Neve 1272 and impedence question

Hi there i have a strange problem

i´ll begin with describing what i have built.
I have a ba283card and a 2 bank 23 pos elma switch. i have a fixed gainboost on the second stage a 470 ohms resistor instead of 1,5k.
then attenuate with the first bank of my swith from 4 db to 55 db after that i boost the first stage up to 27 ohms, with the second bank.

I also have a 604 ohms resistors on the output xlr to match my impedence of my computer.

Now when i plug in my guitar i get the right sound but when i try to plug in my synth the gain gets so high that everything gets red i have my volume on my synth almost set to zero , and on the pre amp the switch is at 4 db and the fader pot almost a zero and still i get the sound so high that it´s crazy i wonder what i´ve done wrong.......also o feel that the preamp is poor on bas frequencess it only goes down to 200 hz with the guitar and a microphone not to talk about the synth why why why why why

Shortgypsy,

First of all, Didn't Geoff Tanner already answer all of the the 1272 questions you had? He is overly generous for someone who's still in business with this stuff.

Let me try to answer:

1) Do you even have transformers hooked up to this 283 card? Your post makes no mention of it so I thought I'd ask. That would cause huge problems and poor sound quality.

2) The impedance to your computer input is most certainly not 600 ohms. The resistor is put there to properly load the LO1166 output transformer when it is connected to MUCH higher impedance inputs such as your computer.

3) Yes, the bass response is going to be horrible with gains that high, which is why people have said over and over that you can't boost a single 1272 past 50dB (including transformer step up). Neve didn't go past 91 ohms in the first (NV) stage, and 1.5k ohms in the second (AM) stage, UNLESS it was a talkback amp which used the values you list. Talk back amps are allowed to sound like shit. If you want to try and correct this loss of bass, you need to connect some really big 25V or 35V electrolytic caps from pins N and M to ground. All of the schematics show these caps as smaler values such as 470uF in the first stage and 1000uF in the second. As Geoff has told us many, many times you should increase these values to as larger as can fit i.e. 4700uF if you want to have decent bass response.

4) The 10468 input transformer should be loaded with approximately 5kohms on the secondary to function at its optimal freq response. On the first click of many Neve switches there is about 5kohms acting as a pad. This is how much attenuation some louder signals require.

5) My advice since you you have a 2 deck switch is to wire it like JLM audio does with the first bank as a proper 5k pad/loading and the second deck used to boost the first stage only up to a reasonable value. If the gain is then too much, reduce the boost resistor in the rear section back to where it belongs. you should not need more than 10 or 11 positions. The 23 position switch was used for a 1073 module which had three gain stages and settings for both mic and line inputs. You can't pull this off with just one 283 card. Also, make sure to add those huge caps I talked about! See here: http://www.jlmaudio.com/JLM1272hotrodmod.pdf

cracking reply - in laymans English - just downloaded it for future reference
what's worrying me is that I actually understood some of it

Oh, and I almost forgot. Have you properly adjusted the bias trimmer for the 2N3055 output transistor? An improper bias would contribute to many of the problems in sound quality you described. There is much information about this procedure if you look around. If in doubt, somewhere around 12 or 1 o'clock would be a rough estimate.

I think this would be the right place to find this out??? I have a 23 step Elma switch I am using for a 1272 type project with a 183 board. @ an estimated gain of 50dB, I am having trouble figuring out my resistor values. I would like to just use 22 of the steps and have step one be OFF. This should leave 2.27db of gain per step. I have tried one of the online calculators, but am still confused????

-Mike

So you want the gain to be from unity up to 50? Why not make the steps a little larger and acommodate line level inputs as well?

I can dig up the formulas for calculating the steps if you like. Give me a a while to do that.

[EDIT] What are you using for transformers? The step up ratios contribute to the total gain and ought to be factored into the switch calculations.

For input I'm using the 31267 in reverse. I only have the one mic input scheduled. Output x-former is the 1+1+1 from JLM audio...apparently very similar to the L01166. In the future, I plan to only use the 12 step switch, but I paid $50 each for the two 23 step 3 pole Elma's I have, so I want to put them to good use.

-Mike

Okay, this should get you started.

know the following:

- The BA183/283AV (fully stuffed) card has two gain stages. There's a preamp stage called NV and the output stage called AM.

- NV stage has 18dB gain without any boost resistor. To boost, tie a resistor from pin T to ground.

- AM stage has 11.5dB gain without any boost resistor. The JLM output trafo wired for 1:2 stepup should give you another 4-6 dB of gain depending on losses. Granted, let's call the AM stage basic gain 16dB for simplicity's sake. To boost, tie a resistor in series with a 25V electrolytic cap (100uF until you get up to around 28dB, then it becomes 220uF) from pin K to ground.

- Ground (aka B-) on a BA183/283AV card is pins A, J, and V tied together.

- The 31267 input transformer half windings in reverse gives you a 1:2 ratio for approximately 6dB step up.

When you add it all up the basic gain of your preamp is about 40dB. To get anything less requires that you attenuate the signal from the input transformer secondary. After that you start boosting. The N*eve 1073 and its relatives keep a fixed amount of boost in the AM stage while boosting the NV stage up to 28 dB before switching in a second NV stage (from a different card).

I will be back with formulas + schematics in a few minutes.

Have a look at the BA183/283AV schematic.

AM:

The gain is being set based on the feedback resistor, R5, and the shunt resistor, R4. As you can see, pins K and J allow you to parallel an external resistor with R4. Recall that the combined value of resistors in parallel must be less than the smallest one. By effectively decresing the value of the shunt resistor more signal is shunted and less of it becomes feedback. Less feedback brings us closer to open loop gain (the maximum amount of gain the amp is capable of).

So, to determine the gain of the AM stage for a given resistor value, use this formula:

N ohms = boost resistor || R4
gain = 20 log ( (R5 + N) / N )

I've written this like you'd type it in a calculator. || means "in parallel with" and / means "divided by".


NV:

The principle is the same. R17 is the feedback and R18||R19 (we can ignore C17) is the shunt. By putting a smaller resistor between pins T and V we have "boost resistor" || R18 || R19 and the combined shunt value is now less. Formula looks like this:

N ohms = boost resistor || 320.5 ohms (the combined value of R18||R19)
gain = 20 log ( (R17 + N) / N )


...so, start trying some values until you get what you want. You'll have to round to the nearest half dB in most cases, but this doesn't matter since the resistor tolerances throughout the circuit will equate to at least 5dB of error.

Here are some values to get you started:

AM (including the output trafo @4.5 dB boost):
1k5 = 20dB
470 = 25dB

NV:
330 = 23dB
120 = 28dB

Okay, now I will FINALLY answer your original question concerning attenuation reistor values.

First, have a look at the schematics for 1063 and 1073 to see how N*eve did it with a 3 deck, 23 pos switch. I include the 1063 because it provides a separate ladder network to get line level signals with the same transformer. Our discussion here will focus on the 1073, however.

Note that there is a position for OFF position in between -20 and +10. No boost is applied, neither of the input transformers taps are connected, and the BA283 input sees the dummy load of R19.

I should mention that the numbers on this sensitivity switch (i.e. -20dB) indicate the level of the signal coming in so the module knows how much boost/cut to apply to bring the level to 0dBm. For example, in the minus 50 position, the input passes straight through with no attenuation and the signal is then boosted by approximately 50dB.

Now, focusing on positions -20 through -50 (the mic pre section using only one NV stage) you'll see that the first two decks compose a voltage divider, attenuating the input, and the third deck boosts the NV stage. You want the total impedance of the voltage divider in any position to be about 5k, since this is what the input transformer would like to be loaded with. You could use a 5k log pot shunting to ground and it would have the exact same effect.

As for calculating the voltage divider values for your preamp, I would do it like this:

1) basic gain minus gain you want = loss in dB
2) Antilog (loss dB/20) --> on the calculator it's 10^(loss dB/20)
3) multiply the result by total impedance (5kohms in this case) and you get the value of the lower resistor (the one connected from the second deck to ground).
4) Now choose an actual resistor closest in value to what you got and subtract it from 5k. This gives you the value of the upper resistor (the one connected from the first deck to the second deck.

As an example, let's say you want the first step to be 3dB gain. 40 - 3 = 37dB loss, or -32dB.
10^(-37/20) = .0141235375446 x 5000 = 70.62 ohms (<-lower resistor)
5000 - 70.62 = 4.93k ohms (<-upper resistor)


That's it. Questions??? I will try to help.

Thanks Jens!! I have read through your responses (many times to get it through my thick skull). I am starting to understand more. Though with more answers come more questions. Good for now though. Tomorrow I'll work on the calculations. Again thanks for the input!!

BTW should this scale in the first stage, then, keep the first stage at full power and then scale in the second stage??? (thus 11 steps for each?)

-Mike

[quote author="Aryuserius?"]BTW should this scale in the first stage, then, keep the first stage at full power and then scale in the second stage??? (thus 11 steps for each?)[/quote]
Sure, if that's how you want to go about it. Geoff Tanner always recommends distributing the gain equally among the two stages, and the Seventh Circle N72 works this way. Myself, I just imitate what N*eve did with their input modules, but whatever.

I know my response could've been much more concise and spoon fed, but this way you can really understand the circuit and come to your own conclusions. If you teach a man to fish...

My version of a 1272 uses a 2x6 Lorlin for gain adjustment from 40-65dB in 5dB steps. The first pole adjusts the preamp stage gain (pin T to ground) while the second pole adjusts the output stage gain (pin K to ground by way of a 100uF/25v electrolytic):

Pole 1
"40dB" (no connection)
"45dB" 330ohm
"50dB" 120ohm
"55dB" 56ohm
"60dB" 56ohm
"65dB" 27ohm

Pole 2
"40dB" thru "55dB" 1.2k
"60dB" and "65dB" 470ohm

In truth the "40dB" through "65dB" labels are not totally accurate, but they're close enough for me, and they are 5dB apart, so I'm cool with labeling them as such. Actual gain is a couple-ish dB higher.

As for the missing input pad required for gain values between 20dB and 35dB, I use a 5k audio taper pot between the secondary positive of the input transformer and the input to the preamp stage of the 183/283 card (pin U). Wiper goes to input of the 183/283 card. This effectively duplicates the function of the pad on the original without the need for a more-expensive 3-pole gain switch, and it also eliminates the need for a separate fader pot, since the pad itself is a continuously-variable pot. I got this idea from Joe Malone's 1272 SimpleMod schematic on his site.

Ditto what was said earlier about increasing the value of the power caps to maintain good LF response. I've been using 4700uF/25v on the output stage (pin M) and a 1500uF/25v on the preamp stage (pin N). From a practical perspective this design isn't requiring anything of the preamp stage beyond what Neve did up to 50dB (more than enough for drums and other loud sources typical for tracking with Neves), and the additional gain required of the higher settings is split between both stages in 5dB steps.

In practical use this circuit sounds excellent, even at higher gains, with good LF response (though different, still, from my original 1084's), and work well for most stuff. I use mine mostly for tracking drums, so I rarely ever need the higher gain settings anyway.

Definitely worth having a couple of them around.

JC

I admit, in most cases I'll just set the gain fixed in both stages and control level via a 5k pot which shunts the trafo secondary to ground. The wiper is the input to the amp card.

Another way to enhance LF response with high gains is change the NV feedback cap, C17, from its current 400uF (or more likely 470uF) to something bigger like 2200uF.

The wiper is the input to the amp card.

Click to expand...

Oops..... that's right. Must be the glue :wink:

(I edited my previous post to fix this mistake... thanks)

JC

hi,
great info going on here....

question:

if you were to use the 1272 without the output transformer(taking the unbalaced signal from point F), what resistor would you use to "emulate" the transformer ? my guess would be the DC resistence of the output primary's?

thnx

No resistor, you need a big inductor. N*eve T1310 or something equivalent. se here

I have a question about the gainsetting on the 1272. Im trying to decide if to use the JLM simple mod, or the Hotrod mod.

Have i understood correctly that the simplemod has the second gain stage gain constant? Would this affect the sound quality of the signal any compared to using the hotrod method of using a 2deck 12 position switch for setting the gain?

I kinda like the simplicity of the simple mod, but i dont want to compromise sound quality.

Any thoughts?
/Jonas

At a certain point you've got to use your ears. Try it and see. IMHO, boosting the gain of the first (NV) stage much past 30dB is most definately a compormise. Both JLM versions go higher than that. Others would disagree.

Even though they go higher than that, im pretty sure i wont use that much gain most of the time. My api 312 channels dont have much more gain than that, and they work fine on most sources. The added gain is just a plus for me even if i might not even use it.

So if we compare the two jlm versions at normal, intended gain levels for the 1272, would there in theory be any shortcomings sonically, using the simple mod versus the hotrod mod.

I might try both, and see for myself, but any input from other people who has tried this would be nice aswell.

/Jonas