Re: Cockroft-Walton question.

On 9 Jul 2005 18:02:52 -0700, Winfield Hill
<Winfield_member@xxxxxxxxxxx> wrote:

>John Larkin wrote...
>>
>> Hey, this is fun:
>> http://www.analog.com/en/prod/0,,768_833_ADG465,00.html
>
> Those I-V curves are not unfamiliar to someone who has used CMOS
> fault-protected multiplexers, such as the HI-5108 (AoE page 150),
> max358, adg438 or dg458. These switches disconnect the signal if
> it goes above the supply rails. But these parts quickly run into
> trouble if the signal gets too high (e.g., over +/-25V).
>
> For a precision series-protection current-limiters I like to use
> a pair of back-to-back LND150 or LN250 depletion-mode MOSFETs.
> These have the advantage of protecting to over +/-500V. They do
> present 2k-ohms of series resistance, which can add some Johnson
> noise. If that's important the 180-ohm dg458, or your new 80-ohm
> adg465 might be preferred, but only provided that you can keep the
> excess external voltage to under +/-35 to 40V. That fairly-low
> working voltage seems rather limiting to me. :>)

Agreed; I'd prefer something that could stand AC line voltage and
transmit 15 or 20 mA of signal. The LNDs are great when you can
tolerate the low Idss. At least the 465 allows my customers to
mis-wire a 24 volt circuit - pretty common in their biz - into my dac
thingies. Claire has a new opto ssr that's self-protecting, maybe
worth it as a limiting element alone.

We were playing with polyfuse/transzorb combinations, but couldn't
come up with a comfortable arrangement, especially with the space
restrictions of this project.

John

.

Quantcast