Electronics > Projects, Designs, and Technical Stuff

Mixer - Schematic

<< < (3/5) > >>

nick_d:
Since there are some objections in regard to the XLR output, I decided to sketch out the earlier described ideas to explain better what I meant. Again, these ideas are ONLY for discussion, I don't necessarily recommend changing to this, since OP's circuit is really good as far as I can see.

I agree that 5V is too low, I checked and XLR equipment uses much higher levels, I didn't know that. I sketched it with 10V, it can be increased.

These are SINGLE-SUPPLY circuits, I have never built anything with dual supplies and I have never found it necessary to do so, with modern parts. At worst you might use an extra op-amp to generate a reference, here 5V. By the way, "virtual earth" is not the correct term for the 5V reference, "virtual earth" is the -ve terminal of an op-amp in inverting configuration, because the op-amp maintains it equal to the +ve terminal.

The first file shows some pieces: an input stage with 20Hz..20kHz band limiting and volume control, a buffered reference, and a "true" XLR output stage which drives the XLR connector differentially. The second file shows a more controversial approach, the "fake" XLR output stage which drives the XLR connector single-endedly, and provides a dummy -ve output. To show why the "fake" XLR output stage works, and works the same as the "true" XLR output stage, I've drawn also a sample of an input stage from another piece of equipment (a differential amplifier).

I personally cannot see any reason to prefer the "true" XLR output stage over the "fake" output stage which uses one fewer op-amp, except that the "true" XLR output stage will be twice as loud. If you want to reduce the supply to the minimum possible (about 5V) then use the "true" XLR. Although, I am keen to be educated if there is something I have missed in my analysis. I have built these circuits several times, but I would by no means call myself an audio or electronics expert, and I also don't have any XLR equipment apart from a collection of dynamic microphones.

As has been noted in this thread, creating an output that can be connected to an XLR microphone input with phantom power is a bit of a hassle. The output absolutely must be high-pass filtered, which is a shame because otherwise the output filter can be omitted. The output filter is bulky and expensive, due to the need for a low-value resistor for decent output drive, and corresponding high-value capacitor to get an RC time-constant of 50ms, for the 20 Hz high-pass cutoff that I have chosen (although you could get away with a higher cutoff in some applications).

As a compromise, I've specified a 50uF BP electrolytic with a 1k ohm resistor. When the 1k ohm feeds the 6.8k ohm resistor in the 48V phantom power circuit, the result will be a 1 / (50u * 7.8k) = 2.6 Hz high-pass cutoff and a signal level of 6.8k / 7.8k = 0.87 what it should be. On the other hand, it could rarely be 12V phantom power with a 680 ohm resistor, giving 11.9 Hz cutoff and a signal level of only 0.40. Hmm. The 50uF is conservative, but I want the cutoff well below 20 Hz, because the other device also has a high-pass and I don't want to be 6dB down at 20 Hz.

The volume control is also slightly tricky. I've specified a 1k ohm pot, because it's feeding a summing junction via a 50k ohm resistor, and I don't want the 50k ohm to load down the volume control pot. A factor of 50 is probably acceptable. You can also do interesting things here, I did the calculations a while ago and I cannot exactly recall how, but you can use the loading-down of the volume control pot to get an almost logarithmic response. If anybody wants me to re-design that circuit (I was doing it with a 50k digital pot) I will try to do so. Or, a larger value logarithmic pot can be used with a further op-amp as a buffer. The added buffer can either be a follower, or it can have the volume control in its feedback loop.

The volume control requires a buffered reference for the "ground" end of the pot. So I used the buffered reference everywhere, in the first diagram. The second diagram is only an output stage, so I used a dedicated 100k/100k divider instead of a shared, buffered, reference. Because of this, and because the 50k resistors used everywhere must be exactly half of the 100k used in the reference divider, I've specified 50k not 47k.

cheers, Nick

Jbliss:

--- Quote from: nick_d on December 18, 2018, 03:33:47 am ---Since there are some objections in regard to the XLR output, I decided to sketch out the earlier described ideas to explain better what I meant. Again, these ideas are ONLY for discussion, I don't necessarily recommend changing to this, since OP's circuit is really good as far as I can see.

I agree that 5V is too low, I checked and XLR equipment uses much higher levels, I didn't know that. I sketched it with 10V, it can be increased.

These are SINGLE-SUPPLY circuits, I have never built anything with dual supplies and I have never found it necessary to do so, with modern parts. At worst you might use an extra op-amp to generate a reference, here 5V. By the way, "virtual earth" is not the correct term for the 5V reference, "virtual earth" is the -ve terminal of an op-amp in inverting configuration, because the op-amp maintains it equal to the +ve terminal.

The first file shows some pieces: an input stage with 20Hz..20kHz band limiting and volume control, a buffered reference, and a "true" XLR output stage which drives the XLR connector differentially. The second file shows a more controversial approach, the "fake" XLR output stage which drives the XLR connector single-endedly, and provides a dummy -ve output. To show why the "fake" XLR output stage works, and works the same as the "true" XLR output stage, I've drawn also a sample of an input stage from another piece of equipment (a differential amplifier).

I personally cannot see any reason to prefer the "true" XLR output stage over the "fake" output stage which uses one fewer op-amp, except that the "true" XLR output stage will be twice as loud. If you want to reduce the supply to the minimum possible (about 5V) then use the "true" XLR. Although, I am keen to be educated if there is something I have missed in my analysis. I have built these circuits several times, but I would by no means call myself an audio or electronics expert, and I also don't have any XLR equipment apart from a collection of dynamic microphones.

As has been noted in this thread, creating an output that can be connected to an XLR microphone input with phantom power is a bit of a hassle. The output absolutely must be high-pass filtered, which is a shame because otherwise the output filter can be omitted. The output filter is bulky and expensive, due to the need for a low-value resistor for decent output drive, and corresponding high-value capacitor to get an RC time-constant of 50ms, for the 20 Hz high-pass cutoff that I have chosen (although you could get away with a higher cutoff in some applications).

As a compromise, I've specified a 50uF BP electrolytic with a 1k ohm resistor. When the 1k ohm feeds the 6.8k ohm resistor in the 48V phantom power circuit, the result will be a 1 / (50u * 7.8k) = 2.6 Hz high-pass cutoff and a signal level of 6.8k / 7.8k = 0.87 what it should be. On the other hand, it could rarely be 12V phantom power with a 680 ohm resistor, giving 11.9 Hz cutoff and a signal level of only 0.40. Hmm. The 50uF is conservative, but I want the cutoff well below 20 Hz, because the other device also has a high-pass and I don't want to be 6dB down at 20 Hz.

The volume control is also slightly tricky. I've specified a 1k ohm pot, because it's feeding a summing junction via a 50k ohm resistor, and I don't want the 50k ohm to load down the volume control pot. A factor of 50 is probably acceptable. You can also do interesting things here, I did the calculations a while ago and I cannot exactly recall how, but you can use the loading-down of the volume control pot to get an almost logarithmic response. If anybody wants me to re-design that circuit (I was doing it with a 50k digital pot) I will try to do so. Or, a larger value logarithmic pot can be used with a further op-amp as a buffer. The added buffer can either be a follower, or it can have the volume control in its feedback loop.

The volume control requires a buffered reference for the "ground" end of the pot. So I used the buffered reference everywhere, in the first diagram. The second diagram is only an output stage, so I used a dedicated 100k/100k divider instead of a shared, buffered, reference. Because of this, and because the 50k resistors used everywhere must be exactly half of the 100k used in the reference divider, I've specified 50k not 47k.

cheers, Nick

--- End quote ---


Hi Nick, Once again thank you! I am revising my schematic as I type this, implementing changes. Thinking of raising the low pass filter on the input to above 20KHz I know it realistically makes no audible difference would just like to keep the box transparent as possible. What comment values would raise the filter to 30KHz

Jbliss:

--- Quote from: Sylvi on December 18, 2018, 02:06:31 am ---Hi

I haven't read the whole thread but looked at the schematic.

It might be a mistake to not have AC coupling between the input pots and the virtual-earth node. Either add a single cap between the junction of the input weighting resistors and the opamp input, or add individual caps for each wiper. otherwise, there might be a scratchy pot issue.

I would leave the EQ stage intact to correct for the inversion of the mixing stage. You can make the EQ "bypassable" simply by adding two equal-value resistors give the stage unity-gain (inverting). The EQ network of the pots, Rs and Cs is still connected at the input and output of the circuit, but a switch is added between the opamp -in and the wiper resistors from the controls. This works very well and is not noisy if there is a high-value resistance across the switch.

The schemo does not show what the output driver chip is. Going by the functional names for the outputs, I suspect that "sense" should be tied to the output side of the cap rather than to the chip-side of the cap. I assume the chip is supposed to provide a "zero z-out" this way -  technique I a always suspicious of as it contradicts the purpose of the cap.

--- End quote ---



Hi Sylvi, Thanks for you help. Great idea regarding bypass Switch would you be able to draw a quick schematic. Would like to make sure I understood.

Thanks JBliss

Jbliss:

--- Quote from: nick_d on December 18, 2018, 03:33:47 am ---Since there are some objections in regard to the XLR output, I decided to sketch out the earlier described ideas to explain better what I meant. Again, these ideas are ONLY for discussion, I don't necessarily recommend changing to this, since OP's circuit is really good as far as I can see.

I agree that 5V is too low, I checked and XLR equipment uses much higher levels, I didn't know that. I sketched it with 10V, it can be increased.

These are SINGLE-SUPPLY circuits, I have never built anything with dual supplies and I have never found it necessary to do so, with modern parts. At worst you might use an extra op-amp to generate a reference, here 5V. By the way, "virtual earth" is not the correct term for the 5V reference, "virtual earth" is the -ve terminal of an op-amp in inverting configuration, because the op-amp maintains it equal to the +ve terminal.

The first file shows some pieces: an input stage with 20Hz..20kHz band limiting and volume control, a buffered reference, and a "true" XLR output stage which drives the XLR connector differentially. The second file shows a more controversial approach, the "fake" XLR output stage which drives the XLR connector single-endedly, and provides a dummy -ve output. To show why the "fake" XLR output stage works, and works the same as the "true" XLR output stage, I've drawn also a sample of an input stage from another piece of equipment (a differential amplifier).

I personally cannot see any reason to prefer the "true" XLR output stage over the "fake" output stage which uses one fewer op-amp, except that the "true" XLR output stage will be twice as loud. If you want to reduce the supply to the minimum possible (about 5V) then use the "true" XLR. Although, I am keen to be educated if there is something I have missed in my analysis. I have built these circuits several times, but I would by no means call myself an audio or electronics expert, and I also don't have any XLR equipment apart from a collection of dynamic microphones.

As has been noted in this thread, creating an output that can be connected to an XLR microphone input with phantom power is a bit of a hassle. The output absolutely must be high-pass filtered, which is a shame because otherwise the output filter can be omitted. The output filter is bulky and expensive, due to the need for a low-value resistor for decent output drive, and corresponding high-value capacitor to get an RC time-constant of 50ms, for the 20 Hz high-pass cutoff that I have chosen (although you could get away with a higher cutoff in some applications).

As a compromise, I've specified a 50uF BP electrolytic with a 1k ohm resistor. When the 1k ohm feeds the 6.8k ohm resistor in the 48V phantom power circuit, the result will be a 1 / (50u * 7.8k) = 2.6 Hz high-pass cutoff and a signal level of 6.8k / 7.8k = 0.87 what it should be. On the other hand, it could rarely be 12V phantom power with a 680 ohm resistor, giving 11.9 Hz cutoff and a signal level of only 0.40. Hmm. The 50uF is conservative, but I want the cutoff well below 20 Hz, because the other device also has a high-pass and I don't want to be 6dB down at 20 Hz.

The volume control is also slightly tricky. I've specified a 1k ohm pot, because it's feeding a summing junction via a 50k ohm resistor, and I don't want the 50k ohm to load down the volume control pot. A factor of 50 is probably acceptable. You can also do interesting things here, I did the calculations a while ago and I cannot exactly recall how, but you can use the loading-down of the volume control pot to get an almost logarithmic response. If anybody wants me to re-design that circuit (I was doing it with a 50k digital pot) I will try to do so. Or, a larger value logarithmic pot can be used with a further op-amp as a buffer. The added buffer can either be a follower, or it can have the volume control in its feedback loop.

The volume control requires a buffered reference for the "ground" end of the pot. So I used the buffered reference everywhere, in the first diagram. The second diagram is only an output stage, so I used a dedicated 100k/100k divider instead of a shared, buffered, reference. Because of this, and because the 50k resistors used everywhere must be exactly half of the 100k used in the reference divider, I've specified 50k not 47k.

cheers, Nick

--- End quote ---



Hi All,

Here is my latest working schematic with a bunch of changes implemented Let me know what you think!
This schematic does not include output filtering, forgot was late last night.
(I have deiced to keep the DRV-134 (line Driver) in the output as I have a bunch lying round)

Jbliss:
Also Let me know what you think of the new PS circuit ?
and input section.

Navigation

[0] Message Index

[#] Next page

[*] Previous page

There was an error while thanking
Thanking...
Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod