the real RCA ba71?

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

hitchhiker

Well-known member
Joined
Jun 5, 2004
Messages
893
Location
vancouver island,bc,canada
Could those of you that can, please explain this design? ie; what the funtions are of the designers' choices.
Is the latter section a type of push pull?

http://www.geocities.com/pepperplace/BA71.html?1097956825668

Years ago I bought an RCA bc7 broadcast console . I was hoping the mic
preamps would be as good as other popular discrete amps of the 60's
and 70's. Well, they don't quite measure up but....

I came accross a schematic and manual for a BA71 and bought it with
the hopes of adjusting what sounded to me as a non-flat design.

To my surprise the schematic was all germanium transistors and a completely different design but with the same transfomers. So I decided to bread board the circuit and it sounds much more like a decent mic pre. I think it would be a very good sounding pre with better transformers because the ones RCA used are not great.

BTW All the BA71 amps that I've seen are the same as that came in my console and they are identical to the BA31 schematic so I think they never produced or issued the real BA71.

Lance
 
[quote author="hitchhiker"]please explain this design? ie; http://www.geocities.com/pepperplace/BA71.html?1097956825668
[/quote]
Q1 + Q2 = cascade common emmiter + common emitter.
Gain is set by negative feedback /R19 forms with R5 voltage divider/
(40) is near 6 dB pad (makes "R19" lower)
This is (inversely) approximate gain of amp (without transformer)
Gain of the transformer is ratio of the square roots of winding impedances.

For exact computing, you must take gain of the feedback divisor

Aur ~= R5/((R8//R16) + R19 + R5)

and to compute gain of amp into after Bode.

Auf= Aua/(1-Aur*Aua)

Aua is gain of the amp take:
Gain of 1-st stage * gain of the 2-nd stage
(if ommit h11,h22 (it can not with Ge transistor, but can scale "incorrect" result with
coeffitient if do it, i.e. 0.5 )

1. compute D.C. current via transistor.
then transconductancy (in mmhos) is 40 * this current.
2. compute amplification coef by multiplying of transconductancy with evetrithing signal
resistance in collector (all Rs direct and after caps in parallel)

this do for 2-nd stage, multiply and scale.

C5 (C6) is derivative in feedback (for stability.)
Q3 and Q5 is emitter followers (gain nearly one). Two followers are
in paralel and DC independent (to avoid failures ?).
Q4 and Q6 are active load for Q3 and Q5. Active load is
kind of push-pull. It uses Q3 Q5 as output and invertors in one device.

It is nice design, stable enough for direct replacement Ge transistors
to Si. Q1 is possible planar Si for good noise figure (it explains why
Vbe of Q1 is so high). If not,
replace it to BC550 (and can change C1 + C2 optionally)
xvlk
 
thank you xvlk.

This is exciting for me. I am now going to dig out my transistor text books
that Ive collected from the used book stores. I've read much of it before
but this gives me incentive to take another run at the theory of BJT design.

I've built this up from NTE cross reference parts and it is actually a very quiet amp with a good sound. I've also had it working well transformerless.

The design uses 1: 1.6 input and 1:2.45 output transformers.

Lance
 
[quote author="hitchhiker"]a very quiet amp with a good sound.
I've also had it working well transformerless.

The design uses 1: 1.6 input and 1:2.45 output transformers.

Lance[/quote]
On the schematics I can read 1500 Ohms as output of transformer.
sqrt(1500)/sqrt(200) is 2.73. It is near to 1:3.
It was often optimal mic transformer for small Si input transistor.

1:1.6 is not input transformer. Why this ratio ?

This topology can not be transformerless because of asymetry.
Only possibility is use asymetrically conected microphones with good asymetrical shielded cables.
Then you must use number of input transistors as square of transformer
ratio (9 transistors same as in BA71 as input, or power transistor as input)
Elegant way to do it is AC paralel (with DC resistors
and output and blocking input capacitors for each transistor).

Input of your transformerless amp will look like output of BA71.
Nice idea, much capacitors and DC power.
Have a nice day
xvlk
 
the UTC input transformers are #AR-535 and 1:1.58 is written on them.

Same transformer schematic is in both the ba31 and ba71 schems.

schematics show taps for 37.5/150/600 ohm : 1500 ohm.

The ba31 type amps come wired 600/1500 . I tried 150/1500 on the older ba31 type and the mics just over load the preamps.

I have never tried higher ratio inputs into my bread boarded ba71 preamp . Maybe the RCA designers were hoping the old stock of
transformers would be close enough. Maybe when they found out
it was not good they decided to not produce the real ba71?

I have long wondered why this better preamp was replaced with the
renamed ba31 which was not very good IMO.

cheers,
Lance
 
[quote author="hitchhiker"]the UTC input transformers are #AR-535 and 1:1.58 is written on them.

Same transformer schematic is in both the ba31 and ba71 schems.

schematics show taps for 37.5/150/600 ohm : 1500 ohm.

The ba31 type amps come wired 600/1500 . I tried 150/1500 on the older ba31 type and the mics just over load the preamps.

I have never tried higher ratio inputs into my bread boarded ba71 preamp . Maybe the RCA designers were hoping the old stock of
transformers would be close enough. Maybe when they found out
it was not good they decided to not produce the real ba71?

I have long wondered why this better preamp was replaced with the
renamed ba31 which was not very good IMO.

cheers,
Lance[/quote]

Yes, sqrt(1500/600) is 1.58 . It is true for tapped transformer.
But for dynamic and ribbons you must transformer rewire to
200 ohm input and then go to the 1:3 optimal ratio for small Si bjt transistors.
{for condenser mics you do not need low-noise preamp, there are
much more interesting distortion than noise}

Transformers with taps never can be optimal in the noise figure.
With one except: transformer with astatic section wound {on the
L core with two coils} where input impedance is changed 1:4
{50 to 200 ohms). Your 600 to 150 is the same. You uses all copper {Cu} on the transformer.
To have some part of wounding unconnected {on the standard tapping} is not noise-optimal. {you are not using all transformer window for Cu}

There are other methods to make tapped transformer with higher
variability of input tapping impedance and no worse noise figure.
But they must use disc-wounding {like on some HV power transformers}
and labor work of wounding is high. {and parasitic capacitance, proper shielding....} They are not used as audio inputs.

xvlk
 
wow thats interesting the only gremaniums that my ba71's and ba 72's have the the 2n404 current source. I like how mine sound. Ba31's were a combination of a ba 71 and a powersupply and the ba 41's were the same with a ba 72. You ba 71's must be an a revision the c revisions I have are all sillicon except for the 2n404. I'll see if I can get you a copy of my schematics. the ba71c and the ba72 are electricily equal with kenyon transformer of the same ratio but in a different chassie


Wil

Wilebee
 
hi xvlk, that's very interesting info, good to learn.

I double checked my cross for 2n1010 transistor #1 at NTE and it is
according to them a ge nte103a the others are nte102.

http://nte01.nteinc.com/nte/NTExRefSemiProd.nsf/$$Search?OpenForm

hi wilebee, I have ba71 ,ba71b, & ba71c here and all are same circuit really.

This manual 'ba71A' for the circuit with all ge looks older than the manual I have recently found for the ba71c that you mention.

Lance
 
the Ba71c's I have 4 transistors I'll find the sehematic and scan it and sent it to you. I really like these pres on drums they have a realy thick bottem end

Wil

Wilebee
 
Wil, If you try out this 6 ge transistor version in an one of your modules
I don't think you'll be dissapointed. It's not as low mid and low end heavy
but the hi mids and hi end are similar to the TG console sound (Abbey Raoad etc)and also the Pink Floyd sound on Dark side of the moon.I've tried running mixes padded down through my various vintage amps and these were one of my faves.

I think if we could get to top notch transformers that fit these would be something for a Lab DIY .

Lance
 
I agree I could get Bud Pervine off Onetics in bothell Wa to look into them I'm just down in seattle

Wil


Wilebee
 
a transformer company. they wind the output transformers for my former employer THD Electronics but he is also winding input transformers for audisar's presonus upgrade kit. he is real sharp

Wil

Wilebee
 
Back
Top