued! :<( A close replacement
is their DN3545. http://www.supertex.com/pdf/datasheets/DN3545.pdf
But although capable of 1.6A 450V, it's only available in wimpy TO-92
or TO-243AA packages. :<((( The latter has a tab, where one can add
some thermal mass, but... They also sell the die, saying "Significant
PD increase possible on ceramic substrate." That's an approach that
companies like APEX regularly use to advantage with Supertex FETs.
It's possible the DN3545 can be made to handle 60 watts long enough
for the relay to open, assuming a fast relay. I'm not sure how to
interpret Supertex' Thermal Response Characteristics curves in the
data sheet, it seems they haven't added anything to the TO-243 part's