From: Winfield Hill
Subject: Re: High-side switch
Date: 31 Dec 2002 10:42:44 -0800
Organization: Rowland Institute
X-Newsreader: Direct Read News 4.20
> the renowned Winfield Hill wrote:
>> I've used the IR2113-2, rated at 600V. I got an efficient
>> 25ns risetime using a common IRF820 n-channel switching FET.
>> Most of the high-voltage driver ICs have one peculiarity or
>> another, in the case of the '2113 it was a high logic-input
>> threshold, 50% of the low-end supply or about 9V in my case.
> Also, don't forget that you can't have 100% duty cycle
> because the charge would no longer be pumped! The LVD should
> kick in and snap the main switch off under those conditions.
Right. Usually anyway; in my design referred to above (RIS-427)
I had an additional transformer secondary to provide a floating
14V for the flying high-side FET driver in the '2113, allowing
continuous dc switching. You may remember that we discussed the
design here on s.e.d. and I posted the schematic. Resulting in
Tony, IIRC, to comment that 250V 15ns FET switching will produce
a capacitively-coupled transient current through the transformer
(a Signal 241-3-10) into the AC mains. Oops! Guilty as charged.
A quick measurement now shows the transformer's winding-winding
capacitance to be less than 8pF (thanks to its split-bobbin
construction), which would make a 135mA 15-ns long common-mode
spike for the fastest case. I suppose we didn't notice this on
our noisy ac-power lines. :>) However, this instrument should
have included an input line filter, as in a computer's SMPS, etc.