Introducing the MAGMA front end for BA-2 and F4

[audiorob]

The MAGMA is a new front end for a BA-2, F4, or even a slightly
modified Zen-5 amp. It was concieved as a new front end for
the BA-2 that used magnetics to step up the input voltage and was
therefore named MAG-BA. Realizing that it could also be used on
the F4 or Zen-5, the BA part no longer fit. Wanting to stay with
the Burning Amp theme, it is named MAGMA.

The MAGMA is similar in concept (I have no idea about sound) to
the M2, in that:

* it uses a high performance transformer to magnetically step up
input voltage
* aside for degenerative source resistors, it uses no feedback
* it uses a complementary OPS architecture
* as designed, it is fully class A
* as designed, it outputs about 25 watts RMS
* single ended input
* using differential input, pairs of amps should be able to be
bridged together for higher wattage

and so on...

The simplified amplifier architecture is a 2 stage design using a
signal step up transformer followed by a current amplification
section. The complete design uses JFET impedance matching buffers
around the step up tranformer to help linearize the circuit. As
seen in the schematic, input voltage is fed to an impedance matching
buffer that drives the voltage step up transformer, which drives
another impedance matching buffer, which drives the current gain
stage.

The BA-2 can be used as the current gain stage (OPS). Simply
replace the original IPS/VAS of the BA-2 with the buffer/xfmr/buffer
of the MAGMA.

The output section of the F4 and BA-2 are almost identical.
Obvious differences include the DC offset adjustments on opposite
rails, slightly different component values, and possibly a
different number of MOSFETs. Note the inputs to the OPS are
slightly different. The BA-2 also has a facility to provide
feedback, which this design does not use.

As the OPS of the BA-2 and the F4 are the same basic design, and
as the F4 already has a buffer on its front end, the input buffer
of the F4 can be used (without modification) as the second buffer
of the MAGMA. So the new circuitry consists of an input buffer
followed by the transformer (and associated RC network) prefixed
to the normal F4 circuit. Danged ol' easy!

The vertical lines on the schematic try to indicate the MAGMA/F4/BA-2
sections.

As the MAGMA using the F4 boards from the DIY Store is an easier
build than the BA-2, and as I don't need the extra output devices of
the BA-2, I built the MAGMA prefixed to the F4. I will leave it to the
more industrious types to use the BA-2.


Parts:
The resistors are 1/4 W with values noted on the schematic
The POT P301 is 20 ohms, 1/4 W.
JFETS
Q1 and Q301 are 2SK170BL (N-Channel)
Q2 and Q302 are 2SJ74BL (P-Channel)
Jensen Transformer JT-13K7-A
Cap C301 1300pF, 35V


Construction Notes:
Before installing P301, center the resistance so that it is about
10 ohms between the center and outside pins.

I apologize for the symbol my spice program uses for the P-Channel
JFET. The Drain pin of the P-Channel JFET is towards the negative
rail, and the Source pin is towards the POT.

I run the JFET devices (Q301, Q302) a little hotter than Nelson did
in the F4. You can reduce the current to these devices by using
higher values for R303 and R304 (Nelson uses 1K in the F4). Or
you can leave them where I have them and optionally use little heat
sink tops.

For C301, if you cannot find a 1300pF cap, you can replace it
with a 1500pF cap. I bought a bunch of 1200pF polystyrene caps
and measured one at 1270pF, so I used it. If you can replace
R2 with a 39.2K resistor, the 1500pF value for C301 is just
right.


Initial Turn-on and Adjustments:
Of course, follow Nelson's instructions in the BA-2 paper. In
addition, with no input, measure the voltage between the input
of the signal transformer to ground. Adjust P301 to zero this
value.

do
{
Adjust P301
Adjust bias
Adjust DC offset
Wait a while
} until settings and heat sink temperature are stable


Some measured specs of my prototype into an 8 ohm load:
Distortion @ 1 watt @ 1KHz ...... 0.025%
Input Impedance ........................... 100 Kohm
Gain ................................................ 4.485 or 13.03 dB
Input Sensitivity ............................ 633 mV = 1 W, 3.436 V = 29 W
Output power 8 ohms ................. 29 W RMS @ 0.3% THD, 1KHz
.................................................. ........ 25 W RMS @ < 0.2% THD, 1KHz
Frequency response ................... +0, - 0.1 dB from 20 Hz to 20 KHz
.................................................. ........ -3dB around 70KHz

I'll try to post some graphs in a bit...


Sound:
As of this posting, I have only built one channel. So I offer
this only as a rough critique. I think this sounds much like
my DIY F5 (like a complementary amp), except with a little more
"curves" and less "angles". It may be a little softer on the
entire spectrum, but it is hard for me to tell without a sterio
pair. But I'm working on that...

I would like to thank Nelson for permission to post this design.
I would like to thank Nelson and all of the others in this
group for the support they have given me.


Robert Fournerat

[Zen Mod]

[Nelson Pass]

Ooooooohh....

Super happy shiny toy!

[DougL]

Robert,
I love what you are doing.
Quick question. Have you tried to wire the JT-13K7-A as an auto-former?
If I understand correctly, you get better coupling and slightly higher gain at the expense of higher loading.
Alternately, since the transformer "blocks" DC coupling, would a DCB1 style Shunt regulator have advantages?

Plotting big trouble for moose and squirrel,

Doug

[audiorob]

Originally Posted by DougL

Robert,
I love what you are doing.
Quick question. Have you tried to wire the JT-13K7-A as an auto-former?
If I understand correctly, you get better coupling and slightly higher gain at the expense of higher loading.
Alternately, since the transformer "blocks" DC coupling, would a DCB1 style Shunt regulator have advantages?

Plotting big trouble for moose and squirrel,

Doug

Hi Doug,

Sorry for the slow reply, I've been building that 2nd channel, which I
can now report is done!

I got the JT-13K7-A directly from Jensen Transformer, and they are in a
sealed (I assume Faraday) enclosure. I looked for some autoformers,
and even built a few for experiments. But I could not find or build (for a
reasonable price) any with the performance of the Jensen's. In software
(which is what I do) there is a saying that goes, "Why re-invent the
wheel". I could not build a better wheel (or xfmr), so I used the Jensens.

I'm not familiar with the DCB1 project. I'll have to read more about it.

Robert

[DougL]

You can wire the Jensen as a transformer or as an autoformer.
As an autoformer, I believe you wire red to orange, rather than orange to ground.
I agree that the Jensen is a fine unit.

Thanks for sharing your work.

Doug

[audiorob]

Shewww, I thought you ment to rewire the core! And I wasn't
looking forward to that task.

That is very interesting news to me. I'll try to check with Jensen
tomorrow. Thanks for the info!!!

Robert

[flg]

I expect you have read this thread?
Transformer as a voltage gain element

[Nelson Pass]

This is an interesting subject. If you want to pursue it, I
suggest that you try as many different samples of transformer
as you can, as they are all different, and there are some
surprises to be had. Also keep in mind that these pieces all
love a low source impedance.

[flg]

Well, sounds like audiorob's proposed JFET buffer input would be a good start. Now, a standard step up connection or the autoformer connection?
Hmmm