Pots used backwards in sends?
- Thread starter cuelist
- Start date
Help Support GroupDIY Audio Forum:
With this arrangement, the noise gain of the summing amps is minimised when aux sends are at zero.
When a send is at zero (off), the injection resistance is 50k. At max it is ca. 7k.
Overall, the noise gain of the summing amp varies in a 1:7 ratio. Noise gain is FB resistor/Equiv res of all bus resistors + 1
Concretely, if you have 24 channels on a summing amp with an EIN of -105dBu, in a standard configuration (bus-injection resistor after pot) the noise gain is always equal to the number of channels plus one, so you end up with 28dB noise-gain => -77dBu at the output.
With the alternative config (pot after res),
With all sends off, noise gain is ca.12dB => -93 dBu output noise
With all sends up, noise gain is ca.30dB => -75 dBu
With a typical config of half the channels at 10dB below max, the noise gain is around -85 dBu.
The drawbacks to this arrangement are:
Pot cut-off is not as good, in particular as the wiper-to-track resistance increases with wear
If there are many sends, they impose too much load to the preceding stage, so require a buffer
Level tends to jump at the end of travel; ideally an S-curve pot should be used.
When a send is at zero (off), the injection resistance is 50k. At max it is ca. 7k.
Overall, the noise gain of the summing amp varies in a 1:7 ratio. Noise gain is FB resistor/Equiv res of all bus resistors + 1
Concretely, if you have 24 channels on a summing amp with an EIN of -105dBu, in a standard configuration (bus-injection resistor after pot) the noise gain is always equal to the number of channels plus one, so you end up with 28dB noise-gain => -77dBu at the output.
With the alternative config (pot after res),
With all sends off, noise gain is ca.12dB => -93 dBu output noise
With all sends up, noise gain is ca.30dB => -75 dBu
With a typical config of half the channels at 10dB below max, the noise gain is around -85 dBu.
The drawbacks to this arrangement are:
Pot cut-off is not as good, in particular as the wiper-to-track resistance increases with wear
If there are many sends, they impose too much load to the preceding stage, so require a buffer
Level tends to jump at the end of travel; ideally an S-curve pot should be used.
I don't know how _exactly_ are the "hot" lugs connected (looks like directly to the bus), but if this is so, you have the almost-same problem as fender jazz bass: if pots are in the "middle" setting, your bus will either (a) allow you to listen to the radio or (b) if the bus is heavily shielded, it will have parasitic capacitance to ground, which will raise NG of the summing circuit at HF.
If summing resistors are in close proximity to the summing amp, as "traditional" resistor-after-pot topo allows, there is "less radio"... naively speaking.
If summing resistors are in close proximity to the summing amp, as "traditional" resistor-after-pot topo allows, there is "less radio"... naively speaking.
I have a Presto 3 into 1 preamp/mixer from the 1940's which does this. 3 tube plates hit caps, then are wired to 3 Daven 250K sweepers. The 3 Daven pots are in parallel, and the top of the pot resistances feed the output stage grid.
![Soldering Iron Kit, 120W LED Digital Advanced Solder Iron Soldering Gun kit, 110V Welding Tools, Smart Temperature Control [356℉-932℉], Extra 5pcs Tips, Auto Sleep, Temp Calibration, Orange](https://m.media-amazon.com/images/I/51sFKu9SdeL._SL500_.jpg)
Yes, the ancients are constantly stealing our ideas.
This is done on modern mixers too. Only downside is you can't differential the bus (with a single pot), but on auxes or a lightly loaded sum bus, the reduced noise gain mitigates local ground issues.
JR
This is done on modern mixers too. Only downside is you can't differential the bus (with a single pot), but on auxes or a lightly loaded sum bus, the reduced noise gain mitigates local ground issues.
JR
I like it very much, John. Is it your creation? I'd like to translate it and put it on my website, in company with Calvin Coolidge's "You lose" and a wonderfully psychedelic sentence from Lynne Levitan.JohnRoberts said:
no some ancient stole the saying from me too ;D
JR
JR
A classic Potentiometer wants a ZERO-impedance source and a infinite impedance load. Or non-Zero/Infinite with some "error" of rotational accuracy.
This works fine from triode plate to grid, op-amp output to hi-Z input, etc.
When you have a Hi-Z source and Low-Z load, it may work better "reversed". When you have more than two pentode preamps to mix together, our Linear pot will act more "tapered" because there are two actions happening: the pentode is shorted-out, and the mix resistor (upper part of pot) is increasing against the suming node impedance.
The pot-law approaches "square": 50% turn gives 25% output, 90% turn-down gives 1% output, etc. The exact law is rarely square, but often a whole lot more useful than linear, and a penny cheaper than stocking audio-taper pots. If you can design your WHOLE board around one pot-part, inventory is simplified and workers won't put pots in the wrong places. KAY used 500K Lin for everything: Vol, Tone, Trem rate and depth.
It is also possible to gimmick the impedances so that one input full-up has gain of 1,000, but four inputs full-up get gain of 500 each. This mirrors the rule of many-mike PA systems: doubling the number of open mikes takes 3dB off your acoustic feedback margin. You can set the system so one mike full-up is just stable, then turn-up additional mikes and per-mike gain is reduced just about enough to keep the system stable. (This effect may be undesirable in other work.)
Which is what abbey said, without numbers, and generalized beyond zero-Z mixers.
> only 8 channels so the drawbacks you pointed out probably don't matter much.
Teac 3340 and Fender Deluxe use it for just two inputs. For Teac it was the same price as "conventional" pot and mix-resistor, but gave a useful taper on a linear pot. Fender had hi-Z preamps and omitted mix resistors, saved a dime right there (though lost it with a bigger coupling cap). Fender's implementation interacts like crazy, giving some unexpected results when one input is driven and the other input's knob is set to intermediate positions. Funky but extrememely long-lived design.
As TV says, similar "backward mixing" is found inside some guitars.
Ralph is often cited:
'All my best thoughts were stolen by the ancients.' (Ralph Waldo Emerson)
http://quote.robertgenn.com/getquotes.php?catid=228
Has often been used by Don Davis:
'I think it’s a quote from Don Davis of Syn Aud Con: “The ancients are stealing our inventions”.'
http://www.prosoundweb.com/article/ancients_stealing_our_inventions_what_we_can_learn_from_rick_nelson_recordi/mm
This works fine from triode plate to grid, op-amp output to hi-Z input, etc.
When you have a Hi-Z source and Low-Z load, it may work better "reversed". When you have more than two pentode preamps to mix together, our Linear pot will act more "tapered" because there are two actions happening: the pentode is shorted-out, and the mix resistor (upper part of pot) is increasing against the suming node impedance.
The pot-law approaches "square": 50% turn gives 25% output, 90% turn-down gives 1% output, etc. The exact law is rarely square, but often a whole lot more useful than linear, and a penny cheaper than stocking audio-taper pots. If you can design your WHOLE board around one pot-part, inventory is simplified and workers won't put pots in the wrong places. KAY used 500K Lin for everything: Vol, Tone, Trem rate and depth.
It is also possible to gimmick the impedances so that one input full-up has gain of 1,000, but four inputs full-up get gain of 500 each. This mirrors the rule of many-mike PA systems: doubling the number of open mikes takes 3dB off your acoustic feedback margin. You can set the system so one mike full-up is just stable, then turn-up additional mikes and per-mike gain is reduced just about enough to keep the system stable. (This effect may be undesirable in other work.)
Which is what abbey said, without numbers, and generalized beyond zero-Z mixers.
> only 8 channels so the drawbacks you pointed out probably don't matter much.
Teac 3340 and Fender Deluxe use it for just two inputs. For Teac it was the same price as "conventional" pot and mix-resistor, but gave a useful taper on a linear pot. Fender had hi-Z preamps and omitted mix resistors, saved a dime right there (though lost it with a bigger coupling cap). Fender's implementation interacts like crazy, giving some unexpected results when one input is driven and the other input's knob is set to intermediate positions. Funky but extrememely long-lived design.
As TV says, similar "backward mixing" is found inside some guitars.
Ralph is often cited:
'All my best thoughts were stolen by the ancients.' (Ralph Waldo Emerson)
http://quote.robertgenn.com/getquotes.php?catid=228
Has often been used by Don Davis:
'I think it’s a quote from Don Davis of Syn Aud Con: “The ancients are stealing our inventions”.'
http://www.prosoundweb.com/article/ancients_stealing_our_inventions_what_we_can_learn_from_rick_nelson_recordi/mm
PRR said:
Thanx for the quote reference, I never bothered to look it up before using it.. So RWE stole it from us. 8)
While this is a slight thread veer, I did something similar to that to make a crude NOM function in a cheap installed sound mixer amp years ago.
I'll try to keep this short.. this family of mixer amps had plug in modular inputs and a common bus between the preamp inputs and power amp outputs that was neither low nor high impedance. Like Goldilocks porridge it was just right. This middling impedance midpoint was useful for meatball unbalanced busing together of multiple inputs or multiple outputs, or both.
To make a pseudo automatic NOM mixer, I used the gate function logic from each mic input module to detect when the mic was active and clamp down on the common impedance node with a fixed resistor. For a handful of inputs this gave a passable NOM characteristic. No Dugan AM but not bad for <$1 per input
JR
Thanks for the reference. I'll post it soon.
