summing large amounts of channels.

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
That's not what I see.
I'm sorry, you're absolutely right. It's one of the features I never use because I use almost all the channels all the time.
This makes me even more suspicious as to what the designer understood (or not) about summing arrangements. There is no doubt that, when properly implemented, connecting the summing amp reference to the ground sensing bus is the best solution.
It maybe that the possibility to connect to the general ground would be for factory testing, and not for permanent use.
Do I understand it correctly, that the sense of the ground reference bus is to have a ground bus that is not carrying all sorts of currents (e.g. power supply) to act as a "clean" reference (regarding ground as a superposition of resistive elements where different more or less small voltages could show up at different points)? So it would be connected to the 0V or main ground at the power supply connector only?
But in order to reap all the noise benefits, you need to replace the ground sensing resistors with a much lower value.
How about leaving it out by replacing it with a wire?
 
I know it's a slight subject of disagreement with JR. My view is that condenser mic noise and room noise have a noticeably different sonic character than noise due to electronics and many musicians discern this difference and understand that room noise is not the console's fault.
I doubt we disagree significantly. I repeat an old warning of mine to lower expectations and avoid too much focus on bus amp noise. Even with a perfect summing amplifier, the console would still have an audible measurable noise floor. IMO a major benefit of managing sum bus noise gain/loop gain margin is improved phase response, and lower distortion.

Some noise sources sound different and some pretty similar. Mic preamp noise will have pretty much the same sound character as bus amp noise. All the mic preamp's noise will combine incoherently so not Nx more like square root of N. Room noise, stage wash, whatever of course varies wildly in sound character. I don't claim that bus noise doesn't matter, only that it must be held in perspective as one of multiple contributors to noise floors.

For today's TMI about noise floors, at Peavey we had a house part number for a 5532 what was selected for a guaranteed noise floor spectral response. Specifically parts with dominant 1/F noise were culled out. This was so when customers listened to noise floors with everything turned up to 11 heard a well behave white noise.
It would certainly; we're counting on you io report the results of the comparison. :LOL:
I am not aware of any way to swap out op amps inside the transamp (a hard potted module). IIRC Paul Buff used a TL072 inside which is fast enough for audio and relatively quiet. Perhaps people are talking about replicating the Transamp topology (aka Cohen but perhaps not the summing application)?
I suggest you go to the proaudiodesign forum. There's a whole thread there.

No. You have to grok the concept of OSI, which has very little to do with impedance matching.

So this makes the impedance of the CB stage about 10 ohms. Actually a little more due to the emitter spread resistance. But it results in a very low noise voltage density.
It's like this 10 ohm resistor was in series with the source, so it should be minimized, until the noise current becomes too high. (noise current increases as base current increases)

I can't read any labels in your pic.
Anyway I'm always suspicious about this type of arrangement. Makes me think that the designer was not quite sure he had made the good choice, so he left that open to someone else.
I've seen several mixers that had multiple "grounds", in particular one bus for the summing amps "ground". The most common mistake is forgetting to put a sampling resistor between channel ground and this bus. Anyway it doesn't play well with the most common arrangement that disconnects the feed resistor from the bus on unused channels.
An extreme case would be only one channel routed; with the summing amp ground sampling all the channels, the residual noise of all the channels is still there.
I've often found that I got a better overall performance using this summing amps bus to reinforce the main ground bus.
I would be cautious of allowing general ground currents to corrupt signals. A "signal ground bus" is more of a signal than a ground, IMO.

wrt to differential sum bus topology, if both + and - buses are active and fed audio signal, 10k resistors on both may increase Johnson noise +3dB, but the coherent signal increases +6dB for a net gain of 3dB to S/N. If the - bus is just summing all the local channel grounds, common practice is to use on the order of 100 ohm resistors. This preserved the differential math while reducing the Johnson noise contribution.
Yes, I noticed that. It may be an attempt at fixing an imaginary offset problem or a stability issue...

Have fun..

JR
 
You have to remember that when gear is designed for 'mass production' there are a couple of other, not necessarily technical voices going on in the background of the design 'office'. One being 'ease of manufacture (cost), another being an accountant (again) who would like you to design it using parts that cost nothing. Using a 75 ohm resistor might have been chosen because they already use a 75 but not 100 in the same unit so saving extra parts inventory, shelf space to store it etc. Similarly, adding chokes or ferites which are a 'good' idea but tend to be quite expensive so if you can get away without it then more profit. A designer I worked with a considerable while ago 'proposed' the use of 0.01% tolerance multi resistor 'packs' (4 resistors matched to 0.01%) who didn't get his way. The whole project was shelved some time later but not for this reason. One consideration was the 'noise' that would be put into the grounding system if each channel module had it's own LM7815 and 7915 regulators.
The concept of 'amplified ground' was around but I don't know of any gear that really used it. Amplified ground being a small guage 'reference' grpound and each channel used a non inverting buffer amplifier to then provide the 'ground' for the remaining audio circuitry.
The definition of noise, even in the previous discusion ought to specify whether it is 'thermal' white noise/hiss or the more loose laymans terms encompassing hums and other 'tones' that are picked up from either internal or external sources. Granted most here are considering thermal noise.
Matt S
 
Do I understand it correctly, that the sense of the ground reference bus is to have a ground bus that is not carrying all sorts of currents (e.g. power supply) to act as a "clean" reference (regarding ground as a superposition of resistive elements where different more or less small voltages could show up at different points)?
That's correct.
So it would be connected to the 0V or main ground at the power supply connector only?
No; it should be connected only to the summing amp's non-inverting input. Actually there should be an additional resistor going to the summing amp ground, for perfect rejection.
How about leaving it out by replacing it with a wire?
No. That's a very common mistake. Due to the resistance of the bus and contact resistance, the channels that are closer to the summing amp would be more sampled than the ones that are further. The resistor must be chosen so that the bus and contact resistance is negligible in comparison. Typically 100r.
 
Using a 75 ohm resistor might have been chosen because they already use a 75 but not 100 in the same unit so saving extra parts inventory, shelf space to store it etc.
The question is not why 75r; the question is why a resistor there? It seems all it does is increase noise.
One consideration was the 'noise' that would be put into the grounding system if each channel module had it's own LM7815 and 7915 regulators.
78xx/79xx are a very bad choice for on-board regulation in distributed power. Managing circulation of noise currents is very difficult compared to LM317/337.
 
I doubt we disagree significantly.
I believe we don't. It is just that, having different histories, we may not have the same sensitivities. I think brings a little pinch of salt in these threads. For sure, we can't diverge much on objective statements, but when it comes to perception, we're all different. Germans, who like bass, may be more critical wrt noise than Japanese... :LOL:
I am not aware of any way to swap out op amps inside the transamp (a hard potted module). IIRC Paul Buff used a TL072 inside which is fast enough for audio and relatively quiet. Perhaps people are talking about replicating the Transamp topology (aka Cohen but perhaps not the summing application)?
Yes, teh idea is to more or less clone the Soundcraft summing amp.
I would be cautious of allowing general ground currents to corrupt signals. A "signal ground bus" is more of a signal than a ground, IMO.
I agree. It was a specific case where the implementation of the ground sense bus was flawed. Actually it was done without sampling resistors. Just pressing on a connector changed the noise level.
wrt to differential sum bus topology, if both + and - buses are active and fed audio signal, 10k resistors on both may increase Johnson noise +3dB, but the coherent signal increases +6dB for a net gain of 3dB to S/N. If the - bus is just summing all the local channel grounds, common practice is to use on the order of 100 ohm resistors. This preserved the differential math while reducing the Johnson noise contribution.
+1
 
I believe we don't. It is just that, having different histories, we may not have the same sensitivities. I think brings a little pinch of salt in these threads. For sure, we can't diverge much on objective statements, but when it comes to perception, we're all different. Germans, who like bass, may be more critical wrt noise than Japanese... :LOL:
For TMI about cultural differences, not about noise floors specifically but musical preferences. Back in the 80s I received a less than glowing review from Hugh Ford, the well respected reviewer working for Studio Sound when he reviewed my Loft Delay line/Flanger. Since he was also well respected by me I actually travelled to London to meet with him and learn more about his criticisms. It turns out that one of his favorite test tracks to evaluate dynamics processing was a very percussive (clicky?) Japanese music track.

The Loft BBD delay depends on companding NR to deliver decent noise performance. Hugh's stress test revealed a control feedthrough problem associated with the inexpensive companding NR chip (NE572?) I used. That sent me back to the test bench to clean up my delay line/flanger's transient handling.

[TMI] To better zero in on transient handling in dynamics processors I designed my own tone burst device. It detected zero crossings and gated on/off only at zero crossings to prevent added step artifacts. Further it only gated on/off even numbers of zero crossings to present only full sine waves and avoid introducing DC components from gating half cycles. My burst device did not have a sine wave generator build in but accepted external audio inputs so I could also gate burst real audio sources, to punch them up to be more dynamic stressing dynamics processors harder. The final feature I added was a dry level bypass so in effect the dynamic range or depth of the gating could be continuously varied between none and full. [/TMI]

I found these extra burst control features incredibly effective at parsing out undesirable artifacts in dynamics processors. Helpful in my development of sundry dynamics circuits over the decades.

JR
 
Which Soundcraft summing amp are we cloning?
500/600, 6000, TS24...
**Edit: I just reread your post, I didn't know that the resistor to ground was critical in the rejecting of noise. I just assumed it was protection.
This shows exactly what I actually know and what I'm bullshiting myself into thinking I know. :D.
We all have our moment of temporary lapse. ;)
 
IMO a major benefit of managing sum bus noise gain/loop gain margin is improved phase response, and lower distortion.
That's exactly what I'm trying to achieve.
No. That's a very common mistake. Due to the resistance of the bus and contact resistance, the channels that are closer to the summing amp would be more sampled than the ones that are further. The resistor must be chosen so that the bus and contact resistance is negligible in comparison. Typically 100r.
Umm, now where should it ideally be?
Generally speaking, the resistors we usually see from each channel's ground sense are as small as possible, on the order of 10 ohms.
In the Midas we've got ground reference, mix reference (to the summing amps), main ground and earth. Earth is connected via 33 Ohms to the channel ground. Ground reference goes into IC18 (op amp) and is fed to channel ground via 100 Ohms but also to the second ground which is used for the op amps that drive the busses through 10k. Mix reference is also connected via 10k to the "second ground". This can all be seen in the channel schematic in case my explanation is too confusing.
The non-inverting input of the summing amps can be connected to ground reference or mix reference. From factory it's jumpered to ground reference. Shouldn't it be set to mix reference?
 
I'm too lazy to read schematics first thing in the morning but in the context of a VE sum amp, the "mix reference" should be the mathematic sum or average of all the local channel 0V references. Signals present at the + input of a VE bus amp will get amplified by N+1, so a mix reference that attenuates the local 0V references by the appropriate ratio will precisely cancel out those local 0V errors.

At least that's one design concept.

JR
 
Umm, now where should it ideally be?
they should be connected to the pan-pot ground, assuming it's clean.
In the Midas we've got ground reference, mix reference (to the summing amps), main ground and earth. Earth is connected via 33 Ohms to the channel ground. Ground reference goes into IC18 (op amp) and is fed to channel ground via 100 Ohms but also to the second ground which is used for the op amps that drive the busses through 10k. Mix reference is also connected via 10k to the "second ground". This can all be seen in the channel schematic in case my explanation is too confusing.
Too many "grounds" IMO. I believe the designer was not quite sure of his choices, so decided to leave that open.
The non-inverting input of the summing amps can be connected to ground reference or mix reference. From factory it's jumpered to ground reference. Shouldn't it be set to mix reference?
I think it should.
What's the point of creating a ground-sensing bus and not use it? Have you tried changing? If it works better with the factory settings, something is wrong in the implementation.
 
Shouldn't it be set to mix reference?
It depends.
You may find that using one way or the other produces the 'least' amount of whichever noise you want to eliminate. In a short frame version, in an environment (studio) which is not plagued with signifivant hum or other interference sources_it probably makes very little difference.
While some concepts work brilliantly in small or single channel situations, translating it to gear with many signal paths which may be several mretres long with possibly many amps of power rail current (hopefully reasonably balanced in draw so there isn't a lot in the ground). Working with mixers with only a single supply rail (24 Volts) where the 'power ground' has 5 or 6 Amps flowing requiring the use of half or threquarters inch thick copper or aluminium 'power ground' buses just adds another level of complexity. Expecting this to be 'immune' to a 2 Watt 'walkie talkie' when it is keyed while lying on the control surface then tests the internal signal loops and RF rejection. Having units that use transformers for module inputs and outputs thus separating amplifier signals from power was one solution favoured in German broadcast equipment design and some American where there were plug in modules the Yard' mixer comes to mind in either transistor or valve variantalthough it's maker's name eludes me at the moment.
 
Thanks Bri
I was having a 'senior moment' and just couldn't think of the name although saying that I am not sure the unit was Gates. The chassis was certainly wired for valve and transistor variant modules (not simultaneously)! I am more inclined to think RCA and it had funky dual transistors. Pairs of transistors in an oval aluminium can. Also rotary stud contact faders
 
Alright guys, thank you so much for helping me and answering all my questions! I will get back and report as soon as I finished all my tests. Might take a while though...
 
Back
Top