Forums

Stereo mixer amplifier

Started by Unknown October 6, 2019
Greetings,

An article about my stereo mixer-amplifier project is now available.

http://crcomp.net/electronic/mixeramp/index.php

Thank you, 73,

-- 
Don Kuenz KB7RPU
There was a young lady named Bright Whose speed was far faster than light;
She set out one day In a relative way And returned on the previous night.


On Sun, 6 Oct 2019 14:54:39 -0000 (UTC), Don Kuenz, KB7RPU
<g@crcomp.net> wrote:

>Greetings, > >An article about my stereo mixer-amplifier project is now available. > >http://crcomp.net/electronic/mixeramp/index.php > >Thank you, 73,
The power supplies look strange to me. Why bias up the + inputs of U2 and U3? Why so many DC block caps after the transformers? Why any? Why split the 12 volt supply at all? Looks like that adds complexity and noise. -- John Larkin Highland Technology, Inc lunatic fringe electronics
On 10/6/19 11:28 AM, jlarkin@highlandsniptechnology.com wrote:
> On Sun, 6 Oct 2019 14:54:39 -0000 (UTC), Don Kuenz, KB7RPU > <g@crcomp.net> wrote: > >> Greetings, >> >> An article about my stereo mixer-amplifier project is now available. >> >> http://crcomp.net/electronic/mixeramp/index.php >> >> Thank you, 73, > > The power supplies look strange to me. Why bias up the + inputs of U2 > and U3? > > Why so many DC block caps after the transformers? Why any? > > Why split the 12 volt supply at all? Looks like that adds complexity > and noise. > > >
And if there is a split rail already available why are the op-amps non-inverting inputs biased thru a divider and decoupling cap? The op amp stages seem to just be there so 4 (8!) transformers can be used, one for each input, and the outputs of the transformer summed independently without signal loss but other than that it and the 50k attenutor just add noise. if they just used one transformer and a relay to switch the inputs you could use a passive attenuator of lower value, plenty of drive available from the line level signal and the TDA amp has plenty of gain.
On 10/6/19 8:55 PM, bitrex wrote:
> On 10/6/19 11:28 AM, jlarkin@highlandsniptechnology.com wrote: >> On Sun, 6 Oct 2019 14:54:39 -0000 (UTC), Don Kuenz, KB7RPU >> <g@crcomp.net> wrote: >> >>> Greetings, >>> >>> An article about my stereo mixer-amplifier project is now available. >>> >>> http://crcomp.net/electronic/mixeramp/index.php >>> >>> Thank you, 73, >> >> The power supplies look strange to me. Why bias up the + inputs of U2 >> and U3? >> >> Why so many DC block caps after the transformers? Why any? >> >> Why split the 12 volt supply at all? Looks like that adds complexity >> and noise. >> >> >> > > And if there is a split rail already available why are the op-amps > non-inverting inputs biased thru a divider and decoupling cap? > > The op amp stages seem to just be there so 4 (8!) transformers can be > used, one for each input, and the outputs of the transformer summed > independently without signal loss but other than that it and the 50k > attenutor just add noise. if they just used one transformer and a relay > to switch the inputs you could use a passive attenuator of lower value, > plenty of drive available from the line level signal and the TDA amp has > plenty of gain.
At the very least if OP got rid of the AC coupling caps C3 and C10 and DC coupled the input stage the total attenuator value could be brought down to like 1k instead of 50k. There's already a 100uF DC blocking cap in series with the input like to the TDA like the datasheet says so why is 0.22uF C11 in series with that? and those other resistors. if the attenuator were lower value you could make it op amp biased to mid-point -> attenuator -> DC blocking cap -> TDA input, that's it
On Sunday, October 6, 2019 at 8:55:54 PM UTC-4, bitrex wrote:
> On 10/6/19 11:28 AM, jlarkin wrote: > > On Sun, 6 Oct 2019 14:54:39 -0000 (UTC), Don Kuenz, KB7RPU wrote: > > > >> Greetings, > >> > >> An article about my stereo mixer-amplifier project is now available. > >> > >> http://crcomp.net/electronic/mixeramp/index.php > >> > >> Thank you, 73, > > > > The power supplies look strange to me. Why bias up the + inputs of U2 > > and U3? > > > > Why so many DC block caps after the transformers? Why any? > > > > Why split the 12 volt supply at all? Looks like that adds complexity > > and noise. > > > > > > > > And if there is a split rail already available why are the op-amps > non-inverting inputs biased thru a divider and decoupling cap? > > The op amp stages seem to just be there so 4 (8!) transformers can be > used, one for each input, and the outputs of the transformer summed > independently without signal loss but other than that it and the 50k > attenutor just add noise. if they just used one transformer and a relay > to switch the inputs you could use a passive attenuator of lower value, > plenty of drive available from the line level signal and the TDA amp has > plenty of gain.
If you used a relay to select an input it wouldn't be a mixer. The Opamps are used as summing amplifiers to compensate for the resistive losses for each input. That allows higher value potentiometers, and near zero interaction between sources. Have you ever examined the design of a commercial audio console? Split power supplies and what you consider as wasted parts are used to produce the required flatness and isolation along with the lowest overall system noise. Even early tube mixing boards used transformer inputs, and the fades had to provide at least 80dB of attenuation at their lowest settings. Find some manuals and study them. The crap circuits used in guitar amps are useless for real audio processing. You can do balanced inputs or outputs with just opamps, but transformers eliminate ground loops which induce hum. A capacitor between a transformer and an active circuit prevent it from affecting the biasing of the input or output stages. Some of the first PA systems I worked on, I've despised unbalanced inputs and passive mixers. Cheap crystal microphones, on long single conductor shielded cables that picked up radio stations, and police cars were common in the '50s and '60s. I have worked with many broadcast and studio consoles, from Mono, eight input tube based Gates from the '40s to state of the art 40 input stereo consoles with six output busses. They all used transformer inputs and outputs, which allowed balanced or unbalanced inputs. Early equipment as 600/150 ohm, selected by using half or all of a winding at the connections. All inputs were line or mic level. Turntables used external magnetic phono preamps. At one very remote site, I had to use a spare mag phono preamp to bring the old WWII vintage telephone circuit up to an acceptable level. There was over 30dB of loss in the last mile of old cable, but the telephone people couldn't wrap their minds around anything without a dial and a bell.
jla...@highlandsniptechnology.com wrote:

-----------------------------------------

Don Kuenz, KB7RPU
> > >Greetings, > > > >An article about my stereo mixer-amplifier project is now available. > > > >http://crcomp.net/electronic/mixeramp/index.php > > > >Thank you, 73, > > > The power supplies look strange to me. Why bias up the + inputs of U2 > and U3? >
** The OP thinks he has to bias the coupling electros.
> Why so many DC block caps after the transformers? Why any? >
** See above.
> Why split the 12 volt supply at all? Looks like that adds complexity > and noise.
** The OP is a radio ham. BTW: The input transformers used are of telephone quality. The OP is a radio ham. .... Phil
On 10/6/19 9:34 PM, Michael Terrell wrote:
> On Sunday, October 6, 2019 at 8:55:54 PM UTC-4, bitrex wrote: >> On 10/6/19 11:28 AM, jlarkin wrote: >>> On Sun, 6 Oct 2019 14:54:39 -0000 (UTC), Don Kuenz, KB7RPU wrote: >>> >>>> Greetings, >>>> >>>> An article about my stereo mixer-amplifier project is now available. >>>> >>>> http://crcomp.net/electronic/mixeramp/index.php >>>> >>>> Thank you, 73, >>> >>> The power supplies look strange to me. Why bias up the + inputs of U2 >>> and U3? >>> >>> Why so many DC block caps after the transformers? Why any? >>> >>> Why split the 12 volt supply at all? Looks like that adds complexity >>> and noise. >>> >>> >>> >> >> And if there is a split rail already available why are the op-amps >> non-inverting inputs biased thru a divider and decoupling cap? >> >> The op amp stages seem to just be there so 4 (8!) transformers can be >> used, one for each input, and the outputs of the transformer summed >> independently without signal loss but other than that it and the 50k >> attenutor just add noise. if they just used one transformer and a relay >> to switch the inputs you could use a passive attenuator of lower value, >> plenty of drive available from the line level signal and the TDA amp has >> plenty of gain. > > If you used a relay to select an input it wouldn't be a mixer. The Opamps are used as summing amplifiers to compensate for the resistive losses for each input. That allows higher value potentiometers, and near zero interaction between sources. Have you ever examined the design of a commercial audio console?
Ok right you are. The "mixer" part seemed to fade from my mind...
> Split power supplies and what you consider as wasted parts are used to produce the required flatness and isolation along with the lowest overall system noise. Even early tube mixing boards used transformer inputs, and the fades had to provide at least 80dB of attenuation at their lowest settings. Find some manuals and study them. The crap circuits used in guitar amps are useless for real audio processing. You can do balanced inputs or outputs with just opamps, but transformers eliminate ground loops which induce hum. A capacitor between a transformer and an active circuit prevent it from affecting the biasing of the input or output stages.
It's just there for a single TL0x dual op amp it has 100dB of PSU rejection. The TEA amp is single supply. A BU634P to make a virtual ground for 1 dual op amp seems like over-kill any way you cut it...
> Some of the first PA systems I worked on, I've despised unbalanced inputs and passive mixers. Cheap crystal microphones, on long single conductor shielded cables that picked up radio stations, and police cars were common in the '50s and '60s. > > I have worked with many broadcast and studio consoles, from Mono, eight input tube based Gates from the '40s to state of the art 40 input stereo consoles with six output busses. They all used transformer inputs and outputs, which allowed balanced or unbalanced inputs. Early equipment as 600/150 ohm, selected by using half or all of a winding at the connections. > > All inputs were line or mic level. Turntables used external magnetic phono preamps. > > At one very remote site, I had to use a spare mag phono preamp to bring the old WWII vintage telephone circuit up to an acceptable level. There was over 30dB of loss in the last mile of old cable, but the telephone people couldn't wrap their minds around anything without a dial and a bell. >
Ok but DC blocking caps on the mixing stage here are superfluous. You could provide the mid-point bias thru a buffered +6 reference to the tranny center taps and eliminate the caps and the split PSU entirely. maybe DC servo them with respect to the op-amp output to keep it at the mid-point. Once the blocking cap on the op amp output is gone you can bring the attenuator impedance down. There's no good reason to use such a high-value attenuator after a low-impedance buffered output it just adds noise.
On 10/6/19 9:56 PM, bitrex wrote:

>>> And if there is a split rail already available why are the op-amps >>> non-inverting inputs biased thru a divider and decoupling cap? >>> >>> The op amp stages seem to just be there so 4 (8!) transformers can be >>> used, one for each input, and the outputs of the transformer summed >>> independently without signal loss but other than that it and the 50k >>> attenutor just add noise. if they just used one transformer and a relay >>> to switch the inputs you could use a passive attenuator of lower value, >>> plenty of drive available from the line level signal and the TDA amp has >>> plenty of gain. >> >> If you used a relay to select an input it wouldn't be a mixer. The >> Opamps are used as summing amplifiers to compensate for the resistive >> losses for each input. That allows higher value potentiometers, and >> near zero interaction between sources. Have you ever examined the >> design of a commercial audio console? > > Ok right you are. The "mixer" part seemed to fade from my mind... > >> &nbsp;&nbsp;&nbsp; Split power supplies and what you consider as wasted parts are >> used to produce the required flatness and isolation along with the >> lowest overall system noise. Even early tube mixing boards used >> transformer inputs, and the fades had to provide at least 80dB of >> attenuation at their lowest settings. Find some manuals and study >> them. The crap circuits used in guitar amps are useless for real audio >> processing. You can do balanced inputs or outputs with just opamps, >> but transformers eliminate ground loops which induce hum. A capacitor >> between a transformer and an active circuit prevent it from affecting >> the biasing of the input or output stages. > > It's just there for a single TL0x dual op amp it has 100dB of PSU > rejection. The TEA amp is single supply. A BU634P to make a virtual > ground for 1 dual op amp seems like over-kill any way you cut it...
If you're gonna use a jellybean 20 cent TL-series dual op amp as your mixer not much point worrying about having a super low-noise audiophile-quality split power supply here for 'em, y'know
On Sunday, October 6, 2019 at 9:56:46 PM UTC-4, bitrex wrote:
> On 10/6/19 9:34 PM, Michael Terrell wrote: > > On Sunday, October 6, 2019 at 8:55:54 PM UTC-4, bitrex wrote: > >> On 10/6/19 11:28 AM, jlarkin wrote: > >>> On Sun, 6 Oct 2019 14:54:39 -0000 (UTC), Don Kuenz, KB7RPU wrote: > >>> > >>>> Greetings, > >>>> > >>>> An article about my stereo mixer-amplifier project is now available. > >>>> > >>>> http://crcomp.net/electronic/mixeramp/index.php > >>>> > >>>> Thank you, 73, > >>> > >>> The power supplies look strange to me. Why bias up the + inputs of U2 > >>> and U3? > >>> > >>> Why so many DC block caps after the transformers? Why any? > >>> > >>> Why split the 12 volt supply at all? Looks like that adds complexity > >>> and noise. > >>> > >>> > >>> > >> > >> And if there is a split rail already available why are the op-amps > >> non-inverting inputs biased thru a divider and decoupling cap? > >> > >> The op amp stages seem to just be there so 4 (8!) transformers can be > >> used, one for each input, and the outputs of the transformer summed > >> independently without signal loss but other than that it and the 50k > >> attenutor just add noise. if they just used one transformer and a relay > >> to switch the inputs you could use a passive attenuator of lower value, > >> plenty of drive available from the line level signal and the TDA amp has > >> plenty of gain. > > > > If you used a relay to select an input it wouldn't be a mixer. The Opamps are used as summing amplifiers to compensate for the resistive losses for each input. That allows higher value potentiometers, and near zero interaction between sources. Have you ever examined the design of a commercial audio console? > > Ok right you are. The "mixer" part seemed to fade from my mind... > > > Split power supplies and what you consider as wasted parts are used to produce the required flatness and isolation along with the lowest overall system noise. Even early tube mixing boards used transformer inputs, and the fades had to provide at least 80dB of attenuation at their lowest settings. Find some manuals and study them. The crap circuits used in guitar amps are useless for real audio processing. You can do balanced inputs or outputs with just opamps, but transformers eliminate ground loops which induce hum. A capacitor between a transformer and an active circuit prevent it from affecting the biasing of the input or output stages. > > It's just there for a single TL0x dual op amp it has 100dB of PSU > rejection. The TEA amp is single supply. A BU634P to make a virtual > ground for 1 dual op amp seems like over-kill any way you cut it... > > > Some of the first PA systems I worked on, I've despised unbalanced inputs and passive mixers. Cheap crystal microphones, on long single conductor shielded cables that picked up radio stations, and police cars were common in the '50s and '60s. > > > > I have worked with many broadcast and studio consoles, from Mono, eight input tube based Gates from the '40s to state of the art 40 input stereo consoles with six output busses. They all used transformer inputs and outputs, which allowed balanced or unbalanced inputs. Early equipment as 600/150 ohm, selected by using half or all of a winding at the connections. > > > > All inputs were line or mic level. Turntables used external magnetic phono preamps. > > > > At one very remote site, I had to use a spare mag phono preamp to bring the old WWII vintage telephone circuit up to an acceptable level. There was over 30dB of loss in the last mile of old cable, but the telephone people couldn't wrap their minds around anything without a dial and a bell. > > > > Ok but DC blocking caps on the mixing stage here are superfluous. You > could provide the mid-point bias thru a buffered +6 reference to the > tranny center taps and eliminate the caps and the split PSU entirely. > maybe DC servo them with respect to the op-amp output to keep it at the > mid-point. > > Once the blocking cap on the op amp output is gone you can bring the > attenuator impedance down. There's no good reason to use such a > high-value attenuator after a low-impedance buffered output it just adds > noise.
It isn't studio or broadcast oriented. It simply allows him to use one set of speakers for four devices at once. Since he is an Amateur radio operator, he has to consider RF ingression. Design something better and post it, if you can.
On 10/6/19 10:01 PM, Michael Terrell wrote:
> On Sunday, October 6, 2019 at 9:56:46 PM UTC-4, bitrex wrote: >> On 10/6/19 9:34 PM, Michael Terrell wrote: >>> On Sunday, October 6, 2019 at 8:55:54 PM UTC-4, bitrex wrote: >>>> On 10/6/19 11:28 AM, jlarkin wrote: >>>>> On Sun, 6 Oct 2019 14:54:39 -0000 (UTC), Don Kuenz, KB7RPU wrote: >>>>> >>>>>> Greetings, >>>>>> >>>>>> An article about my stereo mixer-amplifier project is now available. >>>>>> >>>>>> http://crcomp.net/electronic/mixeramp/index.php >>>>>> >>>>>> Thank you, 73, >>>>> >>>>> The power supplies look strange to me. Why bias up the + inputs of U2 >>>>> and U3? >>>>> >>>>> Why so many DC block caps after the transformers? Why any? >>>>> >>>>> Why split the 12 volt supply at all? Looks like that adds complexity >>>>> and noise. >>>>> >>>>> >>>>> >>>> >>>> And if there is a split rail already available why are the op-amps >>>> non-inverting inputs biased thru a divider and decoupling cap? >>>> >>>> The op amp stages seem to just be there so 4 (8!) transformers can be >>>> used, one for each input, and the outputs of the transformer summed >>>> independently without signal loss but other than that it and the 50k >>>> attenutor just add noise. if they just used one transformer and a relay >>>> to switch the inputs you could use a passive attenuator of lower value, >>>> plenty of drive available from the line level signal and the TDA amp has >>>> plenty of gain. >>> >>> If you used a relay to select an input it wouldn't be a mixer. The Opamps are used as summing amplifiers to compensate for the resistive losses for each input. That allows higher value potentiometers, and near zero interaction between sources. Have you ever examined the design of a commercial audio console? >> >> Ok right you are. The "mixer" part seemed to fade from my mind... >> >>> Split power supplies and what you consider as wasted parts are used to produce the required flatness and isolation along with the lowest overall system noise. Even early tube mixing boards used transformer inputs, and the fades had to provide at least 80dB of attenuation at their lowest settings. Find some manuals and study them. The crap circuits used in guitar amps are useless for real audio processing. You can do balanced inputs or outputs with just opamps, but transformers eliminate ground loops which induce hum. A capacitor between a transformer and an active circuit prevent it from affecting the biasing of the input or output stages. >> >> It's just there for a single TL0x dual op amp it has 100dB of PSU >> rejection. The TEA amp is single supply. A BU634P to make a virtual >> ground for 1 dual op amp seems like over-kill any way you cut it... >> >>> Some of the first PA systems I worked on, I've despised unbalanced inputs and passive mixers. Cheap crystal microphones, on long single conductor shielded cables that picked up radio stations, and police cars were common in the '50s and '60s. >>> >>> I have worked with many broadcast and studio consoles, from Mono, eight input tube based Gates from the '40s to state of the art 40 input stereo consoles with six output busses. They all used transformer inputs and outputs, which allowed balanced or unbalanced inputs. Early equipment as 600/150 ohm, selected by using half or all of a winding at the connections. >>> >>> All inputs were line or mic level. Turntables used external magnetic phono preamps. >>> >>> At one very remote site, I had to use a spare mag phono preamp to bring the old WWII vintage telephone circuit up to an acceptable level. There was over 30dB of loss in the last mile of old cable, but the telephone people couldn't wrap their minds around anything without a dial and a bell. >>> >> >> Ok but DC blocking caps on the mixing stage here are superfluous. You >> could provide the mid-point bias thru a buffered +6 reference to the >> tranny center taps and eliminate the caps and the split PSU entirely. >> maybe DC servo them with respect to the op-amp output to keep it at the >> mid-point. >> >> Once the blocking cap on the op amp output is gone you can bring the >> attenuator impedance down. There's no good reason to use such a >> high-value attenuator after a low-impedance buffered output it just adds >> noise. > > > It isn't studio or broadcast oriented. It simply allows him to use one set of speakers for four devices at once. Since he is an Amateur radio operator, he has to consider RF ingression. Design something better and post it, if you can. >
If he's worried about RF ingression there should probably be something in the circuit that would help with that.