Dual/Quad 1081 Mic Preamp

[alexc]

The short version is, that it sounds to me like you want to do :

xlr mic input connector
phase switch
pad switch
phantom switch

mic transformer

stepped rotary make-before-break single deck 12 pos 'gain' control  in the -ve leg of the 'discrete op amp'
ba338 discrete amp module as non-inverting opamp config and +24V unipolar supply

10K audio pot "level' control as voltage divider between stages

b340 line amp module and +24V unipolar supply
output transformer
xlr connector


That would represent a reasonable '1081 mic pre' implementation. 

It varies only in the 'gain' control simplification.


Quite reasonable to do and not too hard!

 

[justinheronmusic]

The short version is, that it sounds to me like you want to do :

Yes, to narrow it down further.

- xlr mic input connector

- phantom switch

- mic transformer

- stepped rotary make-before-break single deck 12 pos 'gain' control  in the -ve leg of the 'discrete op amp'

- BA338 discrete amp module as non-inverting opamp config and +24V unipolar supply

- 10K audio pot "level' control as voltage divider between stages

- b340 line amp module and +24V unipolar supply

- output transformer

- xlr connector

To be honest, I never use phase and pad, so for me those things would be excess. Everything else sounds right.

 

[justinheronmusic]

This appears to be the output section of the original 1081 schematic. I have outlined the parts that I don't understand. Otherwise, minus the input and gain switching, is this a good representation of what I am looking to do?

Pad 21 is the B+
I can't make out the number but below pad 26, there is a pad for unbalanced output.

Pad 23 and 24 are my balanced output

Then pad 25 is my B-

What are the other pads?

I can probably do away with the "TO EQ SW". I don't understand what "Fader Wet" and "Gain Boost" is, but I am assuming I can replace that section with a 10k pot. I am just not sure what parts to do away with in that section.

 

[alexc]

The pads 18 (send), 19 (return)  are for a potentiometer - 'output level'. Likely a 10K log pot. Fader wet is the 'faded' signal return  ie. the wiper of the pot

The gain boost pad allows for connection  of an additional resistor to the feedback network of the  b340 to increase gain. Its not required but can be used if desired. I forget how it connected - need to do some more checking.

pad 21 is B+ 24V, pad 25 is B- ie.  ground

pad 22 is unbalanced out

pads 27, 26 are for power filtering - not required

 

[Rocinante]

Alex. That was awesome.
Iirc the 'sensitivity' switch is a complicated affair.
I now remember why i went running for the hills when playing around with ba338 ba340.
Still it could be a cool project.

 

[alexc]

No doubt at all!

It's amazing to think about how a couple of relatively simple amp modules and some good transformers can truly be the heart of the most desired bits of kit there ever was. ANyware.

I like Ed Anderson traffos - I'd be keen to try his mic traffo in a build. And for outputs, well one could go any which way, really!

Myself, I use Haufe traffos *alot*.  Mostly because I like to do 'reclaimed parts' build where ever I can. It amuses me

My ba338 front end is using an expensive Haufe from our German friends   

The back end is another Haufe, from a Neuman fader module.  Pretty impressive performance generally speaking, where they are properly deployed.

As always, I'll be interested in the thd measures for the various modules. And comparing with other similar modules.

I think the JLH circuit modules, class A and class AB are super interesting. Some are quite Neve-ey and undoubtably British in their quality and ..  conciseness!

Doing some 'channel strips' with pre+eq+comp  with a  10Wrms 8ohms capable power amp  on board    is something I'll do next.

I just have to complete a couple of summing boxes, a stereo eq and a big tube power supply ..  first. 

 

[justinheronmusic]

The pads 18 (send), 19 (return)  are for a potentiometer - 'output level'. Likely a 10K log pot. Fader wet is the 'faded' signal return  ie. the wiper of the pot

The gain boost pad allows for connection  of an additional resistor to the feedback network of the  b340 to increase gain. Its not required but can be used if desired. I forget how it connected - need to do some more checking.

pad 21 is B+ 24V, pad 25 is B- ie.  ground

pad 22 is unbalanced out

pads 27, 26 are for power filtering - not required

Ok, so I've made some notes to the schematic. I think I am getting closer. I can borrow the EZ1290 input and +48v phantom power part of the schematic. I've attached some more notes and whatnot. Feel free to chime in. Things I need to learn:

- How to build in the 10k pot and the 12pos stepped pot into my new schematic

- What in the output section can I do away with and how do I do away with it? i.e. What is absolutely required after the BA340 Op Amp before it hits the output transformer? What is absolutely required before and after the 12pos switch?

- What switch do I go with? I've been looking at switches on Audio Maintenance, will single gang work?

 

[justinheronmusic]

Some more markings for the input section; using cropping to kind of piece together an idea of what the final schematic will look like.

 

[alexc]

The Output Level' pot goes across pins 18, 19.

The potentiometer is wired up with the fader connnected to one end, as a 2 terminal device.

The 12 pos attenuator can be a simple 1 deck affair - best to have 'make-before-break' to help avoid pops when switching it.
It is done to be like a potentiometer wired up as a 'two terminal device'.  ie. fader connected to one end

You would want a 'reverse log law' as you are wiring it into the -ve leg of the ba338 as per  JLM or api in the 'modern inverting opamp' style.

Use an excel spreadsheet to calculate the steps and resistance values.

Before the attenuator rotary switch is simply the mic transformer secondary with any terminating network across it.

That depends on your choice of transformer. I would think that a modern rework of the 10468 mic traffo of the original product would use the same network as on the original 1081 schematic.

It just optimises the response of the transformer into the ba338  to avoid 'ringing' and whatnot. Called a 'zobel network'


The primary side of the mic transformer is fine to take from say a 1272 or 1290. 

ie. xlr connector, a pair of phantom resistors 6K81 1% at least in accuracy, a small reservoir cap and a switch, which needs a single pole.

Sometimes a good idea to wire in a LED too. Should be able to see how from existing mic pre circuits.


After the b340 class AB line amp module is needed a cap to block any dc offset from the output transformer primary. It's shown on the schematic.

 

[justinheronmusic]

The 12 pos attenuator can be a simple 1 deck affair - best to have 'make-before-break' to help avoid pops when switching it.
It is done to be like a potentiometer wired up as a 'two terminal device'.  ie. fader connected to one end

You would want a 'reverse log law' as you are wiring it into the -ve leg of the ba338 as per  JLM or api in the 'modern inverting opamp' style.

Thank you for being patient with me and helping me to understand the different aspects of the circuit. A few stupid questions -

- for the 10k output, I need 1st lug and second lug, but the 3rd lug I don't really need a pad for on the PCB? Or would I have the 3rd lug go to ground on the PCB?

- At this point I am pretty close to being ready to do the schematic. Doing the schematic for the 12pos switch is still intimidating me like crazy. Is there an Eagle LBR file template for that sort of thing or do I just have to draw it into the schematic like everything else?

 

[alexc]

No probs

A potentiometer has 3 terminals :  the 'start' (of resistance track), the 'wiper' (variable resistance) and the 'end' (of resistance track).

You can use all 3 terminals, with the 'start' terminal as the 'hot' signal, the 'wiper' terminal as the 'attenuated' signal and the 'end' terminal often goes to ground.

You can also use it as a 2 terminal device, where the 'wiper' terminal can be connected to either the 'start' or 'end' terminals.

In the case of the 'Output Level' potentiometer before the b340 line amp module, you want to connect a pot, probably 10K log taper law (not reverse!) as a 3 terminal device.

ie. pot  'Start' terminal goes to pad 18, 'Wiper' terminal goes to pad 19, 'End' terminal goes to audio ground.

So that is operating as a 'shunt-to-ground' fader across the output of the ba338 and providing attenuated signal to the b340.


Hope that clears it up for you.

The 'Gain' rotary switch, in front of the ba338 is a little different - here it operates as a 2-terminal device : the variable resistance in the '-ve' leg of an opamp wired into a non-inverting variable gain stage.

So you would need a 10Kohm pot, reverse log taper 'law' with the 'wiper' and 'end' terminals connected together and going to ground thru a cap.

Check the api style mic pre doa '-ve' leg and you'll see what I mean.

Furthermore, you want to emulate that pot as a rotary switch of 12 steps : so you need to wire up a single deck, make-before-break switch, say a Grayhill to emulate the 10Kohm reverse log taper pot.

AGain, you can see exactly how this is done from an api mic pre doa.

Which is why I was asking 'how historical do you want to go' ?  You are basically doing an api mic pre using a ba338 module as a 'discrete op amp non-inverting gain stage' with an additional b340 class AB line amp + traffo backend.

As far as eagle footprints and lib components and hat not, you are essentially asking if anyone has a api mic pre project done in eagle that you could cut-n-paste.

Probably there is - spome more research and asking nicely I suppose would give you an answer. I would say however, if you are going to do that, then why not simply get an already fabricated pcb?

If you are more of a 'recording' guy than a 'electronics' guy, I would recommend getting a good kit with pcb+parts. There's heaps of them available.

You still have the thrill of diy, at a reasonable price, and wind up with a really nice product. You'll probably wind up with a more 'focussed' end result too  ie. a distinctive 'neve' sound or a distinctive 'api' sound.  ...  rather than something in-between, which may wind up being very good performer but without any real uniqueness ...  kinda.

However, if you are already diy fevered to the max, go for it and fab your own pcb!

But anyways, it's all good.

 

[justinheronmusic]

The 'Gain' rotary switch, in front of the ba338 is a little different - here it operates as a 2-terminal device : the variable resistance in the '-ve' leg of an opamp wired into a non-inverting variable gain stage.

So you would need a 10Kohm pot, reverse log taper 'law' with the 'wiper' and 'end' terminals connected together and going to ground thru a cap.

Check the api style mic pre doa '-ve' leg and you'll see what I mean.

Furthermore, you want to emulate that pot as a rotary switch of 12 steps : so you need to wire up a single deck, make-before-break switch, say a Grayhill to emulate the 10Kohm reverse log taper pot.

Ok, so I found this website that is a Stepped Attenuator Resistor Calculator. If I am using a 12 step reverse log taper, then step 12 on here would be step one of the pot, correct? Will this calculator work?

http://rssconsultancy.co.uk/atten.html

I think I found a Grayhill 71 series LBR footprint for Eagle, I just have to check it vs the data sheet to make sure. If not I will just create one - I watched a video on it this morning.

Side note: I understand what you are saying about kits. There is something about me that you should know; I'm very stubborn. I did a kit once and it was over too quickly. It didn't give me the heartache I was looking for. Most of these kits only take a few hours. I need a project to outweigh me, destroy me, make me feel like a failure so that I can rise up from the ashes and conquer it. If it's all laid out, I won't get the chance to drag myself through the mud. That's the best part. For me, DIY is about wrestling my demons of self doubt, inadequacy, and self loathing. With each completed build, I overcome. Maybe one day that will change, but for now, I am out for blood. Right now, I need it to be hard for me. Harder than my last build. This 1081 project is all I could think of that I could actually use (I am obsessed with the 1081 sound). I don't even care how many I make, I'd be happy with two. The goal isn't in the completed build for me, the goal is in the process.

 

[ruffrecords]

You need a stepped series resistor rather than a pot. Pot calculators will give  you the right values for a potential divider but you want a series resistor which is quite different. The gain at any step is given by (Rf+Rpot)/Rpot where:

Rf is the feedback resistor in the BA338 and Rpot is the value of the switched resistor. This allows you to calculate the resistor value for each gain setting. You can then wire these individually or convert them to a ladder.

Cheers

Ian

 

[justinheronmusic]

The 'Gain' rotary switch, in front of the ba338 is a little different - here it operates as a 2-terminal device : the variable resistance in the '-ve' leg of an opamp wired into a non-inverting variable gain stage.

So you would need a 10Kohm pot, reverse log taper 'law' with the 'wiper' and 'end' terminals connected together and going to ground thru a cap.

Check the api style mic pre doa '-ve' leg and you'll see what I mean.

Furthermore, you want to emulate that pot as a rotary switch of 12 steps : so you need to wire up a single deck, make-before-break switch, say a Grayhill to emulate the 10Kohm reverse log taper pot.

But anyways, it's all good.

Ok, forgive me for not seeing this before. I haven't looked at the JLM circuit in a while, but after looking at it I feel stupid.
Is all the calculations for my gain switch already done for me in that schematic? I will happily learn to calculate my own gain switch, but it appears he may have done it already. See attached schematic:

Again, thank you guys so much for helping me to understand. It's very humbling and I appreciate all of your knowledge. You guys know far more than I do.

 

[alexc]

The goal isn't in the completed build for me, the goal is in the process.

Well said!


You can use the 'jlm 1081' attenuator as is    That will work nicely.

....

That 10K resistor you circled isn't the feedback resistor, its the 'load resistance' - gives the ba338 gain block an easy impedance to work into  ie. a 10K load is very 'light' - requires very little current to drive to desired voltages.

In a real mic pre, that would generally be replaced by the 'Output Level' pot of similar value.

.....

In the 'jlm 1081',  Rfb (feedback resistor) is one section of the rotary gain switch - the part enclosed by the circuit 'output' and the '-ve leg' of the 'opamp'

It sets the gain of the non-inverting op amp stage with the relation 

gain = Vout/Vin = 1 + Rfb/Rin

where Rin is the resistance to ac ground of the resistor in the '-ve leg' of the opamp

In the 'jlm 1081', Rin  is the other part of the rotary gain switch

As shown in the 'jlm 1081', drawing, 

Rin (max)  = 22+18+30+56+100+200+360+750+1K8 -> 3336 ohms   
Rin (min)  = 22 ohms

and  Rfb = 20K+11K+6K8+3K5+2K2 -> 43.5K

gain (min) = 1+ 43.5K/3.336K -> 14.0    which can be expressed in dB  as 20log(gain) -> +22.9dB

gain (max) = 1+ 43.5K/0.022K -> x1978  or +65.9dB    -- seems a little high! 

Should be something like x50 or x100 for the gain (max)


**  I may have the details a little wrong with the ladder  **  -  Ian Ruffrecords can correct

....


In the api circuit  the '-ve leg' usually has 10K potentiometer + 1K fixed resistance to (ac) ground.  The Rfb is usually something like 50K or so.

So the gain (non inverting)  at max = 1+ 50K/1K  (ie. pot resistance is at zero end of travel) -> x51 -> +34.2dB

and gain(non inverting) at min = 1+ 50K/(1K +10K) -> x5.5 -> +14.8dB

so the range of gain control is +14.8dB .. +34.2dB

Of course, there is the mic traffo gain as well  .....  often the ratio is 1:5 step up or +14dB, giving a total gain for the 'mic pre front end' of +28.8dB (min pot setting) to +48.2dB (max pot setting)

..

That's the basic idea. 50dB or so is the typical gain for a front end.  The 'back end' b340 probably adds some more 'voltage gain', maybe 24dB or so, as well as the 'current gain' required for a line stage with decent load-driving capability (current).

...

In the 'jlm 1081' circuit, the rotary attenuator varies both the Rin and Rfb together

As Ian stated, it's a 'ladder attenuator'. You can do some searching on that, but I think it's fair to say the JLM circuit is as good as any

 

[Rocinante]

Justin there is a grayhill 71 series .lib footprint for eagle in the .lib file i sent you via email.

 

[justinheronmusic]

Got it! Thanks, Rocinante! 

I think I am ready to start on the schematic. Not sure how long it will take, and it may take me a few tries. Wish me luck! Talk to you guys when I've got some stuff to share. Thank you for all your contributions.

 

[Whoops]

It's nice that you didn't go the OPamp with pins route.  Spreading the components is better for heat dissipation and troubleshooting, also not having unnecessary pins and sockets avoids also unnecessary intermittent connections.

As for the input switching you really should add a Phase switch and PAD switch, you will need that in the furture and you will regret not having them also we are talking about just 2 switches and 3 resistors more, a marginal cost for something that is really useful. It's also really easy to implement.

Check the JLM Go between schematic for the Phase switch and PAD switch:
http://www.jlmaudio.com/GO%20Between%20PLUS%20Schematic.pdf

 

[justinheronmusic]

You can use the 'jlm 1081' attenuator as is    That will work nicely.

So here is what I have so far - I just wanted to check in to see if I have wired up the Grayhill switch assembly properly, and then the output potentiometer following the O/P wire on the BA338.

The JLM interpretation is pretty different than the original schematic. I am following the original, but trying to work in the JLM attenuator, I've also subbed a few parts per Tommytones recommendation.

If you have any other guidance, please let me know.

 

[alexc]

Just having a looksy now ...

Referring to your designations  for the ba338 amp module and jlm mic pre  scheme

- C5 470uF 25V is wrong way around. The electro + should go to pin 2
- the top junction of C4 100p and R108 2K2 should not connect to the back of D2 NTE519
- pot P1 10K pin 3 should go to rail B- (ground)
- feedback cap-to-ground C7 is wrong way around -ve leg should go to rail B- (ground)

And

-  One can use a non-polar electrolytic  in your C7 position.  If you can get one use it, otherwise a polarised will do
-  One can use a non-polarised electrolytic in your C5 position, although a polarised will do

and also

JLM micv pre schematic uses the ba438 module, the later iteration of the ba338 and shows no 'Boost' pin 4  that the ba338 and your version is showing.

I would need to check how that is handled in this hybrid implementation - I think it's left floating but can be connected up for various modified outcomes.

Other than that, looks good

...

These make very nice sum amps and can be used as 'virtual earth' (also known as 'current' summing) or 'passive' (also known as 'voltage' summing) mixing.

It's pretty amazing the performance one can get with so few active devices, and even with the less-than-perfect build values inherent in my efforts

> 100dB of signal-hum margin, THD can had in the 0.002 % range with a boat load of headroom to boot.

Add it to your 'to do' list  to make a passive sum bus, not too many channels, to keep it simple -  and hook it up to this mic pre.

Awe inspiring sound. The 'minimal' approach really shines through.

My 'opinion' is that too many opamps in-a-row , like in most mixers, really does detract from the subjective sound quality.

In my music setup, I use a small virtual earth mixer, built around these modules, as the master 'mixer of mixers'.

My diy stuff sounds far and away mixed thru a neve styled sum bus than thru the digital console alone. Much better.

That's why I like building sum bus units    Can be done fairly cheap with heavy weight sound.

I feed the 'master sum bus'  my DAW  'ITB mix' stems,    the Yamaha digital desk stems (for keyboards, edrums and fx)  and feed in yet another mixer (y.a.m    with the sum of all my analog diy channels (for monitoring).

I feel I get the best of all worlds