A little console to simplify tracking, somewhere between mixer and monitor controller

[Chrisfromthepast]

Oh thats tidy
Buffering the signal is a good idea, that corp chip seems expensive in this role, but, thats not all the input conditioning that needs to be done, right? The output of the 1200 still swings above and below 0v?
Are you doing anything to rectify and integrate?
Maybe Im muddying the waters with my ignorance, but
I thought you needed the obsolete lm3915 to get the range you printed on your panel.

If what you have built works great, please tell me so I stop with the nonsense.
If you have a hard time calibrating, there is a pic16f1503 package that could still fit inside your board outline with some minor changes to nets.
Flashing PICs isn’t too hard, with this binary designed to replace the LM chips
https://electricdruid.net/led-bargraph-audio-level-display/

 

[Dave_Eropkin]

Oh thats tidy
Buffering the signal is a good idea, that corp chip seems expensive in this role, but, thats not all the input conditioning that needs to be done, right? The output of the 1200 still swings above and below 0v?
Are you doing anything to rectify and integrate?
Maybe Im muddying the waters with my ignorance, but
I thought you needed the obsolete lm3915 to get the range you printed on your panel.

If what you have built works great, please tell me so I stop with the nonsense.
If you have a hard time calibrating, there is a pic16f1503 package that could still fit inside your board outline with some minor changes to nets.
Flashing PICs isn’t too hard, with this binary designed to replace the LM chips
https://electricdruid.net/led-bargraph-audio-level-display/[/URL

I'll let you know when I setup the board. Just about everything I do takes a lot of work - I tend to miss small details and then beat my head on the wall till I figure it out! ha ha. Good to put stuff on here to have other eyes on it, rather than doing a bunch of iterations with no oversight.

 

[Dave_Eropkin]

Oh thats tidy
Buffering the signal is a good idea, that corp chip seems expensive in this role, but, thats not all the input conditioning that needs to be done, right? The output of the 1200 still swings above and below 0v?
Are you doing anything to rectify and integrate?
Maybe Im muddying the waters with my ignorance, but
I thought you needed the obsolete lm3915 to get the range you printed on your panel.

If what you have built works great, please tell me so I stop with the nonsense.
If you have a hard time calibrating, there is a pic16f1503 package that could still fit inside your board outline with some minor changes to nets.
Flashing PICs isn’t too hard, with this binary designed to replace the LM chips
https://electricdruid.net/led-bargraph-audio-level-display/

I believe the LM3914 does have a diode, to shunt the negative part of the waveform to ground, along with a buffer built in, unless I'm missing something. It's on page 8 of the data sheet. I am going to have to adjust the input to account for the higher dB levels though. I think my original intent was to use a LM3916, which is obsolete now. I'll have to go back to the drawing board on reducing the input signal to match the 5V max display or change the scale.

 

[Chrisfromthepast]

I think you need a log amplifier before the lm3914 to get the readings you are printing on your front panel. Those require rectified audio.
I think you should still definitely offer the transformer secondary a high impedance buffer. However, with a conventional dual op amp you could knock out all of the buffering, rectification, and log amp in the same sized package, and not affect the audio coming out of the other tap of the transformer.

Its a shame the 3915 isn’t being made by TI, because they were able to control the behavior of the log amp within the feedback loop of that on board buffer in the picture you attached. I think
(Edit) looking at the lm3915 datasheet, the log calculation is not done in the feedback loop, but in the voltage divider network between comparators.

 

[Dave_Eropkin]

I think you need a log amplifier before the lm3914 to get the readings you are printing on your front panel. Those require rectified audio.
I think you should still definitely offer the transformer secondary a high impedance buffer. However, with a conventional dual op amp you could knock out all of the buffering, rectification, and log amp in the same sized package, and not affect the audio coming out of the other tap of the transformer.

Its a shame the 3915 isn’t being made by TI, because they were able to control the behavior of the log amp within the feedback loop of that on board buffer in the picture you attached. I think

Yeah - it's a shame they don't make it anymore. The reason I went to the 3914 was not wanting to use obsolete components etc... but ended up causing more issues. I'm thinking of scraping that version and building one using a couple of LM339N's - which could be used to essentially re-make the 3915's or 16's and still have components that are able to be replaced if they fail etc. I was also looking at the log amp today before the LM3914, seems like I'd need to do an inverse log setup to make the log values that I want to display come out at set intervals into the LM3914 so that it could be scaled properly. I appreciate the second set of eyes - although it wasn't really what this post was originally about, I have zero people to double check anything I do! ha ha.

 

[Dave_Eropkin]

I think you need a log amplifier before the lm3914 to get the readings you are printing on your front panel. Those require rectified audio.
I think you should still definitely offer the transformer secondary a high impedance buffer. However, with a conventional dual op amp you could knock out all of the buffering, rectification, and log amp in the same sized package, and not affect the audio coming out of the other tap of the transformer.

Its a shame the 3915 isn’t being made by TI, because they were able to control the behavior of the log amp within the feedback loop of that on board buffer in the picture you attached. I think
(Edit) looking at the lm3915 datasheet, the log calculation is not done in the feedback loop, but in the voltage divider network between comparators.

Ok, getting much closer here. Used the LM339N's to make it an 8 LED VU meter. I'll bread board this in the next few weeks and get it finalized. I liked the flexibility of this method much better, and looks like they aren't going to discontinue these IC's anytime soon. Still need to drop the THAT 1200 and use a cheaper op-amp. I had some of those lying around so I figured I'd just use them, but they are a bit expensive for making multiple units etc. Thanks for the input and feedback on this PCB/Circuit - made some good progress.

 

[Chrisfromthepast]

Im not a professional EE, but it looks pretty standard.
Im impressed by your layout speed, so why not add 2 more cents?
In my mixing experience, its important to have a -30db signal present led and a clip led that stays lit for a second or three, so you can see what clipped without watching meters all day.
I can see how this could exceed your design requirements. If you were only worried about clipping the daw inputs, and you can see the low levels on your computer screen, these features could be unnecessary.

 

[ruffrecords]

I am concerned about the position of D2. I think is needs to be after D1 not before it. In its present position it could add significant distortion to the THAT output.

Cheers

Ian

 

[Mixermend]

Do have a look at the LED metering circuits in some commercial mixers (Mackie and A&H for example) - then consider building a separate precision rectifier stage and you can also then add the time constants to give the required VU 'ballistics'.

Most commercial products like to use the LM339 comparator - but drive the LED's in series from a constant current source - this avoids the change in current demand as each LED turns on - which can cause clicks in the audio path.

 

[bjosephs]

This project is back on my mind. I went off and built that monitor controller but never filled the mixer role and realized that after 10 years of using it, I just hate the digital mixer on my interface for adjusting the monitoring of inputs for any purpose. I need a knob. The good news is that I upgraded my computer since the original post so getting low latency "wet" signal isn't as impossible as it once was but can still be a problem late in production.

New vision: 4-8 channels of input with level and pan (potentially LCR), 1 or 2 stereo returns for main mix at the very least maybe also with a level control but maybe not. Some way to route signals at the mixer to recording inputs (either sends on each channel or just a normalled direct out on each channel.

As I said so long ago I never record more than 2 channels at once but I done have 6 good analog inputs so maybe I just split them at the input and hardwire. It's that or I have some sort of selection facilities too route a channel to record output which is also appealing because I'm a fan of some multimic guitar recording techniques that blend signals right after the preamp. I could blend them in this box.

Is it a terrible idea to have an insert at the input and/or direct out just be passive switching? I think it's fair these days to trust that the source (a preamp, synth, modeler, etc) can drive a little extra cabling for an insert or a the direct out in parallel with the fader amp, right?

 

[bjosephs]

Quick Block Diagram of what I'm working out.

• Input jack, a switchable insert that can be used as split point (trying to avoid patch bay), then direct out. This whole section will be differential. Perhaps this is all passive or I will at minimum buffer the input before the send jack.
• The fader amp will be a balanced to unbalanced stage configured for some gain. I have a schematic going showing an OPA134 but it may be smarter to use a dedicated IC for the (INA137 can be wired for +6dB). Neither can be expected to drive the load though because...
• I'm playing with the idea of using a stacked pot for level/pan. CAPI sells a 1k/dual10k concentric pot that is small and will save space. Or I could use a larger level resistance and buffer again for the pan. Trying to keep the active stages down.
• One feature I'm trying to work out is an additional routing option like a second bus or two channel "record out" bus. As drawn I need an interface input for every source to be able to select them for recording in DAW. I'm not sure my next interface will be more than 2 channels and being able to route an input to the record output as needed seems appealing. Maybe a 3/4 bus switch on each channel, maybe it's 2 mono sends? Maybe something in between. Not sure, looking for advise and I can break it down if that's not clear.

 

[bjosephs]

And rough draft and incomplete schematic. That lower section is my currently ill defined alternate routing option. As drawn each mono channel could be optionally routed to 1 of 2 record outs. This method lacks level controls like a send would have but it could be used to, say, sub mix down sources (like guitar cabinet mics).