Clone Marantz 7 M7 Line Amp

 

The heart of Marantz 7 M7  line amp is a two common cathode stage with negative feedback followed by a cathode follower.

The design is simple with few components, but  both the tone control and filter circuit  deteriorates the sound details a lot  also  so many  switch contacts inserted further down grade the  sound quality , clone results from different  DIYers  varies.

The circuit is simple,  but the result could be  great  if  built correctly.

The B+ supply voltage seems random but actually is carefully selected to  handle up to  +20 db signal peak at very low frequency end.

The 500K trim pot works for better  S/N ratio  but also cutting down the high frequency response as a side effect.

In an original M7 machine, there is no any  B+ supply stability arrangement and too much switch contacts inserted,  the over all result is a bit foggy  sound and  not enough detail  in midrange,  also  a bit weak at high frequency end.

The heart design is supper to see whether we can dig the treasure out or not.

Some  Diyers agree  a point of  simplest is the best,  if this is always  true then  a  single triode common cathode amp  could dominate .  But the fact  is  different  designs share the market in competition.

Actually the  famous  Mertisse  Reference  line amp  has the similar  philosophy  with M7 .

How can  a line amp  of two common cathode stage with negative feedback plus a cathode follower win  a position in  the high end world for decades ?  Let’s go and see.
 
 

 
 
The electron flow inside a triode see picture:

electron02

The negative bias at grid wire creates an electric field  near around  the grid wire, and this electric field size varies  as  bias changes , the result is changing the electron path width between grid wires as well as plate current, see picture:

electron03

Due to the mass of electrons the final value of plate current responses to bias changing will not  be reached immediately after the bias changed. There will be an overshoot  before the final value , an appropriate negative feedback arrangement will improve the over shoot a lot, see picture:

electron05

 

Please note the situation is intentionally exaggerated to clearly show the case.

In case of no NFB,  for low frequency the bias changing is slow enough to be followed so that there will be fine, for high frequency side the input and output stray capacitance will automatically correct the over shoot and  even further  turn down the high frequency response.  So that the overall result is an excellent base,  a too bright mid  and a bit darken high.

electron13

 

This end result often can be find  in single triode line amps, even  a  cathode follower  added,  but not in power amps , this is because the line amp output signal will be further amplified whilst the power amp is  almost  the last stage  to the speakers.

So that the theory of  non  negative feedback  applied at  single end tube power amp is correct,  but for preamp the story might be different. I  understood  some  DIYer  prefers  non NFB triode preamp for the attractive bright middle human voice  as  a special taste.

The negative feedback circuit compares the output and input signal and corrects the overshoot far before the peak wave form comes, in high frequency side,  NFB also detects and corrects the under shoot to maintain a perfect response. The NFB loop is short and the response of 12AX7 is fast enough to cover up to over 100KHz and fulfills the  high end standards.

 

The M7  NFB loop was intentionally designed not including the cathode follower to get a shortest loop path so that it is fast enough for transience response. Any similar line amp design NFB loop including the cathode follower could not get a better result  as  the response speed slows down.

The only shortcoming of  original  M7 line amp is  a  +6 db response at  about  3 Hz at output point.  This problem was overcome  by correct choose the B+ voltage and plate operating point. Yet  +6 db in 3Hz might be welcome for increasing  low punch, no longer a shortcoming.

A well fine turned DIY M7 machine stuck to this rules could become  one of the best tube line amp in the world.

Please see the original  M7 design  with  30Hz high pass filter for record turntable rumbles as in 1960s high end turntable was not popular as to day.  Again the 5000Hz low pass filter is for tracking dust noise from play black plastic records. A  30Hz to 5000Hz response is very similar to that of AM broadcasting.  So that bypass switches can override the filters to handle  better quality sound sources.

electron06

Nowadays we no longer need these filters please see picture:

electron08

The overall gain of V1 and V2 is about  18 , the 500K trim pot works with 180K will give a  0.632   gain  when setting at the center  point , and the cathode follower gain is 0.97.

18×0.632×0.97= 11

This matches  the line amp standard gain of 10.

By  examine  the trim pot  circuit:  (500K+180K)/4= 170K

If the input circuit and grid to cathode stray capacitance of the cathode follower reaches 30  pf , a typical value, the high frequency response will be  about  30KHz  – 3db. It seems ok.

But if we don’t mind  the overall gain  18×0.97= 17.5 , we could override the 500K trim pot and achieve a high frequency response of  100KHz  -0.6 db.

I suppose most of the DIYers  would  choose override the trim pot, please see picture:

electron11
 
 

 
 
This is the original designed  component value of M7  line amp, if  we  follows these values could get a very good result:   10Hz to 50KHz +/-  0.5 db or  better, trust me,  this is very true.

I understood most  experienced DIYers  didn’t feel comfortable  with the  0.01 uf  coupling capacitor at  V1 plate,  copes with 1M  grid  resistor  the low frequency  -3db  point is about 16Hz,  too high for high end purpose. Yet some DIYers would like to modify the machine for better performance,  then  enjoy the excellent  sound with satisfaction.  The most common suggestion is using a  0.1 uf  to replace the 0.01 uf  cap  so that the -3 db point will be  1.6Hz, much better.  But,  don’t  do  this  please ,  it will  ruin the machine in both high and low  frequencies.

To clone a  M7  line amp  , NEVER  do the following  two things:

  1. Never get the output signal directly from V2 plate so as to avoid  two  capacitors  in series as well  as  saving  the signal path,  because  this connection  might  overload  the cathode follower , see picture:

electron15

  1. Never change the 0.01 uf capacitor to 0.1 uf  with other capacitor remains unchanged.  This will  make  V2 plate   See picture:

electron16

The V1 and V2 form  the heart  amp ,  high  speed negative feedback  loop expends  the wide range frequency response,  the cathode follower provides impedance matching,  high input impedance doesn’t burden the heart amp  (very important point )and low output impedance could fight  for signal cable stray capacitance to maintain high frequency driver capability. Further  more,  the cathode follower  could  improve horn driver metallic  sound  so M7 is good for Altec 288/ 290  and Tannoy drivers because the mid range under control not too bright.  A perfect design.

 

Key  point :  The  M7  line amp secret  formula

There are  two  RC networks  in the heart amp :

  1. C1, R1  are forward  RC  network
  2. C2,  (R2+R3)  are  feedback  RC network

See picture:

electron12

The  RC  time constant of  forward and feedback  network should  be equal to each other  for excellent  performance ,  so that  the secret  formula  is :

C1 R1 =  C2  (R2+R3)

If we choose 0.1 uf  as C1  then:

(0.1 uf)(1000K) =  C2  (82K+ 4.7K)

Then:

C2 =  (0.1 uf)(1000K)/ (82K+ 4.7K)

C2 = (0.1 uf)(1000K)/(86.7K)

C2 = 1.15 uf   We can use  1.2 uf

That  means if we change C1  to  0.1 uf,  we have to change  C2  to  1.2 uf  otherwise RC  network out of balance causing frequency response  out of balance as well.

Please see original M7  circuit:

electron18

R2 = (4.7K+82K)// 150K// 270K// (500K+180K) =  42.7K

Where  C1 R1 = C2 R2

( 0.01 uf) (1000K) =  C2 (42.7K)

C2 = (0.01uf) (1000K) / (42.7K) =  0.23 uf   The M7  designer  uses 0.22uf

I suggested the final version of  M7  line amp DIY :

electron17

The V1 open loop gain :

270K // 1M +  80K  +  (4.7K  X  101)

=  212.6K +80K + 474.7K

= 767.3k

V1 gain  =  100 (212.6/767.3)

= 27.7

The V2 open loop gain :

100K //  (82K +4.7K )  +  80K  + (1K  X101)

= 46.4K  +80K +101K

=227.4K

V2 gain = 100 (46.4/227.4)

= 20.4

 

A = V1, V2 open loop gain

= 27.7 x 20.4

= 565.08

B =  4.7K /  (82K + 4.7K)

=  4.7/  86.7

= 0.0542  (known as feedback factor)

Close loop gain  =  A  /  ( 1 + AB )

=  565.08/ (1+  565.08X 0.0542)

=  565.08 / 31.62

= 17.87

Over all gain = 17.87 X 0.97

= 17.33
 
 

 
 
Frequency response  :

5 Hz  to  50KHz   +/- 0 db

5 Hz  to  100KHz   – 0.6 db

5Hz  to  300KHz   -3 db

Square wave response :

 

10KHz  output  6 Vpp

10KHz_
20KHz output  6V pp

20KHz_

 

50KHz output  6V pp

50KHz_

Testing on experiment board.

m7_ equipment

m7_test

The actual machine building

m7_board

 

 

m7clone1
 
 

 
 
Power supply arrangement:

For preamps, stabilized B+ supply is a must or bad performance even good design.

Those close loop voltage stabilizing circuits are  slow, any  active power  supply claimed fast becomes slow motion in front of  300KHz.

A very simple method is to use Zener diode, 40 pcs of 7.5V 200mW  Zener connect in series will get 300V.  7.5V Zener diode  has  good sounding.

An example of  7.5V Zener diode array :

electron21

Filament supply strongly recommend DC stabilize using 7812  TO-3  metal package with heat sink.

electron22

Power supply :

electron20

Two 12K resistor use wire wound type better,  7.5V Zener  and 12K  try different brand name to fine tune the sound.  The 200 uf capacitor is carefully selected so don’t go away  from 180  to 220 uf range.  Two  0.1 uf bypass  cap  use high quality capacitors.  V1, V2  and  cathode  follower share  a  same B+  supply is  intentionally arranged not for saving  component budget but for good sounding.

In the line  amp ,  all coupling capacitors  have to use high quality  component,  I  would  not suggest any brand name but  they should have  the loss factor equal to or better than 0.0005 , the bottom line.  Also  the  1960’s  old  type  12AX7 or ECC83 strongly  recommended.

If we change the 82K resistor to 47K , 1.2uF capacitor also change to 2.2uF, the overall gain will be about 10, could give a better sound, worthwhile to try.

Point to point wiring  gets better  sound, print circuit board  is more convenience  to build,  also  neat  component layout  for  visual effect or shortest  path  for excellent sounding is  a  personal  choice, enjoy.

 
Thank you  for  visiting.