another DIY option which has been discussed elsewhere on the forum:
http://www.hoffmann-hochfrequenz.de/downloads/lono.pdf
Not sure whether this is the best design where it comes to distortion. I see a whole lot of MLCCs in the signal path which have a voltage dependant capacitance and thus the amplifier has an amplitude dependant frequency response.
The early versions had foil capacitors, truckloads of 'em, and it bought nothing. And capacitors with
no voltage across them don't change. When there is AC across them, they are too small.
Really, performance at the low end suffered because even the bank of foil capacitor at the input was
too small, resulting in a much worse than 1/f rise of noise below 60 Hz. The current op amp version
has a wet a slug tantalum at the input but this is problematic, too. It needs some sequencing when you
attach the preamp to a low impedance DUT with some DC, and using it with non-low-impedance DUT
is pearls before swine.
When you mis-sequence, the transient through the huge tantal may zener the op amp's input transistors.
There is now a version with many small jFETs (forget IF3601/02 !) but it is next to impossible to
make it exhibit no negative real part of the input impedance somewhere. Another wording for instable,
and shared by all published designs that I analyzed. When you are aiming for 1 MHz BW, the feedback
simply comes to late. It seems that i can get 100 KHz with REALLY fast op amps in the feedback loop,
but their 1/f corners do hurt, even with the FETs as a first stage. Between a rock and a hard place.
( I talk about CS FET, Cascode, op amp for loop gain, feedback to source)
Cheers, Gerhard
(above web site is mine)