How to Build a DIY Passive Summing Box

I’ve been curious for a while about the fuss surrounding analog summing. But with commercial boxes priced $600+ I wasn’t that curious. Luckily, passive summing is a pretty simple concept and one that’s very wallet- and DIY-friendly to achieve. Check out the video above for a quick run-through of how I built my 16-channel summing box.

Contents:

How It Works

Most analog summing designs take a number of mono inputs and assign them to the stereo bus via either pan pots or L-C-R switches. This route, while superior in functionality, is rather complex and expensive. The design below leaves out panning/assigning and makeup gain in order to avoid the need for power and forgo toggle switches (≈$7.50/channel). What we end up with is a passive bus mixer: 4 or 8 stereo pairs are summed via L and R buses, which are then sent to external microphone preamps for makeup gain. This basic design has been proposed by “New York Dave” and Fred Forsell, and can be found in the monitor section of old Neve consoles.

The greatest benefit of this approach, in my opinion, is the tonal flexibility. Every pair of preamps will impart their own sonic signature on your mix.

Passive Summing Schematic

Inputs are summed to the bus wires via resistors. All of these resistors must be the exact same value with as low a tolerance as possible. Recommended values are between 5-10k ohms depending on the input impedance and volume drop you’re aiming for. The value of the shunt resistors depends on the number of inputs used and the value of your summing resistors. NYD gives us the formula for calculating it, where I is the input impedance, R the desired output impedance, and N the number of channels: (I/N)*R/((I/N)-R)

Putting It Together

Building a passive summing box always seemed like a bit more of a hassle than it should be for such a simple circuit until I found Rhythm in Mind’s idea for mounting everything on a front panel. It’s a great approach for this project, as all of the components weigh in at a grand total of around 3 lbs. Simply drill a panel as I’ve drawn below (or order one!) and mount components right to the panel as shown. I used nylon standoffs to suspend the bus wires with lock washers to hold them taut.
diy summing box

I’ve had no problems with noise or interference with this setup. But if you do find it to be noisy, you can always add a case.

Bill of Materials

Qty Value Note
8 1/4″ TRS jack Panel mount, solder terminals
2 182Ω for 16ch 165Ω for 8ch Metal film, 1/4 watt, 1% tolerance resistor
16/32 6.8kΩ Metal film, 1/4 watt, 1% tolerance resistor
4′ Bus wire At least 24 AWG thickness
6 1 3/8″ Nylon standoff Female->male, 6-32 or 4-40 thread (must match other standoff)
6 3/8″ Nylon standoff Female->male, 6-32 or 4-40 thread (must match other standoff)
6 Machine nut 6-32 or 4-40 thread (must match standoffs)
6 Lock washer #6 or #4 size (must match standoffs)
2 Male XLR jack Solder cup terminals
4 4-40 nuts and bolts For mounting XLR jacks

Using It

Since the inputs of our summing panel are hard wired to either the L or R bus, you can’t simply send a mono signal (unless you want it panned R or L). Rather, assign your tracks to a number of stereo buses and then send those buses to the summing panel as shown below:
passive summing diagram

Update: Adding Mono Inputs

Since posting this article, I’ve received a lot of questions about adding mono inputs for tracks like lead vocal, bass, kick, and snare that are often panned dead center. Luckily, this is really easy to implement with just a little more math. Of course we know that in the world of stereo, a signal that is “dead center” is actually two identical signals of the same volume in each channel. So instead of sending our mono inputs to either the L or R buss wires, we will send them to both.

However, by doubling the signal, we’ve made our mono input 3dB louder than the others. We can rectify this by changing the input resistor values for the mono channel to 1.4x the value of the regular input resistors (1.4:1 being 3dB expressed as a ratio). For example, I used 6.8k Ohm input resistors for the “Endless Summer” kits. 6,800 x 1.4 = 9,520. The closest value stocked by Mouser is 9.53K (very close!), which are now included in the kits.

Keep in mind that if you change the number of channels in your summing box, the ideal values for the output shunt resistors will change as well. Use NYD’s forumula at the bottom of the “How it Works” section to calculate the correct value for your number of inputs.

Any questions, ideas, or suggestions for this project? Let me know in the comments below.

Let's discuss...

*
*

  1. al
    Posted December 18, 2011 at 12:44 am | Permalink

    First of all, this site is amazing. And keep the DIY youtube videos coming. They’re great.

    Question about the shunt resistor.

    I dont know a whole lot about impedance, what did you use for your 8 channel shunt resistors? and did you go with 10k’s for the inputs?

    thanks.

    • Posted December 19, 2011 at 12:30 am | Permalink

      Hey Al, thanks for the kind words. A good output impedance for most mic preamps is 150 ohms. Actually, I ended up using 6.8ks for the inputs to get a little less dB loss, but 10ks work just fine. With 10k input resistors, and 8 channels the formula above for the shunt resistors spits about 159.5. Anything fairly close to that is good.

  2. Sunil
    Posted January 15, 2012 at 2:18 pm | Permalink

    Hi Dave,
    Thanks for the info. New to the DIY community. I’m just wondering are these balanced outs ? Do they make a difference ?
    If I wanted to pick up a case to complete the architecture where would you suggest?

    Thanks in Advance
    Sunil

    • Sunil
      Posted January 15, 2012 at 2:19 pm | Permalink

      Sorry I mean Peterson, you looked like a Dave in the vid.
      Sorry again.

      • Posted January 16, 2012 at 5:12 am | Permalink

        Haha, no problem ;) Yes, they are balanced outputs. Balanced outputs are not always necessary if they will be feeding a short cable run, but are highly recommended.

        For a case, I known of Par-Metal, Collective Cases, Hammond, and Middle Atlantic in the US. Where are you located?

        • Don
          Posted June 27, 2012 at 1:34 pm | Permalink

          Hi Peterson. Can you provide part numbers for those cases?

  3. Sunil
    Posted January 16, 2012 at 12:35 pm | Permalink

    Thanks, will check these out. Are you ready to be shipping kits soon ?

  4. Sunil
    Posted January 16, 2012 at 12:36 pm | Permalink

    Miami

  5. Sunil
    Posted January 16, 2012 at 12:37 pm | Permalink

    Will this yield similar results to other passive summing boxes, eg folcrom ?

  6. Posted January 16, 2012 at 3:45 pm | Permalink

    Yep, kits will ship this week. In terms of features, something like the Folcrom has L/C/R assign switches which this does not. In terms of sound, I’m not sure it’s possible to make something more sonically transparent than this project.

  7. Sunil
    Posted January 16, 2012 at 9:09 pm | Permalink

    Thanks for the info. How would it handle channel panning? Say A channel that is panned slightly to the left\right or different delays to left\right channel or channels that are dead center ?

  8. Posted January 17, 2012 at 3:47 am | Permalink

    Since most mixes have more than 8, or even 16 channels, this project is usually used to sum buses. So, say you have four guitar tracks that are panned in different places throughout the stereo field. You would send all four guitars to a “Guitar Bus” in your DAW or console, and then patch the L and R outputs of that bus to two consecutive channels of the summing box.

    For example, I typically split up my busses as: 1)Vocals/harmonies 2)Bass/Drums 3)Guitars and 4)Keyboards/effects sends. Then I run these buses through different stereo compressors on the way to my summing box.

    Does that clear things up?

    • Posted January 17, 2012 at 3:48 am | Permalink

      I should have added that all of the panning is done in your DAW/console on the individual tracks before the get sent to the buses.

  9. sunil
    Posted January 17, 2012 at 5:03 am | Permalink

    Thanks alot Peterson. The reason I am asking is because I also split my buses similar to you but I have some going to external mono compressors and some go to stereo compressors.
    Eg.
    1)Kick and bass > mono
    2) Snare, hats some percs > mono
    3)Percs\toms\synth blips etc. > stereo
    4)Synths\Effects\pads\Guitars\Piano\Vocals etc > stereo
    Thus I was wondering how this would be affected in the mixer. What would you advise in this situation? Most of my music the kick and base needs to site center.

    • Posted January 17, 2012 at 2:57 pm | Permalink

      Sunil, in your position I’d make the first two channels hard wired mono. I haven’t tried this myself, but I think that you could simply wire the output from jacks 1 and 2 to all four bus wires. Keep in mind, however, that these channels will be 6dB louder than the others if you use the same value resistors. I’m sure someone more knowledgeable than I am could chime in with a formula for calculating the resistors to make those inputs the same volume ;)

  10. sunil
    Posted January 18, 2012 at 9:28 am | Permalink

    Hmm,, that would be great if someone can chime in my. I’d love the flexibility of having mono channels summed as well. Most of my 500 series gear is mono.

    • Posted January 18, 2012 at 4:00 pm | Permalink

      I would think that if most of the channels you want to sum are mono, you will probably want pan controls, right? Then we are talking an active circuit and more complex/expensive.

      • Posted January 19, 2012 at 4:07 am | Permalink

        I did a little research today and it looks like adding a “mono” switch to a couple of channels, or just hard-wiring them to be mono would be quite easy. I’ll try it out and report back.

  11. Sunil
    Posted January 19, 2012 at 6:16 pm | Permalink

    Peterson, Thanks so much! That would be great. Although having the flexibility of a panning switches would be great, usually once I cam out of a DAW mono, I usually don’t need it panned just dead center. Let me know how it goes.

  12. impol
    Posted January 25, 2012 at 4:09 pm | Permalink

    Hello! Why use analog summing mixers? Why dont we just sum them within the DAW and then run the stereo audio track to a mic preamp? What are the differences? :)

    • Posted January 27, 2012 at 4:13 pm | Permalink

      Good question. Some people prefer analog summing for their workflow if they use a lot of outboard gear. Other people think it sounds better than digital summing. Others mix on a console with an inferior master section and use a unit like this to bypass it.

    • Andrew
      Posted March 24, 2014 at 9:46 am | Permalink

      Wow, I’m surprised in 2 years nobody has answered you properly. The difference between summing electrons and summing numbers is huge. This is the equivalent of having an analogue mixing desk, without the space requirements. Many producers will have a desk simply to run busses out of because desks tend to produce much cleaner sound. The math is complicated, but remember that the real world works at almost infinite detail and the computer operates at up to 192khz 32bit FP.

  13. Dan
    Posted January 26, 2012 at 1:52 am | Permalink

    Hi there. I would like to make a custom 16-channel version with D-Sub inputs [similar to the folcrom. Could anyone help me with the resistors needed? what shunt resistors should I use that go before the XLRs? and the other resistors? I need 150 ohm output. I just need the values, because I’m not all that great with electronics! any help would be greatly appreciated.

    Thanks!

    • Posted January 27, 2012 at 4:18 pm | Permalink

      Hey Dan, for my kits I chose 6.8k input resistors and 182 ohm shunt resistors. This will give you an output impedance of 150 ohms and about 26dB signal loss. Good luck with your custom build!

  14. sunil
    Posted January 26, 2012 at 5:39 am | Permalink

    Peterson, any update on the mono channel mod ?

  15. Posted January 27, 2012 at 4:23 pm | Permalink

    Yeah! You can make any channel mono by connecting it to both L and R buses via resistors that are 1.4x the value of the regular input resistors. So, if you are using 6.8k inputs you can use 9.5ks to both buses to make it mono. I’ll be including a bunch of 9.5k resistors in the kits from now on so people can have the option to make some channels mono. Thanks for the idea!

  16. Tim Beach
    Posted February 1, 2012 at 3:30 am | Permalink

    If using 10K input resistors on an 8 channel matrix, wouldn’t you want something like 217 ohm shunt resistors? or I am thinking about this wrong. Is this the right formula for figuring the shunt: (20,000/8 channels)200 / (20,000/8 channels) – 200 = shunt resistor value 217.39.

    Thanks in advance.

    • Posted February 1, 2012 at 4:59 am | Permalink

      Tim, that’s exactly right for a desired output impedance of 200 ohms. I did my calculations, however, based on a 150 ohm output impedance. In truth, there’s probably not a noticeable difference between both. That is, they should both be low enough to make any professional mic preamp input happy.

      Funny, I just noticed today that the SM57 specs for output impedance say “Rated impedance is 150 ohms (310 ohms actual) for connection to microphone inputs rated low impedance.” So that should go to show how (un)important a few ohms of output impedance is when working with modern preamps.

  17. Sam
    Posted February 2, 2012 at 9:45 am | Permalink

    Hi there, I was hoping to order your kit, but the pulldown menu to select 8 or 16 channel won’t let me select either…..so it then won’;t let me add it to my cart.

    Thanks!

    • Posted February 2, 2012 at 5:30 pm | Permalink

      Hey Sam, thanks for pointing that out! I’ve fixed that little bug so you can order away.

      • Sam
        Posted February 3, 2012 at 8:37 am | Permalink

        thanks so much, I look forward to assembling this!

        Sam

  18. Mathieu
    Posted February 6, 2012 at 4:54 pm | Permalink

    Hi there, I’d like to use a summing box like one of these to free some inputs on my mixer.
    My plan would be to plug all of my effect returns into the summing box so they would only use 2 channels on my mixer.

    The only thing though is that I’d really like to plug the summing box’s outputs into the aux return channels of my console that are line level (10Kohm).

    Most of my effects unit output around 600 ohm. Would it be possible with some resistors combo to use the summing box into a line in?

    • Posted February 6, 2012 at 7:43 pm | Permalink

      Hi Matieu, do you have any outboard mic preamps? The problem with using only a passive mixer for your application is that, even if you get the impedances right, you’ll still be loosing 20-40dB. If you do have a pair of preamps however, you go could Effects -> Endless Summer -> Preamps -> Aux Return.

      • Mathieu
        Posted February 17, 2012 at 5:46 am | Permalink

        Sorry it took me so long to reply, I had forgotten about this project.

        I don’t have any external preamp, I only use the console’s pres.

        Isn’t it possible to build a box that doesn’t lose as much dB?

        I know nothing about those things…

        • Posted February 17, 2012 at 6:06 am | Permalink

          Console pres should do the job just as well, I don’t know why I said “outboard.”

          No, there’s no way to do it passively without losing volume while keeping the right impedances.

  19. Mathieu
    Posted February 17, 2012 at 4:00 pm | Permalink

    Alright, thanks Peterson.

  20. tk
    Posted February 25, 2012 at 12:06 am | Permalink

    what are the exact values of the resistors in and out(i.e. tolerance, impedance, watts) for a 16×2 configuration like in your video?

    • Posted February 25, 2012 at 8:41 pm | Permalink

      The exact specs are, input: 6.8k ohm, output: 182 ohm. All resistors should be 1% tolerance, metal-film, 1/4 watt

      • Tk
        Posted February 25, 2012 at 10:22 pm | Permalink

        Thanks man. Keep it up! People like you are well needed in the field!

  21. Simone
    Posted March 16, 2012 at 10:21 pm | Permalink

    Hey
    the project is awesome!
    One question: any suggestion for some cheap and good Mic preamp ? (or an easy DIY) ? Thank you!

    • Simone
      Posted March 16, 2012 at 10:26 pm | Permalink

      I mean… to combine with the summing mixer of the project ;)

  22. Chris
    Posted March 19, 2012 at 7:11 pm | Permalink

    This is great – thanks for the info/video!

    I’m going to build a summing mixer very similar to this but I’ll probably case it and have the inputs/outputs on the back just to keep my rack a bit tidier on the front…

    I was thinking about making it with veroboard rather than copper wire just for ease of soldering, I’ve seen a couple of people do it this way – is either way a better way to go?

    Also, you mentioned hardwiring mono inputs but I wondered if you’d had any more thought about how easy it is to include some mono switches for a few of the inputs?

    Thanks again!

    • Posted March 19, 2012 at 10:05 pm | Permalink

      Hi Chris! Regarding veroboard vs. copper wire: in my opinion copper is copper and either way there is plenty of it for the application. You shouldn’t notice a difference using veroboard.

      For the mono/stereo switches, it would be pretty easy to wire up a DPDT toggle to switch between Center and Left/Right operation for each jack. You’d wire the switch like this:
      -Run hookup wire from the input jack to the two center pins of the switch.
      -From one of the other pairs of pins, run your standard input resistors (6.8k for the Endless Summer kits) to the L or R buss wires as usual. This is the normal “stereo” assignment.
      -From the final pair of pins, run four resistors to each of the bus wires. So each pin will have two resistors: one going to the L bus wire and one going to the R. For the hot (+) pin, run them both to the hot bus wires and the cold wire for the cold (-) pin. These resistors must be 1.4x the value of your normal input resistors (9.5k in the Endless Summer kits).

      Does that make sense? If I have more time this week I can draw a diagram.

      • Chris
        Posted March 19, 2012 at 10:28 pm | Permalink

        Hey Peterson, thanks for clearing that up… it makes complete sense!

        That’s a pretty simple addition so I think it’s well worth doing. I’ll assign the first 4 channels to be switchable (just like they are on the Tube Tech SSA-2B, for example.)

        Cheers!

    • nikola
      Posted May 16, 2012 at 9:53 am | Permalink

      but shouldn’t we change the shunt resistors values then?

      • nikola
        Posted May 16, 2012 at 9:55 am | Permalink

        if we use center l/r switches a mean.

        • Posted May 16, 2012 at 1:53 pm | Permalink

          Yes, technically it will change the values in the formula. But in practice I found it to work fine, even with four mono switches engaged. Of course the performance depends upon your situation and the gear you are using it with.

  23. Simone
    Posted March 22, 2012 at 9:52 pm | Permalink

    Mono Switch! Great idea! :)

  24. zino mikorey
    Posted April 3, 2012 at 10:43 pm | Permalink

    Hi Peterson,
    when will it be available again ?
    thank you

    Zino

  25. Keith
    Posted April 20, 2012 at 9:58 am | Permalink

    This is great! I purchased some things at the local Whack Shack to make something similar to this. A few questions if I may. I’m just getting into electronics and making stuff like this, so please forgive my ignorance. Are the setups universal? I won’t be using a traditional “mic pre” so to say. To make up for gain The signal will be ran through a Lexicon Alpha. My problem is, I don’t know if I need different value resistors due to my driving device. I will be using the TRS jacks since I only have one mic input on the device. The Alpha specs says it has a gain of +50. The line input impedance is 20k ohm balanced, and, 10k ohm unbalanced. I will only be using a total of 4 inputs. But I also wanted to know if I could add a stereo plug to the buss with no problems either. I hope I make sense and I thank you for your patience. As always thanks.

  26. Posted April 20, 2012 at 3:26 pm | Permalink

    Hey Keith, unfortunately I don’t think the Alpha will work for this setup. Typically, line inputs do not provide any gain, and you will need significant gain after the passive summing to properly drive line inputs.

    What are your four inputs? It looks to me like the Alpha only has two outs?

    • Keith
      Posted April 20, 2012 at 8:01 pm | Permalink

      Thank you for your response. I wanted to achieve as follows. I wanted to plug my MPC into one set of L-R and a keyboard into the other L-R, to make up the 4 inputs. I then wanted to add a single stereo plug to the L-R bus for a device like an ipod or such. (Actually 5 inputs). I wanted to then take the output from the passive box into the line in’s on the Alpha device via TRS cables. Thanks again for your help.

  27. Posted April 20, 2012 at 8:26 pm | Permalink

    Keith based on what you wrote I don’t think passive summing is going to do the job you’re looking for. In a passive system like this one, all of the inputs affect each others’ input impedances, so you want to have inputs with relatively uniform output impedances. In your scenario, a keyboard, sampler, and iPod might have quite different output impedances, not to mention output volumes. They really need the buffering that active circuitry provides to play nicely together. And then there’s still the problem of feeding the Alpha’s line inputs what is essentially a mic-level signal.

    All in all, I think you’d get much better results in your situation from a cheap active mixer than a passive one.

    • Keith
      Posted April 21, 2012 at 12:06 am | Permalink

      Thank you. I appreciate you taking your time to offer your insights.

  28. Dan
    Posted April 24, 2012 at 7:14 pm | Permalink

    hi guys. just finished my take on Peterson’s passive summing box. It sounds pretty great. Especially considering it cost me under $50 in parts! The Rolls Music Folcrom goes for about $750 and is basically the same thing.

    The way I wired it up is basically identical to Petersons’ design, except I used a board to attach the resistors to.

    http://i65.photobucket.com/albums/h239/cottonear/image-1.jpg

    http://i65.photobucket.com/albums/h239/cottonear/image.jpg

    http://i65.photobucket.com/albums/h239/cottonear/image-2.jpg

    • Posted April 25, 2012 at 2:27 pm | Permalink

      Hey Dan, looks great! Thanks for sharing. Nice idea to use snake cable and DB25 connectors, too.

  29. e.scarab
    Posted May 26, 2012 at 5:53 am | Permalink

    Peterson, I am picking parts to assemble my version of your basic mixer. I have a question for you: what made you choose the 6.8kohm resisters I realize using a lesser resistor will reduce the amount of gain lost, supposedly at the expense of more crosstalk. However, I thought that if wired as an 8 in, 2 out balanced configuration there is effectively no crosstalk, especially when using the appropriate shunt resistors.
    I will be going into a Langevin DVC myself and am trying to select the optimum buss and shunt resistors.

    I can tell you this, the DVC has has a 2400ohm input impedance and is optimized for 100 – 600 ohm mics. I have decided that perhaps 240ohms will be my output impudence (10% of the total as is the “norm” with most pres) or perhaps 200. But I cannot decide the best actual load resistors to put on the inputs. Although the DVC can easily replace the 40db lost when using 10k resistors, I figure since the DVC is very low noise and because I am using some .1% resistors I found at Mouser, I shouldn’t have any crosstalk issues if I go below 20kohm and so to reduce the amount of makeup gain needed slightly I am considering using 9kohm resistors with 267ohm shunts for 240ohm output or 221 ohm shunts for the 200 ohm output.

    Is there an actual formula to optimize the channel buss resistors? Again, was there a specific reason why you chose the 6.8kohm resistors for the kits and is the logic (and math!) behind my plan to use 9kohm resistors and a 240ohm output impedance (based on the DVC’s specs) sound?

    Thanx for any light you might shed on this for me!

    • Posted June 7, 2012 at 2:37 pm | Permalink

      To the extent of my knowledge, there is no formula to optimize the resistor values for your particular setup. There is just trial and error. But, to be honest, you might be over thinking it! As long as your output impedance is within the standard microphone range, and you’re not loosing more than 40dB of signal, you’ll be fine.

      I choose the 6.8k resistors just by guessing and experimentation. Both the unit I built with 6.8k and 10k sounded fine, but the 6.8k had a little less voltage drop, so I went with that.

      • e.scarab
        Posted June 8, 2012 at 3:51 am | Permalink

        Hey, thanx for the reply. I kind of figured that. I decided to cut the output impedance back to 200ohm. I’m not worried about gain loss because the Langevin Dual Vocal Combo has a total of about 60db of gain if you include the 10db available from the limiter. I picked 18k just to drop the gain loss a small amount. I’ll probable end up trying several versions. One nice theing about the DVC, you can truly run the gain pretty much wide open. I read on another site where someone got good results with a Langevin using the standard 20K inputs with a 200ohm out so 18K should work fine. MUST HAVE HEADROOM! Unfortunately, I won’t be buying one of the kits, however, I still want to say thanx again for sharing ideas and sources and giving clear directions and such.

        I’m going to use combo on the inputs jacks for adaptability and for a box, I decided to go “all analog” and am building a 5″ deep wooden box from a piece of 3/16″ walnut I had laying around and will mount the pre-punched face plate. Just for fun I am even considering laminating a piece of burl maple or rosewood or something on the front panel for fun and looks! If I do I’ll post some pix.

  30. fabio Alex
    Posted June 7, 2012 at 1:51 am | Permalink

    Hello.
    Can I use unbalanced sources on a balanced summing mixer also? (the source with cold and earth solded together on the plug)
    Or it will be necessary build an unbalanced mixer version?
    I will build a balanced mixer but I’ll use an unbalanced audio card…
    Thanks

    • Posted June 7, 2012 at 2:43 pm | Permalink

      Hi Alex,
      You’ll certainly want to build an unbalanced mixer, instead of plugging unbalanced inputs into a balanced mixer. “Ground” in unbalanced and balanced systems are actually completely different things that you generally don’t want to mix.

      The other issue you’ll run into is makeup gain. Using mic preamps will not be ideal as they’ll expect a balanced input.

  31. Nick
    Posted June 26, 2012 at 11:19 pm | Permalink

    Hi Peterson,

    I noticed that the 16-channel version seems to be out of stock…Is this a temporary issue or did you plan to do only a limited run? Also, would there be the option of cascading multiple units together? I’m very excited at the prospect of building this box!

    Best,

    Nick

  32. Don
    Posted June 27, 2012 at 1:44 pm | Permalink

    I would also like to order the 16 channel unit ASAP.

    • Don
      Posted July 11, 2012 at 3:55 pm | Permalink

      Thanks Peterson! The 16 channel is back in stock! Commence ordering!

  33. Mark T
    Posted July 13, 2012 at 6:38 pm | Permalink

    Thank you for the very nice writeup. I just have a few questions:

    About the input resistors, you write “recommended values are between 5-10k ohms depending on the input impedance and volume drop you’re aiming for”.

    Could you elaborate on this? Is there a formula for determining the 5-10k ohm value? What specific input impedance and volume drop were you aiming for when you chose to use 6k8 resistors?

    Second, how would the values of the input resistors change if you wanted to sum 16 instead of 8 channels?

    Finally, you say that the values of the shunt resistors can be determined using this formula: (I/N)*R/((I/N)-R), “where I is the input impedance, R the desired output impedance, and N the number of channels”. Could you show exactly what values you used when applying this forumla? What was the input impedance and desired output impedance used when you wound up using the forumla to select 182 ohm shunt resistors? Obviously, the number of channels used in your case were 4 or 8, but then shouldn’t you have two values of shunt resistors, 182 ohm for 8 channels and something else for 4 channels (or vice versa)? How many channels was the 182 ohm value calculated for?

    Just looking for some details to put all the pieces together. Thanks!

    • Posted July 14, 2012 at 7:06 pm | Permalink

      Let’s break it down!

      Input impedance: There’s a wide range of values that will work. The trade off is between how much of the input voltage is lost (higher value is better), noise (lower value is better), and how much gain we lose with the summing (lower value is better. I was aiming for about a 20dB voltage drop with the 6.8k resistors. The values do not change with the number of inputs.

      Shunt resistors: For my formula I was 13,600 (6.8k*2) (decided based on previous considerations), R was 150 (typical mic output), and N was 8 or 16. The formula puts out 182 for 16 channels and 165 for 8 channels.

      • Mark T
        Posted July 14, 2012 at 9:03 pm | Permalink

        Thank you for your reply. This definitely helps, but I have some more questions now..

        You wrote, “The trade off is between how much of the input voltage is lost (higher value is better), noise (lower value is better), and how much gain we lose with the summing (lower value is better.”

        How are those three values calculated?

        “I was aiming for about a 20dB voltage drop with the 6.8k resistors.”

        Why 20dB in particular? I understand that it’s a compromise between the three things mentioned above (input voltage loss, noise, and gain loss), but why choose this particular tradeoff over another one? Why not a 40dB or 10dB voltage drop?

        • Posted July 15, 2012 at 11:35 pm | Permalink

          No problem, I’m glad to discuss the finer details of this project. Let’s look at the three formulas:

          1. Voltage Drop. The input impedance of the summing box and the output impedance of your line outs essentially form a voltage divider (http://en.wikipedia.org/wiki/Voltage_divider). The formula for the voltage on the output of a voltage divider is Vout=(Z2/(Z1+Z2))*Vin, where Z1 is the output impedance and Z2 in the input. As you’ll see plugging values into the formula, as long as Z2 is 8-10 times higher than Z1 the voltage lost is negligible. That’s why any value between 5-10k is safe for the input resistors.

          2. Noise. I don’t know the formula for noise, but suffice it to say the higher the value of the resistor the higher the noise. Either way, it’s not a major concern.

          3. Volume drop. I wish I could find this right now, I know NYD posted it one time, which is what I used to calculate the voltage lost with 6.8k resistors. I chose 20dB because it straddled the line well between needing too much and to little makeup gain from the preamps.

  34. Mark T
    Posted July 13, 2012 at 6:44 pm | Permalink

    Oh, just one more question.. when calculating input impedance, did you just use the stated output impedance of the equipment you intended to connect to the inputs of the Endless Summer? Or did you include in your calculation of the input impedance the values of the input resistors in the Endless Summer itself? If it was the latter, what formula did you use to calculate the total input impedance? Is it just a sum of the values of all the input resistors and the output impedance of the equipment being connected to the Summer’s inputs?

    Thanks again!

    • Posted July 14, 2012 at 7:26 pm | Permalink

      I’m not sure what you mean. Are you trying to calculate I? Just multiply the value of your input resistors by 2.

      • Mark T
        Posted July 14, 2012 at 9:22 pm | Permalink

        Sorry. I’m just really new to all this, and that’s probably why my questions seem really stupid.

        I understand that you generally want to impedance matching for all your inputs and outputs. For instance, regarding the (I/N)*R/((I/N)-R) formula, you say that you used 150 ohms for R because that’s the typical mic output, which makes perfect sense to me. But then the input impedance is 13,600 because it’s twice 6.8k (the value of the input resistors used in the summer). But shouldn’t the input impedance match the equipment you intend to plug in to the summer’s inputs?

        • Posted July 15, 2012 at 11:38 pm | Permalink

          You don’t want to match impedances! That belongs to the realm of telecommunications and has no place is modern audio. What you want to do is “bridge” impedances, meaning the “load” (input) is arbitrarily higher than the “source” (output of the previous stage). There are a few reasons for this, but the main one is the voltage divider issue addressed above: the lower Z1 and the higher Z2 is, the more voltage is transferred.

          Great question–this is a very common misconception.

          • TDX
            Posted September 4, 2012 at 2:17 am | Permalink

            This was also very useful. Thanks Peterson!

  35. Andy
    Posted July 18, 2012 at 6:50 pm | Permalink

    hello peterson

    inspired by your project i just built a 4-channel summing box and it works well and sounds great. thanks a lot!!! and btw… great site – increased my love for the whole DIY-idea!

    greetings from germany

  36. TDX
    Posted September 4, 2012 at 2:12 am | Permalink

    Hi Peterson. Firstly, thanks for putting this together so well. It has been very useful!

    I need to sum four channels, 2 balanced, and 2 unbalanced (equalling a total of three stereo line inputs). The sources are just two PCs and a USB audio interface with balanced outputs. I have a strong interest in DIY electronics but not a lot of theory, so I have no idea how the math works out on this. I’ve drawn up a circuit diagram based on your design, and I’d really appreciate it if you could have a quick look to see if I’m on the right track: http://i.imgur.com/UoazV.png (The colours help me to make sense of it all, sorry.)

    I’m not sure what to do with the resistors with this design. I think I understand that the resistance on each… pin(?) needs to match so that the resulting output attenuation matches, which in the case of this design means that not all resistors will be the same. For example, the ring on channels 3 and 4 are the only thing running to Pin 3 on the outputs. Then again I could be completely wrong – no idea really.

    I’d really appreciate a nudge in the right direction. Thanks alot!

    • TDX
      Posted September 4, 2012 at 2:22 am | Permalink

      Would it perhaps be easier to just split the output of my sources into unbalanced left and right TS Jacks and just build the summing mixer according to your specs?

      • Posted September 4, 2012 at 1:58 pm | Permalink

        Hi TDX, in a passive system like this you don’t want to mix balanced and unbalanced signals. You could without much trouble simply treat the balanced signals as unbalanced, and make a fully unbalanced mixer. However, if you’re planning on using a mic preamp for makeup gain you could run into trouble because it will be looking for a balanced signal.

        Funny enough, there’s no easy solution for passive unbalanced summing like there is with balanced. This is because of the lack of good, commonly available unbalanced gain stages. You can check out this unbalanced summing schematic from Fred Forssell: http://www.forsselltech.com/media/attachments/8chsum_2.pdf, but keep in mind that it will be looking for a load impedance of 10k on the outputs, which a mic preamp will not provide.

        Best of luck and thanks for the kind words!

  37. Stex
    Posted September 4, 2012 at 1:30 pm | Permalink

    Hi! One question…what happens if I only use eight of the sixteen channels for summing in the 16×2 version? Any problem? Thanks!

    • Posted September 4, 2012 at 2:00 pm | Permalink

      Using any less than all of the channels may increase crosstalk between the channels. I highly recommend building an 8-channel version if you have only 8 channels.

      • David Mansfield
        Posted August 18, 2014 at 3:18 pm | Permalink

        Would this also be a risk if, say, I had two eight channel audio interfaces connected and one of them was powered down? Or is this only an issue if nothing is connected to ports 9-16?

  38. Ben Rosenbach
    Posted September 30, 2012 at 1:02 pm | Permalink

    Can you really heard the difference between those 4 stereo busses being summed in the box vs analog? I’m wondering because I don’t have any outboard gear at the moment so I’m curious if this is worth my time to essentially do half of my summing in analog (since the other half happens in the DAW on the stereo busses). Will I notice a difference?

    • Mark Vitek
      Posted October 20, 2012 at 4:02 pm | Permalink

      If I wanted to add a switch to optionally disable each pair of inputs – could that be done without significantly changing the difficulty of the project? Would it necessitate that the resistor values be changed?

      Thanks!

  39. Irénée
    Posted March 10, 2014 at 9:04 pm | Permalink

    Hi, and thanks for this great tutorial.

    However, I get in trouble about the XLR output/Mic level. The output isn’t balanced, only summed, so wouldn’t it be better to out in TRS/line level ?

    My aim is to sum 2 stereo keyboards (unbalanced TRS/line level). What do you think about this idea ?

    Because in this case, I need to change the output impedance (~10K?).

    Thanks !

    • Irénée
      Posted March 10, 2014 at 9:19 pm | Permalink

      Ooops, I just read the others comments and notice that it dont fit for unbalanced inputs (and i meant TS, not TRS btw).

      I’ll check the good topic ;)

  40. guingo mendes
    Posted March 15, 2014 at 1:33 pm | Permalink

    bom dia. para usar apenas 3 microphones num mesmo channel?

  41. Annyms
    Posted March 16, 2014 at 8:47 pm | Permalink

    Hey Peterson, I just had a couple of questions. If I were to mount everything in a small case, would I be able to use PCBs in this build instead of buss wire? Also, if I were to add the switch for L/R-C, would that not change the number of channels to the L and R buss technically (9 channels each), or does adding one channel to each buss have a negligible impact on the output impedance that you can still use the 165 ohm resistors? Thanks in advance.

  42. Andrew
    Posted March 24, 2014 at 9:48 am | Permalink

    If I wanted to make a 32 Channel version of this… say I was going to make two “16 channel to mono out” versions of these boxes, would I need to change my choice of resistor?

  43. Felipe Navarro D.
    Posted April 1, 2014 at 6:20 pm | Permalink

    Hi.

    This is my first time posting to a forum. Hoping you all can help me out.

    I am interested in building a passive mixing desk for my recording studio that I am building. I recently invested in a custom Tree Audio ‘roots’ console for tracking into a JH24 2″ tape machine that I also recently picked up. I’d like to make a 24 input passive mixing desk with individual channel faders and individual channel pan pots with a stereo output for mixing and simple tape playback monitoring.

    If you can shed some knowledge on what I would need to built this, that would be awesome.

    Thanks in advance,
    F. N. D.
    All Welcome Records
    www. allwelcomerecords. com

  44. Stainless
    Posted May 2, 2014 at 12:50 am | Permalink

    This is great! I’ve just about got all the parts and was wondering ig there would be any advantage in shielding the bus wires. Just something that pooped into my head when I was watching the video. I’d read somewhere about long bus rails being an issue in large mixers for introducing noise)… this may not be long enough to be a concern).

    I have an old Fostex Line Amp 2u case that will be perfect for this!

    Also reading above about signal loss- would a line matching transformer be advantageous? I have some of the old Altec Lansing octal socket line Xformers

    Great column! I just wonder where I was 2 years ago!

    • Posted May 2, 2014 at 4:01 pm | Permalink

      Thanks, glad you enjoyed it! Regarding your questions:

      -Yes, shielding the wires will help. Since everything remains balanced throughout the summing mixer, any common mode noise should be cancelled out by the makeup gain stages. But it can never hurt to stop noise before it gets into the wires.

      -Not really. The signal loss is just part of the beast. Introducing step-up or step-down transformers would make the impedance calculations more complicated and probably wouldn’t do much good.

  45. David Mansfield
    Posted August 16, 2014 at 9:11 pm | Permalink

    Anyone know where I can find a patch panel like this one? I don’t imagine you can make it yourself without a drill press??

    • Mike Falcone
      Posted August 29, 2014 at 2:32 pm | Permalink

      I’m looking for a panel, too. I may just have to pull out a hand drill and do my best. I’m not quite sure how this works but I found this price list at Par-Metal: http://www.par-metal.com/price.php

    • Mike Falcone
      Posted September 1, 2014 at 8:53 pm | Permalink

      I ended up using Front Panel Express which is a company that provides software (free) for you to design your rack panel the way you want it. Depending on how you make it the cost will vary, but it’s a great tool if you’re doing a lot of personal designing of rack equipment. The panel’s aren’t as cheap as buying them blank and then drilling them yourself ($10 for this summing panel supposing you have the correct drill bits and skill), but using Front Panel Express it could be anywhere from $50-100 (depending on how many whistles, colors, engravings, etc. you put on it). http://www.frontpanelexpress.com/download/front-panel-designer/

  • Last updated January 9, 2014

  • Get the DIY Newsletter

    Be the first to know about new DIY gear, giveaways, and exlusive offers
    I will never spam you.