PCBs for scalable DIY Neve style summing mixer possible feeler

[Potato Cakes]

Hello, everyone,

I'm thinking about adding a summing mixer of some sort to my console to add options for things like mixing stems to better suit a variety of sonic demands/requests from clients. The CAPI SumBuss looks awesome and my plan was to go that route, but the mods on my console's channels and the mix buss lend itself in that direction. Lately I've been enjoying my Neve style preamp/EQ builds and I've decided if this idea moves to board design it will be in this direction. Here's the initial design ideas:

Boards will be in banks of 8 input channels. Each input channel will have a footprint for a VTB9045M or mounting holes for an EA10468 and a BA189AV circuit for adjusting gain to user specific settings. PCB dimensions will be so that two of them and two BA283AM circuits could fit in a 1U case approximately 11 inches deep with relative ease. The 283AM circuit will not be on the board for as there are already various ones of them available that work great and this also allows for using various mix buss circuits like the ACA-BO. My initial thought is to have a push button toggle on board for mono/stereo settings for each pair of inputs. If someone desired they could incorporate pan, volume, etc., controls but they will have to be wired off board to so as to keep the PCB size to a minimum.

I'm going to start casually compiling my notes from previous Neve builds come up with some crude layout ideas to then take the board designer I've been using. Probably the one item of business that I need to make certain of is the buss resistor value. I've previously used 15k with a fair amount of success but noticed a fair bit of cross talk, which may be normal for older Neve mixer circuits. Has anyone found better/different good results with other values? Everything else I am trying to do seems a bit straightforward.

I'm curious if this is of some interest to anyone besides myself. I've used these same buildings block circuits on other projects and the results were exceptional, so I have no reason why this won't sound great as well.

Thanks!

Paul

 

[Thro]

Great idea, i was just thinking about doin this with the ez 1290 pcbs and line input transformers

 

[KDE]

Love the idea and would be interested in making a 32 channel box. My questions also is why not use the line input transformer?

 

[Potato Cakes]

The line in version is an option, but as seen in a number of Neve style preamp designs the line input transformer is done away with and unity gain is still achieved. I'm also thinking about a way of making each channel input have a slightly higher gain setting then attenuate it before it goes to the summing amp, that way at unity gain on the output, the channel is saturated slightly. If I'm using the mic transformer version then I can set the initial channel gain at the transformer. This is the part that will require some experimenting and listening to see where this saturation/distortion is to occur, or if it's even desirable at all.

As mentioned, this is a little ways down the road. I'm merely trying to gauge if this is something that I'm going to do just for myself.

Thanks!

Paul

 

[KDE]

Gotcha ya put me down i'm interested.

 

[ilfungo]

Great Idea!

 

[SIXTYNINER]

Great Idea!

+ 1 Here 

 

[madreza]

Been Dreaming of that for a while ..... never went into designing !
I'm for sure in
so , no insert point , pan and volume onboard , right ?

 

[Potato Cakes]

Been Dreaming of that for a while ..... never went into designing !
I'm for sure in
so , no insert point , pan and volume onboard , right ?

I've only slated the board to have stereo/mono switches per each channel. If someone wanted fader, mute, etc they would have to add that.

Hopefully I can get back to working on this in November.

Thanks!

Paul

 

[Potato Cakes]

Hello, everyone,

I'm slowly circling back to this idea and am about to start doing some rough sketches for board layout. Where I've landed on design is that each board will be essentially eight channels of BA283 with the input transformer on board. The output transformer is omitted as that will be after the summing amp of the end user's choice. There will be a switch point for mono and stereo per each pair. If the end user desires to add faders, pan, insert, etc, then that will have performed by integrating their scheme for such things with the schematic for this circuit. The input will be connected via a separate small board that will be a DB25 connector and header of some sort to simplify wiring to the input channels and allow for more flexibility in case and panel layout. All of the channels will feed a buss resistor network that would then feed a summing amp of the builder's choosing. There will also be solder points for cascading other summing PCBs together. The thing yet to be determined is where to boost signal to each channel. If using the VTB9045 transformer, the highest gain with the input of 1k2 ohms achieved is +6, but since this circuit takes one side of the secondary winding to ground, I think this puts the total gain going into the circuit back to unity, which I could be wrong. If I am correct about what actual signal levels are getting to the input of each channel circuit, then the next place would be to boost the gain at the preamp stage and reduce the gain after the output circuit before the buss resistors. Or maybe this reduction could take place with the selection of the value of the buss resistors. I would like have the signal be at unity before it hits the summing amp. I also want to make sure there is a boost in this circuit so as to not require the various converters to need to be driven harder at their outputs to achieve the main goal of this design.

I do have some questions I'd like to get some input from others who have worked with this style of circuit before:

How necessary is it to have perfectly curved solder traces on the board? I have seen some discussions mentioning this is important but to save some space I might not be able to maintain the wide, flowing curves found on original cards.

Will the MJE3055T work well as a replacement for the 2n3055? The latter is getting harder to find and more expensive. Also, to save space the MJE3055 would be ideal.

Which buss resistors values are preferred for this style of summing circuit? I have read and have used in the past 15k with success, but have noticed a decent amount of cross talk with this amount of resistance.

I'll do some more digging on my own. If anyone has some helpful insight or links to info that would be beneficial to this design I would appreciate having it sent my way.

Thanks!

Paul

 

[trashcanman]

How necessary is it to have perfectly curved solder traces on the board? I have seen some discussions mentioning this is important but to save some space I might not be able to maintain the wide, flowing curves found on original cards.

Not at all necessary, unless you’re superstitious. Any competent layout should be fine.

 

[Potato Cakes]

Not at all necessary, unless you’re superstitious. Any competent layout should be fine.

Definitely not superstitious. Other really smart people made their board layouts of this flavor of circuit with the same long curves in their design, which prompted to think there was an important reason to do so. Even though the PCB designer I use is very good, I still have to provide schematics and direction of what do for the layout and I want to be able to give him the best information to guide his decision making.

Thanks!

Paul

 

[alexc]

It's a great idea ...  I would just say don't skimp on the trimmer pot on the singled-ended nve-ee style output - try to use a good quality one. 

Also, I like bigger traces rather than smaller and and not fussed with 'curvature'  ..  if possible give the little transistors some space on the legs, preferably with those nice '3-pin small signal transistor sockets  ' - it aid's any 'repair-replace'  ops,  don't you know.

Essentially, if it's possible to service and the 'connectors' are good, then - fabulous -  I could use a couple right now!

Some reasonable psu board, with linear regs of sufficient capacity and physical size to allow for easy, proper heatsinking on the reg +24V supply    is important .....  one can easily use a 'jlm 5 rail'  type of thing for the 'auxilliary rails' as required.

[I like DI, +48V phantom, vu meters, leds meters and 'relaiy bypass'  on my neve--ee style pre-amps  personally]

Anyway - good luck with it

 

[Rocinante]

So I did this sorta. Kind of a blend. API meets Neve.
Check out
https://groupdiy.com/index.php?topic=62302.0
I used 990s and app DOA's to keep in line with the 24v needed for the Neve boards I used for summing. Ian Thompson Bell (former Neve tech) pretty much walks me through everything

 

[Potato Cakes]

So I did this sorta. Kind of a blend. API meets Neve.
Check out
https://groupdiy.com/index.php?topic=62302.0
I used 990s and app DOA's to keep in line with the 24v needed for the Neve boards I used for summing. Ian Thompson Bell (former Neve tech) pretty much walks me through everything

I saw and read your build thread a while back and really enjoyed following along. And I do know who Ian is and I try to stick closely to his Grounding 101 Paper as close as I possibly can on each build.

This is originally designed as an mix option for my console's buss outputs being routed to an external black box with some mono/stereo switches. Other who may build this can add what ever they want in terms of controls, but that will be up to them to know how to read a schematic and implement the various add on circuits therein. Before I modded and cleaned up an old console I now use, I mixed with an SPL MixDream XP and loved the simplicity of just having the mono/stereo controls, which allowed for easy session recall and still summing in the analog domain, which is where I have been getting the best results (for me). For this design I'm sticking to that approach as it will allow the foot print to be smaller and allow builders to build 32 channel units with transformers in a 2U case, at least I hope.

The current thing I have to figure out is which part to drive the signal for harmonic content which would then be attenuated before it hits the summing amp. Can this be done with just a 283NV circuit or would I need the full 283AV to properly do this?

Thanks!

Paul

 

[Rocinante]

I believe the 283nv would be fine. Are you using 283s and xfmrs for each channel? That is a big driver of the 283nv I believe.
Or are you just running a switchable line circuit to get summed by the 283s?

 

[Potato Cakes]

I believe the 283nv would be fine. Are you using 283s and xfmrs for each channel? That is a big driver of the 283nv I believe.
Or are you just running a switchable line circuit to get summed by the 283s?

Correct. Input transformer for each channel then sent to a summing circuit, which would be the builder's choosing. This way, the end user can set different drive levels for individual channels according the tastes of the end user and not have to push the circuit on the summing amp itself. The thing I haven't figured out is where on the NV channels circuit to adjust the input level to create the varying degrees of saturation. On a 2520 style opamp there is a feedback loop that does this and I'm guessing something similar for the NV circuit.

Thanks!

Paul

 

[Rocinante]

Isn't it at the 5k trim pot by the 2n3055?

 

[Potato Cakes]

Isn't it at the 5k trim pot by the 2n3055?

That is for biasing, which is on the AM (output) part of the 283AV circuit. If I am using just the NV, which would simplify things significantly, then I believe I would have to add something akin to a feedback circuit. I haven't had time to sit back down with the schematics and do some signal tracing to figure it out.

Thanks!

Paul

 

[Potato Cakes]

Doing some more scribbling and here is where I am currently. On the full 1073 schematic, R38 (18k) seems to be a feedback resistor for the NV portion of the 283AV card (connections U and S), which I believe could be made to something like a 25k trimmer used for gain adjusting the circuit. The other consideration is the input transformer. According to the schematic, for 0dB per the gain switch labeling signal enters through the line input transformer (wired for -13dB) then goes to a 2k2 resistor to ground. From the same connection point of the input transformer to the 2k2 resistor, signal goes through a series of resistors between the switch positions until it arrives at the pole where it is connected to the 0dB position and then it goes to the U connection of the 283AV card. That same transformer could be rewired for -6dB to introduce more level into the circuit as well. Either way, I think I have a solution for driving the circuit to achieve the varying levels of saturation/distortion/whatever.

The last thing I need to get a better understanding is selection of buss resistors. Previous designs I've done in this style of circuit I've used 15k (which is what I read was historically used in these particular brand of consoles) but I've found that maybe the crosstalk between left and right is a little too high for certain applications. I don't know if arbitrarily raising the resistor value would cause unwanted impedance issues. Once I determine what value to use I think I can proceed with assembling notes to give to my PCB layout guy and get some boards made.

Thanks!

Paul