From: email@example.com (john jardine)
Subject: Re: Wideband gain measurement
Date: 9 Oct 2002 07:18:00 -0700
References: <3DA28DE4.1125B4FC@earthlink.net> <firstname.lastname@example.org>
NNTP-Posting-Date: 9 Oct 2002 14:18:00 GMT
Winfield Hill wrote in message news:...
> john jardine wrote...
> >John Woodgate wrote
> >> You will need to check the attenuation over the upper part of the
> >> frequency range. Even a 10:1 attenuator usually needs capacitance
> >> balancing at 1 MHz, and you'd be surprised what a 1 kohm resistor
> >> may measure as, at 1 MHz.
> > I was very surprised (honest!) to find that the 5%, 0.4W, carbon film,
> > '1k'(1020 ohm, cropped leads), resistor, measured recently on my
> > friendly, indispensable, 50 year old 'Boonton RX250' meter, at 20MHz,
> > dropped in value to 1005 ohms. (slight '-0.3pf' capacitance).
> Likely bad wiring. :-) Much of that is resistor lead capacitance,
> rather than resistor body capacitance per se. Here's how to make
> resistive attenuators with good high-frequency performance.
> . =====++=========++=========++============
> . || || ||
> . _||_ _||_ _||_
> . ---|____|--+--|____|--+--|____|--
> . || | || | ||
> . || | | || | | ||
> . || |_| || |_| ||
> . || | || | ||
> . =====++====+====++====+====++============
> Place the resistors through holes in intercompartment ground shields,
> this greatly reduces end-to-end capacitance. Limit the attenuation
> to 20dB per stage. This technique works well to 200MHz at 50 ohms.
> - Win
Yes, it was(is) 'bad wires' ;-)
Fortunate to have a home workshop I once milled and drilled my way
through 8" of square bar to form construction of a 600ohm 120Db
atten. Resistors were all nearly totally encased in solid metal. Nigh
on perfect to way beyond 100kHz. Definitely not cost effective!.