From: Mike Monett
X-Mailer: Mozilla 2.02 (Win16; I)
Subject: Re: Prof Searle's 100MHz 30dB 1968-era 3-trannie amplifier
References: <3E2358BE.firstname.lastname@example.org> <3E2A0C89.7FAB@sneakemail.com> <3E2A4E8D.78BE@sneakemail.com>
Date: Sun, 19 Jan 2003 06:05:49 -0500
NNTP-Posting-Date: Sun, 19 Jan 2003 06:05:14 EST
Organization: Bell Sympatico
Winfield Hill wrote:
> Mike Monett wrote...
> >Winfield Hill wrote:
> Actually it gets better not worse. Remember Bob Pease and his classic
> rat's nest lashups to test some IC-designs on the bench? It keeps the
> capacitance down and I imagine low inductance is a possibility with some
> effort at places you really need it.
Actually, you are right. I always get parasitic oscillations with fast
transistors and long leads, but I guess a couple of nanohenries really isn't
going to be a problem up to 100MHz. The reactance is so low compared to the
rest of the circuit it won't do much. But I was horrified when I first saw
it. This ain't gonna work. But it worked fine.
> > I tossed this in SPICE using three MMBR941's, the last one as an emitter
> > follower. After fiddling a bit with the bias, here's my results:
> > Parameter 1KHz 200MHz
> > Open Loop Gain 46.14 dB 37.3 dB
> > Closed Loop Gain 30.50 dB 29.99 dB
> > Maximum Output 9.11V p-p
> > No compensation cap was needed, but it definitely needed the 100uF from
> > Q2 emitter to ground.
> Hmm, pretty nice! That was with a 50-ohm load, right? Did you find
> the -3dB down point? Also, what supply voltage did you use, are you
> sure each part's Pd was reasonable?
The parts were cooking. No load, but I set the bias for max swing just to
find the range. Vcc is +15V. The -3dB point is 558MHz.
> Hmm, did On Semiconductor pick up or continue the '941 after Motorola?
I haven't been able to find anything at ONSemi. MaCom supposedly picked up
the line but I haven't been able to figure out how to navigate their web
site to find out.
> I see APT and Microsemi 2nd-source the part. Philips has the PBR941.
For some strange reason, they offer the PBR957 as the replacement for the
MMBR941 instead of the PBR941. They also have some nice high frequency ones
in the BFG series. Here's a short list:
These have a VCEO of 4.5V. I don't know what the base-emitter breakdown is,
but it can't be more than the MMBR941 at -1.5V. So power on/off transients
could be a problem if there's large caps in the circuit. It may not hurt the
device, but I understand all it takes is one event to degrade the noise
> But Mike, the MMBR941 wasn't available in 1968 was it? When did it
> come out? I mean, I don't remember any jelly-bean 8GHz transistors
> that far back, and having only 0.3pF of capacitance is a big help!
True, but the 2N2222 didn't do very well, and I happened to have the model
for the MMBR941 already loaded and wanted to see what it could do. In real
life, there would be three parasitic oscillations all running happily:)
> Let's see what we can accomplish with a true vintage component, like
> the 2n918, or better if it was available.
Yes, that's a much more difficult task.
> > Now I'd like to see how it really should be done!
> Nah, your circuit the looks like the winner. But I'll show you my
> 2n918 circuit if you show me yours. :>)
LOL! I'll upload it tommorow and let you shoot holes in it. But I really
want to find out how you can get away without using a bypass cap.
> - Win