MAVIS preamp

[mutetourettes]

Hi - first post here so go easy on me!

I have a channel from a MAVIS desk circa 1971/2, a Bill Hough design, class A mic preamp (so I'm told).. I'm trying to sort out a trace of the circuit and if possible, actually understand what he's done..

first up, it's using three TI BC384L transistors,which are, as far as i can establish, NPN BJTs in TO92 case. The scraps of indication i could find from the interwebs had this as a to92b with "american" pinout of EBC (viewed from the bottom, flat side up) but drawing the circuit this made no sense (to me) and so I reckon they must be "Japanese" pinout ECB... can anyone confirm this?

with any luck i've attached my first trace of the circuit topology..

can anyone educate me about this circuit, maybe tell me some keywords to look for? I think Q1 and Q2 are a voltage amplifier with negative feedback of bias (and signal?)

and Q3 a common collector current gain stage with emitter load and bias-only feedback to the voltage stage? signal in that feedback loop goes through C3 to ground, right? why doesn't input signalalso just go to ground this way?

i don't really get how the two feedback loops interact around the base/emitter of Q1..

ok so you can tell i'm clueless, just trying to match what i see here to bits and bobs in textbooks and webpages...  can anyone give a few pointers so i can learn? Is this a standard or unusual design?

I'll add values when i've got more brain time to devote to colours... especially tant cap colours/polarity...

the input transformer is a 1:3.2 beyerdynamic..

cheers
M

 

[mutetourettes]

ah ok, thanks for moving this to the right forum!

 

[RuudNL]

I don't think C6 is 10 nF...
(By the way: what are the other values?)

 

[radardoug]

D1,D2,D4 look wrong to me, are you sure they are not caps?

 

[PRR]

Topology is standard 3-transistor mike or phono preamp.

Resistor values would aid understanding.

"Diodes" are surely caps.

C6 at 10nFd does seem mighty small; 10uFd would be screamingly acceptable.

> how the two feedback loops interact around the base/emitter of Q1..

R10 R8||R11 set AC/Audio gain, with some effect on DC conditions.

Short R11 to kill that NFB loop's DC effect.

Now Q1 Emitter is at zero.

If R2 is not-huge, then Q1 Base is at R3 voltage. R3 voltage is Q2 current times R3. Since Q1 Emitter is at zero, Q1 tries to force Q2 to force R3 to put Q1 Base at 0.6V. If design is happy, Q2 current is now solidly fixed (some tempco) and R7 can be selected for the specified supply voltage. With reasonable choice of values plus Q3 buffering any heavy load, gain is far over 1,000.

Un-short R11. DC conditions change a little, some change of R3 will restore them. Closed-loop gain is now the R10/R11 ratio, typically very low (perhaps 10). Add C3 and R8, closed-loop gain can be higher, up to open-loop gain.

 

[mutetourettes]

ok, thanks everyone. the 'diodes' are glassy looking things (of different sizes) with a red line at one end, so i thought diodes, but happy enough to be told otherwise, i guess i've seen caps in see-through resin before..

the 10uF has slipped through, just 5spice's default value, i intended to post no values at that point!

ok folks i'll update the thread with as many values as i can work out, and PRR thanks for that explanation, I'll go over it until I understand fully!

glad to hear it's a standard topology... increases my chances of working it out!

 

[mutetourettes]

Ok, I've got as many values as possible from purely visual inspection. note that the party is now joined by C7,8,9,10, all of which are metal film and unidentifiable in situ..
C1 and C6 are tantalum beads. C4 is a ceramic disk, others are electrolytic.

I'm not sure what the M marking on the trimmer R8 signifies.

On the previous drawing i forgot the 'Attenuation' knob.  I've also added R12, which i'm guessing is the welcome into the eq circuit.

so, pardon my ignorance, is Q1-Q2 voltage gain, and Q3 a buffer? - and does the load being on the emitter means this circuit is non-inverting?

 

[mutetourettes]

here's a pic of the unit, just for reference.  also further pictures of a similar console (from a 2004 thread here on group diy) here: http://members.chello.nl/m.vanoijen/index2.html

 

[mutetourettes]

component side of mic pre circuit

 

[mutetourettes]

trace side (a vertical flip of above)

 

[mutetourettes]

OK, found a datasheet that confirmed BC384L as TO92B and ECB pinout, along with a lot of other specs..

 

[PRR]

> 'diodes' are glassy looking things

Plastic. Polystyrene. VERY clean and stable. However soldering can damage them.

 

[mutetourettes]

fair enough, if it ain't broke don't fix it... so unless there's something very wrong i won't shift them, and their values will remain a matter for conjecture.

Plastic. Polystyrene. VERY clean and stable. However soldering can damage them.

 

[RuudNL]

Strange they used low quality carbon resistors (they are the worst I have ever used!) in an input stage.

 

[abbey road d enfer]

i don't really get how the two feedback loops interact around the base/emitter of Q1..

Q1 is subjected to both voltage FB via R10, 11, 8 and C3 and current FB via R4, 3 & 2.
This principle has been favored by some designers because  it provides some gain degeneration within the FB loop, which some deem desirable and reduces the input impedance , with the advantage that when the preamp input is unloaded, it is less noisy than when there is only pure voltage FB. This is quite debatable, because it is no without drawbacks.
BTW, I'm not convinced R2 is 5k8 (value not normalized, and seems too low - 56k would be more likely).

 

[mutetourettes]

ok - checked again under daylight..  good catch! R2 (the one nestled under the silver cap under the trimmer) is blue grey orange - 68k. damn, better check the others while the sun's out.

ok wow under really bright sunlight there's a fair bit clearer!

R1 and R12 I'd now call 300K
R4 I'd call 470
R5 I'd call 47K  -  big change!

will update schematic when I'm on the right pc...

 

[RuudNL]

I don't think R1 and R12 are 300 K. IMHO the first assumption that they are 100 K is correct.
All resistors are standard E-12 range values, so why would they choose a value outside the E-12 range for a non-critical value?

 

[mutetourettes]

ok - I'll go by your recommendation, my identification of brown or orange on that brown background is sketchy at best. So have you worked on one of these before? When did you experience the poor resistors?

and should i plan to replace them with something better?

 

[RuudNL]

I used those resistors once in one of my own projects when I was a student.
They are noisy and fragile. (They are made of a baked carbon/clay mix, better quality resistors contain a ceramic tube with a thin layer of carbon damped on and metal caps at the end.)
Looking at the schematic I personally think it is a bit strange to have a pot immediately after the (microphone) input transformer.
Usually the gain is controlled by changing the feedback in the first amplifier stage, not by attenuating the level at the microphone input.
A pot at this point may be noisy.

 

[abbey road d enfer]

Looking at the schematic I personally think it is a bit strange to have a pot immediately after the (microphone) input transformer.
Usually the gain is controlled by changing the feedback in the first amplifier stage, not by attenuating the level at the microphone input.
A pot at this point may be noisy.

Many prestigious preamps use a similar arrangement, Neve, Schlumberger, though with rotary switches, not pots... The idea is that the signal level will be almost constant there for a given output level, thus the S/N ratio will be more or less constant, as long as the impedance presented by the pot does not vary too much. If the S/N is good enough for low levels, it shoud be good enough for high levels too, shouldn't it? You may disagree (perhaps I would too!).