Can anyone help me identify the function of trimmer R21?
Thanks!

PS: in the frame pin 1,2,3,4 are shorted on every module. 1&2 is chassis. 2&3 is 0v. I guess it would be better to keep them seperat and connect 0v and chassis at one point only? There is phantom power as well.

Looks like DC feedback around the whole ring of three. Changing it will change the gain but it will also make a change in the DC operating points of T1, T2, and T3.

Chassis and signal grounds should be kept separate except at one point in the console where they should be bonded together.
--scott

Thanks Scott, the effect on the overall gain of R21 seems minimal. Should i just keep it in the middle?

So I will keep chassis and 0V separate on the frame holding the modules. The PSU is external and I would then connect the two within the PSU (feeding 0V, chassis and +24 V to the units frame) Sounds good? The Phantom power is taken from the 24v line and has no separate ground. I guess I should just connect all pin1 input XLRs to chassis to take care of that?

Adjust it for lowest distortion if you can. Or set it so that the voltage on the emitter of T3 is whatever the manual says it's supposed to be. It is affecting the operating point of all three transistors.

Are you trying to rack up a module out of a gutted console? If so you might want to consider real 48V phantom power since 24V won't drive a lot of microphones.
--scott

R13 varies the DC balance …. adjust for similar clipping point on the top and bottom of the output waveform at emitter T3.

As far as the grounding goes … can’t comment on the exact whys and wherefores without physically examining how the modules, chassis, and backplane go together, so I’d leave it stock. I’m assuming T6 was included to prevent current flowing from the supply unless pins 3/4 are firmly grounded … like some protection for module hot-swapping?

Also if R53 and R54 are indeed the same value, then the two highest-gain settings of S1 are the same? Maybe a misprint?

No, I think it's a dual-gang switch. The top resistor in the divider is set by that gang and the bottom resistor in the divider is set by the gang to the left of it. On the one to the left, the resistor values are different on the top two steps. I really hate that design with the divider in the feedback loop because the frequency response and distortion change so much as you adjust the gain. At least that 47uF cap is big enough that the low end won't drop out, although I might move that up a lot more.

What is this thing? Was it intended as a preamp or a make-up module? And what's inside the mystery box?
--scott

It's actually a three-deck switch with two poles per deck (six poles total), with one pole unused. The lowest three gain settings introduce an attenuator formed by the two 499 ohm resistors (both labeled as R45?!) in series with the primary, and the additional secondary loading of R48/49/50. But as I read the schematic and the switch-layout diagram at the bottom, the two highest-gain positions have the four sections on the left (labeled a, b, c, and f) all shorted together, leaving section e (the feedback series resistor) switching between R53 and R54, which are shown as the same value. Why would they switch between two identical resistor values when they could simply short those two positions together like everywhere else? Still thinking it's a misprint . .

This sure looks like a mic preamp to me; assuming V1 is a hybrid module of some ilk. As far as the design of the gain-ranging goes, it's not too dissimilar to many of the Neve modules that everyone swoons over. By switching both the series and shunt parts of the feedback ladder, the impedance can be kept lower at the high-gain settings (for better noise performance), but without causing excessive load on the output stage at lower-gain settings . . . the latter being limited by the standing current through the follower T3 (one of the several circuit parameters adjusted by R21). I'd think changing the value of C20 to be wholly moot given the liberal sprinkling of electrolytics throughout . . . if you're really concerned about the accumulation of LF poles then the whole thing is a non-starter, like most single-supply designs. They're even using electrolytics for filtering in the high-pass CLC, so it's probably better to come to peace with it and call it "mojo".

Agreed, but i don't like it on those either. I don't understand the appeal of those Neve modules.

Yes. It's a reasonable approach.

I am tired of so much mojo! And mojo that changes when you adjust the gains drives me up the wall!

I bypassed all those tantalums in a neve channel strip once back in the eighties and nearly got fired. It sounded so much better.
--scott

Sorry guys, i meantioned the wrong trimmer :(
I meant the one in red R13. Sorry for the confusion. Still kirkus mentions R21changing the balance. But i think thats what R31 (labeled symm.) does. I dont really undestand what R21 does…
Still, i guess kludges first reply was talking sbout R13?

And yes, its a mic pre. When talking about phantom i did not mean that im running 24v, but that the 48v are created from the 24v line and therefore have no seperat ground. I attached some info on the unit creating the 48v. Mine has a slightly different model number (KA added) but this is the closest i could find.

And yes, these are a bunch of racked up sitral modules. But i think this racking job was not done properly. The whole frame had no chassis connection at all, pin1 on all output XLRs has no connection whatsoever and some of the solder joints look very dubious. I doubt that whoever racked these knew what he was doing.

R21 is part of the filter network. If it says "symmetry" I might suspect it adjusts the filter symmetry. R13 is in the DC feedback loop of the ring of three.
--scott

I too got confused on the reference . . . my comments this far have been about R13. I think I looked at R13 on the diagram because it was circled, but typed R21 since I had the reply window open on another tab and just looked at your text. I'll correct on the original post.

On R13 . . . the circuit is pretty straightforward; the only thing (somewhat) less common is the separate feedback paths for the AC and DC. There are tons of tube preamp circuits with the same topology (two plate-loaded sections plus cathode follower and global NFB); virtually always with no DC feedback at all, and that's also true for many early transistor circuits. In this case the input transistor is lightly biased (0.28mA, to reduce input current noise) and lightly loaded (39K resistor, to increase open-loop voltage gain). This leaves the DC parameters suceptible to drift from variations in the beta (hFE) and Vbe in the transistors due both to sample-to-sample variation as well as temperature. The single 19V supply rail doesn't leave a lot left for the bias voltages to wonk around on their own before they'll negatively affect headroom. Thus, the addition of a DC feedback loop will likely stabilize it enough for thermal drift, leaving the trimpot mainly for differences in transistor manufacturing.

The basic topology introduces a DC shift of some 11 volts input-to output, and R13 varies the how much of that is fed back to the input transistor's emitter via the divider with R9 and R5. This has the side-effect of altering the closed-loop DC gain in addition to the offset/balance, but that's unlikely to be of any consequence. I'd adjust it for about 6 volts on the collector of T1 . . . or 11.3-ish on the emitter of T3 . . . or the highest AC signal headroom (before clipping) on T3 emitter . . . or whatever combination of these makes you feel most warm and cozy. It's not going to be a super-exact thing with a satisfying deep null in distortion or anything like that.

On R21. . . that is pretty weird, especially given that it can be cranked down to a dead short if you like. If it hasn't been twiddled yet I'd measure and make a note of the resistance (with the HPF switch in the 80 or 140 position). My guess is that they wanted a maximally-flat Butterworth filter and ascribed particular importance to this especially for the 40Hz position. I'd start by applying a squarewave of about 400Hz and looking at the output on a 'scope, then adjusting R21 for the straightest tilt of the top and bottom. Double-check the frequency response with your AP, re-twiddle if you like. If none of it makes much sense, then put it back in the original place and don't sweat it.

Thanks so much everyone,
Kirkus, i really apprecciate you laying out the circuit like that. Awesome
One more thing: how do i go about the phantom power? At the moment its fed with the +24v and 0v rails that it shares with all modules. Should i connect its 0v pin to chassis instead of 0v hoping to get less interference? 0v and chassis are connected further down at the star point. I do have separate rails for each now, they are not shorted at the module connectors anymore but only at the star point. All XLR input pin1s are connected with a copper rail which is then connected to the star point as well.

EDIT: After reading Ian Thompsons grounding 101, im not sure if it would be better to connect the phantom power units 0v to the copper rail that connects all pin1 on the XLR (and is connected to the star ground). I also attached another document that hints towards connecting it to chassis i guess.

It's really tough to make definitive grounding suggestions simply based on the information in these schematics, as some things aren't clear . . .

- Which grounds are actually connected together in the modules themselves? I.e. pin 5 is shown as an input shield connection and connected to 0V pins 3/4 on the schematic, but 5/7/21/23/25/27 are listed as "screens". Of those only pin 5 is on the schematic at all, and only 5 and 27 are labeled on the plug diagram. I'd guess that pins 1 and 2 connect only to the metalwork of the preamp module itself? Are pins 5/7/21/23/25/27 just connected all together on the module, or on the backplane? Do they really connect with pins 3/4 at any point?

- On the phantom power supply . . . from the twelve 6.8K dropper resistors, I'm assuming that there are a total of six preamp modules in the rack, and they fan out to pins 10 and 12 on each of the modules. That is, pins 5 and 7 on the phantom PSU connect to pins 10 and 12 on the first preamp module, pins 9 and 11 connect to pins 10 and 12 on the second module, etc. etc. So on the negative side, it would make sense that pins 8/12/16/20/24/28 fan out to a particular pin on each of the preamp modules, though that's not shown on the schematic. Also on the plug diagram (on the schematic) there are shown two 150uF electrolytics in series (the equivalent of a single 75uF bipolar cap) across pins 10 and 12 to AC bypass the 6.8K dropper resistors in the phantom PSU . . . I'm assuming these are wired on each preamp module's plug on the backplane?

- I'd likewise guess that there is/was a 24v power-supply module with each of the preamp module's 24v and 0V connections running back separately, where they're all connected together? Do pins 1 and 31 of the phantom PSU go back to there as well? The main thing with grounding this particular phantom supply is that you do NOT want its input current (the 24v and ground connections to the 24v PSU) running through any of the mic shield connections or traces, as the phantom-supply oscillator could potentially crosstalk into other stuff.

- Are there individual XLR input plugs on the main chassis with shielded cables running to each module, connecting to pins 5, 6, and 8 as shown on the schematic? Same on the output with pins 24, 26, and 27? Are the XLR shells common to the chassis metalwork, and does that also screw to each module's metalwork?

You'll need to check and document all of this sort of stuff if you want to use logic to derive a grounding plan. The other way is just trial and error . . . which may work but it's less interesting to read about, so I'm solidly in the "do it with logic" camp. My instinct is to go with the original documentation as closely as possible, then fill in any gaps if necessary . . . even if that's not exactly how I would personally do it if I was designing the unit myself.

Thanks Kirkus!
Ill try to address these point for point:

- pin 1&2 are connected on the pcb and are chassis (all metal work). Pin 3&4 are connected on the pcb and are 0v. All the pins labeled “screens” (5-27) are connected on the pcb. But they are not connected to either 0v or the chassis connection at this point. They were used as screens for the wires going to the XLR connectors though. There ssems to be a bridge suggested on the connector diagram of the schematic, connecting all of these together. It is not installed at the moment.

- im confused about this point. I dont know why you mention pin 10&12. These are the input transformers connections and are shorted. At the moment the Phantom power PSU is connected to +24v and 0V for input. The +48v output is then connected to toggle switches for each channel so phantom power can be switched individually. The output of these toggle switches sends the 48v to XLR pin 2&3 via two 6.8k resistors. There are no capacitors installed at pin 10&12. Are these mandatory in order to protect the input transformers?

- there are three rails running across the connectors. 24v, 0v and chassis. So these short the according pins for all modules. The rails are then connected to the power input XLR. The PSU is external. Pin31 of the phantom psu is connected to the 24v rail and pin1 is connected to the 0v rail.

- yes, that is how the XLRs are connected. But as mentioned earlier, the screens are not connected to either chassis nor 0v at this point at the module connector. All XLR shells are shorted to pin1 and also connected via a copper rail that connects all of them to the star ground.

Another thing. The rack holds two 276a preamps only. Then there are two v272 preamps and one v270. All of these have the phantom switches. But the v272 has no mention of extra capacitors for protection. I never found a schematic for the v270. I attached the v272 schematic as well.

At this point 0v and chassis are only connected at the frames star ground. Not in the external psu.

I hope this helps

Oh, and you are correct regarding R21. It effects the low end response. One module did have a tiny rise at around 40hz and this trimmer took care of it. Awesome!

Kludge's rule of grounds: every ground point in the system has one and only one path to ground to every other ground point. Never zero, never two. So there is one point where everything is bonded.

There are many topologies that can be set up to make this work, and I don't know what topology the console that this was taken from uses. If you can get the manual for the console frame it should show you how grounds are routed.

But it doesn't really matter which topology you use as long as you manage to follow the rule of grounds above.

It is very common to apply phantom to the center point of the input transformer and use a single resistor. Whoever did this used two, one on each edge, and that is fine. If those resistors are balanced perfectly there will be no offset on the transformer. Plug a phantom-powered mike in and measure the offset voltage across the transformer primary and it should be too low to worry about.


Draw it out. Are there multiple ground paths or missing ground paths? If it follows the rule, it's fine. There are many different ways to follow the rule.
--scott