Mark, I spiced this thing (Fender Twin Reverb AB763 TABv3 p.483) up
good and your numbers for the -3db points and notch frequency are
right on even with all your approximations. Nice analysis, very
impressive and enlightening when I took the time to understand it and
draw even further reduced circuits for high and low frequencies. When
I included a non-zero source impedance on the input side, a finite
load impedance on the output side, and threw in all the parts you
threw out, it only made a difference in about the third significant
figure. The notch is indeed at 295Hz.
In general, the non-zero source impedance cuts highs (it is a divider
with Rb), and the finite output impedance cuts lows (it is a divider
with Rt). In the Fender circuit, the cuts are about equal so the
whole response drops fairly evenly.
>Writing a program to do the math above would be pretty easy. Then
>just vary Rb and see the resultant curve: As Rb gets smaller, the
>notch frequency will move upward and the attenuation in midband will
>decrease.
Yes, that is what I saw, but the other equally important effect is
that as Rb is decreased the treble response really falls off as the
treble gets shunted to ground through Rb and Cb. When Rb gets on the
order of 5K or so the "notch" appears more like this:
__
\
\
\ ____
\/
Not a Good Thing.
My new grokking of the Fender/Marshall type tone control is now:
Turning up the bass and treble *increases* the bass and treble
frequencies from a flat response, leaving a notch in the middle. The
corner and notch frequencies can be calculated like you said. The
midrange control is not *really* a midrange - it sort of sets the
"floor" for the whole response (not surprising from its position at
the bottom of the whole tone divider circuit) - as it is turned up, it
tries to compress the whole frequency response upwards to a flat line,
with the greatest affect (most compression) on the most deeply cut
frequencies. Since these are usually the frequencies in the notch,
this means it is usually turning up the midrange, but it is just as
capable of turning up the bass and treble when those controls are
turned down. The midrange control cannot boost the notch frequencies
high enough to flatten the response, though; the only way to get a
flat response is to turn the bass and treble all the way down to
eliminate the notch. The corollary to this is that turning up the
midrange will *not* boost the mids *above* the bass and treble.
I've got this weird feeling that I would have been better off not
knowing all that, but too late now :-)
And Rb, the "sweep" control? You're right, it looks better to change
the corner frequencies by adjusting Ct and Cb because of Rb's
deleterious effect on the treble. (Note that the 100k Fender value
actually seems to boost the highs somewhat; 50k gives a flatter
bass/treble balance in simulation).
--
Tom.
From PYLO--(at)--su.edu Mon Nov 27 16:47:41 CST 1995
Article: 5850 of alt.guitar.amps
From: PYLO--(at)--su.edu (Ruth and Dale VanZile)
Newsgroups: alt.guitar.amps
Subject: Re: Tone circuits: part 3
Date: Sun, 26 Nov 1995 13:03:29
Organization: Arizona State University
Lines: 25
Message-ID:
References: <499a1r$1o--(at)--anix2.panix.com>
NNTP-Posting-Host: ppp2-13.inre.asu.edu
X-Newsreader: Trumpet for Windows [Version 1.0 Rev A]
In article <499a1r$1o--(at)--anix2.panix.com> mgarvi--(at)--anix.com (Mark Garvin) writes:
>Please try to post follow-up queries. No problem with email generally,
>but I will not have as much time to reply to separate email questions.
>This post has already taken long enough. Darn ASCII diagrams.
>Anyway, I hope this has helped to answer a few of the email queries.
>Sorry for the delay.
I think also that the Fender circuit has a bit less insertion loss. I'd be
interested to see your thoughts and analysis on the Hiwatt tone circuit and
the 6G7 Fender Bandmaster. I've done some freq. response analyses on these
circuits using a program called "ECAP". The Hiwatt actually has a bit of a
mid-boosting action to it in the area of 400Hz to 2500Hz (estimated #s, as
it's been a while since I ran the circuits) or so, along with a serious scoop
at about 300Hz. The 6G7 has a big dip at about 800 Hz, and both have better
bass response than the later Fenders. IMO, they are really nice sounding
circuits. The 6G7 sounds brown and almost Marshall-esque, and the Hiwatt is
crisp but still packs some punch in the lows, along with a "clean" sounding
low mid area. Do you have a print of these? I can send a copy of each if you
need it. Or, if you're in the NYC area, I can drop one off over the holiday
break, as I'm going to be back that way for a wedding....
Dutch
From mgarvi--(at)--anix.com Wed Nov 29 00:29:58 CST 1995
Article: 5911 of alt.guitar.amps
From: mgarvi--(at)--anix.com (Mark Garvin)
Newsgroups: alt.guitar.amps
Subject: Re: Tone circuits: part 3
Date: 28 Nov 1995 21:05:16 -0500
Organization: PANIX Public Access Internet and Unix, NYC
Lines: 76
Message-ID: <49gf4s$ga--(at)--anix2.panix.com>
References: <499a1r$1o--(at)--anix2.panix.com>
NNTP-Posting-Host: panix2.panix.com
In <499a1r$1o--(at)--anix2.panix.com> mgarvi--(at)--anix.com (Mark Garvin) writes:
>TONE CIRCUITS, PART 3. Fender controls vs hifi controls.
>----------------- BASS and TREBLE BOOST (MID CUT) ------------------
> In >---+-------------------------------
> | |
> / |
> / R1 ---
> / --- C3
> / |
> | R4 |
> +--------------//////----------+--------> Out
> |
> ---
> --- C2
> |
> /
> / R3
> /
> /
> |
> -----
> ---
> -
By the way, assuming that the 'left' and 'right' R-C sections do not
interact (left RC net is low impedance compared to right-side RC net),
the center frequency (and therefore the attenuation) can be
approximated as:
159000
------------------------------
_ /-------------------------- <-attempt at square root symbol
\/ R1 * C2 * R4 * C3
where Cap values are in microfarads
R3 is ignored, since it is usually small compared to R1
>----------------- BASS and TREBLE CUT (MID BOOST) ------------------
> In >---+
> |
> /
> / R1
> /
> /
> |
> --- C1
> ---
> | R4
> +------//////-------+--------> Out
> | |
> / ---
> / R3 --- C4
> / |
> / |
> | |
> ----- -----
> --- ---
> - -
Pretty much the same here, ignoring R1
159000
------------------------------
_ /-------------------------- <-attempt at square root symbol
\/ C1 * R3 * R4 * C4
>Mark Garvin
>Composer/Design Engineer
>New York City
Index:
