Unbalanced Mixer Direct Out convert to Balanced

[Jackie The Shepherd]

At the risk of confusing things further - I don’t think this has been said so far - don’t worry about levels when plugging your mixer directly into your audio interface using TS-TS cables. Although it might seem that you are plugging a line level source into a mic input, they are designed to cope with this. Plugging a jack into a combo socket should pad the signal down.

As many have said - I would try the TS-TS cable option first, and deal with any issues that arise. I have done pretty much what you are attempting in the past (taking direct outs from a behringer mixer into an RME interface when recording a live concert) and had no issues at all.

Yes that's what I said in my post. But I also said, irrespective of what direction I will finally take, and irrespective of negative opinions about the DI800, I would like to know whether someone can clarify my doubt on a technicality. See below please.

Connecting the unbalanced console direct outputs into a DI800 and thence to your interface mic inputs will work

This is where I am stuck, because of a dichotomy between what really a DI Box does, i.e. attenuates the input after processing, to a balanced mic level signal (i.e. +4 dBu?), to be fed into the mixer XLRs and what the DI800 specs tell me that the max balanced output level is @+23dBu.

I think only one of this can be true, not both. Am I right? And if +4 dBU is what it is, then DI800 is not good because then as you earlier pointed out as well, I would need to increase the gain to restore to acceptable levels. So then the figure of maximum output level of +23 dBU means what?

Sorry for going round in circles, but it is evident to you all that evidently I am worse than a noob

Shall appreciate helping me to clearly understand this aspect, because notwithstanding what option I finally decide to employ for my situation, it would be dangerous to act based on half-baked or zero knowledge.

Thanks

EDIT: I also noticed that there is a +20 dB switch which increases the INPUT LEVEL SIGNAL by 20 dB. Again, how useful would that be, if anyway the output is going to be only at mic level? Or, is it because of this feature that the manual states max output level @23 dB balanced, after adding 20 dB to say +3dB for mic level? Pardon me for all of this! Need help.

[Hugh Robjohns]

This is where I am stuck, because of a dichotomy between what really a DI Box does, i.e. attenuates the input after processing, to a balanced mic level signal (i.e. +4 dBu?), to be fed into the mixer XLRs and what the DI800 specs tell me that the max balanced output level is @+23dBu.

Looking into the DI800 a bit more it seems to be a very non-standard 'DI box'.

For starters, it doesn't use a step-down /isolation transformer at all. Like the Behringer DI120, it is an entirely active device.

I found the DI120 schematics and it appears that both its input buffer and active balanced output stages have unity-gain, and are built using cooking-grade op-amps. I suspect the DI800's circuitry is very similar.

The Behringer specs suggest the DI800 can deliver up to +23dBu at its output, which implies it's intended as a line level device... yet the manual talks about connecting to mic inputs. This strange gain structure might explain the noise complaints associated with these devices.

In reality, I suspect the design simply relies on most instrument sources typucally being around -20dBu and, by outputting much the same level, hopes the destination mic inputs have enough headroom to cope.

I've not tested it, so this is guesswork based on the specs, the DI120 schematics, and received feedback... but my impression is that if you put line level in, you'll get more or less line level out.... which is not what normal DI boxes do!

BTW, mic level can be anywhere between -70 and 0dBu, but most DI boxes and devices delivering a nominal mic output level typically aim for around -30dBu or so. High enough to not need much preamp gain, but low enough not to eat into preamp headroom.

+4dBu is the nominal line level standard reference.

Shall appreciate helping me to clearly understand this aspect, because notwithstanding what option I finally decide to employ for my situation, it would be dangerous to act based on half-baked or zero knowledge.

You're going to have to 'suck it and see', Im afraid. Test your system wiring the mixer's direct outs into the DI800, and the DI800 outs into the mic or line inputs of your interface to see which option works best.

It may well be that you can connect the DI800 outputs into the interface line inputs (rather than mic inputs), and that might give a better overall noise performance than going via mic inputs.

And by experimenting with the -30dB pad and +20dB gain switches you have some options for optimising the level too. (The combination allowing -30, -10, 0, or +20dB gain adjustments.)

But at the end of the day, you will have two or three cooking-grade op-amps in the DI800's signal path when you don't actually need any at all... something you can check simply by moving the inputs to the DI800 straight over to the interface line inputs.

[Jackie The Shepherd]

Redding your reply I’m now far more relaxed.

Well as I said, I’ll go direct to audio interface and if this helps, good.

As for the clarification on the confusing schematic of the Behringer, thanks a ton. May be, SOS should do a review of their V2 officially do that you can tear apart the device and see for yourself. Shall wait for that to happen.

For now I’ll close this chapter from my end. And oh yes, the ART order has been reactivated so there’s a good chance of me getting those two. I understood from some forums that they do add “good warm” colour but that’s not what we want, do we? I thought you said your have those. What do you have to say on this topic?

[Hugh Robjohns]

Behringer rarely choose to submit products for review, so I think it's unlikely I'll get one on my test bench anytime soon.

Paul White reviewed the T8 and said:

In practice the T8 sounded surprisingly transparent, and the only very very subtle difference that I could detect was that the bass end actually seemed slightly more solid via the T8! I find it remarkable that such a low-cost transformer-isolation system can work so effectively while retaining a commendably high degree of audio transparency. I have no doubt that more sophisticated transformers can provide a better technical specification, but I'm equally convinced that any audio deterioration produced here will be far less of a concern than the quality of the output electronics in much budget studio gear.

I think that's a fair assessment, especially the part I italicised.

Unless the transformers are driven at silly high levels most people won't hear any difference at all, and those who do generally like the very subtle bass warmth or thickening effect. As a rule, the benefits far outweigh the negatives.

I do have genuinely transparent line isolators with Lundahl transformers, but they cost £370 for two channels — ie. more than eight times the per-channel cost of the T8!

95% of the time the DTIs are more than good enough and I use them far more often!

[Jackie The Shepherd]

That sums it up. T8 it is.

[Jackie The Shepherd]

Gents, good morning to all.

Would like to know if there exists any guide or book or similar on patch bay applications, suggested routings like Direct Outs, Processors and which configs to choose, e.g. Open, Normalled, Half etc and when.

I understand that each setup will be different from the other, so all I am looking for is a generalised guideline. I did find some tips in the Behringer PX2000 and PX3000 manuals but they didn't sound very convincing, example they mention that Direct Outs should always be placed one below the other and in Open configuration, so all the time there will have to be a patch cord in the front and on both top and bottom rows, connected to the recorder / audio interface.

I also did go through some videos and literature but most of them deal with basics like what is normalled and kind of stuff. What I am looking for is application oriented guidelines.

Any leads would be helpful.
Thanks in advance.

[shufflebeat]

I found the DI120 schematics and it appears that both its input buffer and active balanced output stages have unity-gain, and are built using cooking-grade op-amps…

Ah, that explains the sound of frying eggs.

[Hugh Robjohns]

Would like to know if there exists any guide or book or similar on patch bay applications, suggested routings like Direct Outs, Processors and which configs to choose, e.g. Open, Normalled, Half etc and when.

I'm not aware of any specific books, but there are several articles in the SOS database about patchbays, such as:

https://www.soundonsound.com/techniques ... ern-studio

https://www.soundonsound.com/techniques ... -patchbays

https://www.soundonsound.com/techniques ... -patchbays

https://www.soundonsound.com/techniques ... -patchbays

https://www.soundonsound.com/techniques ... ide-part-2

https://www.soundonsound.com/techniques ... dio-part-5

But there are relatively few hard and fast rules about how you setup a patchbay. Everything depends on how you plan to use it and what needs to be accessible.

... but it helps if you have some degree of logical signal path in the way sockets are allocated rather than scattering sources and destinations randomly. That might be from left to right across acsingle strip, or bottom to top (or vice versa) in a big multi-panel patchbay.

Ideally, though, everything should work correctly with no cords plugged in. So if you expect your direct outs to feed interface inputs 99% of the time, arrange them in normalled vertical pairs with outputs above inputs (ie. Direct outs on the top row, interface inputs on the bottom row).

On the other hand, if the direct outs are only going to be used occasionally with no preferred destinations, you could bunch them together on a bunch of isolated sockets out of the way somewhere.

And careful planning is the key. Draw the patchbay on paper, allocate the sockets and normalling as you think best, and then work through as many scenarios and practi8cal applications as you can imagine to make sure you can overplug and patch to achieve what you want in every situation (and with the patch cord lengths you have!). If it doesn't work, modify and try again....

And don't forget relevant handy functions such as tie-lines to other areas of the room or studio, parallels (mults), polarity inverters, attenuators, tone sources etc... (depending on your specific requirements, of course).

[Jackie The Shepherd]

Would like to know if there exists any guide or book or similar on patch bay applications, suggested routings like Direct Outs, Processors and which configs to choose, e.g. Open, Normalled, Half etc and when.

But there are relatively few hard and fast rules about how you setup a patchbay. Everything depends on how you plan to use it and what needs to be accessible.

... but it helps if you have some degree of logical signal path in the way sockets are allocated rather than scattering sources and destinations randomly. That might be from left to right across acsingle strip, or bottom to top (or vice versa) in a big multi-panel patchbay.

A question that comes to mind therefore is, how to deal with inserts. Say I have 4 dynamics processors but an entire array of inserts on my mixer. Ideally, how should I be connecting, how do inserts connect? Top and bottom rows I guess. But then what about the mode, ie normal or half normal? I read that this should be half normal but didn’t understand why and why not normalled. An explanation would help.

[Hugh Robjohns]

It's all detailed in those articles I linked.

If you don't want anything to live semi-permabently in specific channel inserts, then connect insert out to the top row, corresponding insert in to the bottom row, and normal between them to maintain the signal path through the channel strip.

If you want to connect outboard semi-permanently to specific inserts, then you would have insert out on the top row normalled to outboard input on the lower row.

Then, on the next patchbay down you'd have outboard output on the top row normalled to the insert return on the bottom row.

This last approach makes sense if every channel has outboard connected, but can get messy if its only a few channels...

So the first option with patchcords to the equipment I/O elsewhere on the bay is probably clearer and easier to use.

[Jackie The Shepherd]

Would like to know if there exists any guide or book or similar on patch bay applications, suggested routings like Direct Outs, Processors and which configs to choose, e.g. Open, Normalled, Half etc and when.

I'm not aware of any specific books, but there are several articles in the SOS database about patchbays, such as:

https://www.soundonsound.com/techniques ... ern-studio

https://www.soundonsound.com/techniques ... -patchbays

https://www.soundonsound.com/techniques ... -patchbays

https://www.soundonsound.com/techniques ... -patchbays

https://www.soundonsound.com/techniques ... ide-part-2

https://www.soundonsound.com/techniques ... dio-part-5

But there are relatively few hard and fast rules about how you setup a patchbay. Everything depends on how you plan to use it and what needs to be accessible.

... but it helps if you have some degree of logical signal path in the way sockets are allocated rather than scattering sources and destinations randomly. That might be from left to right across acsingle strip, or bottom to top (or vice versa) in a big multi-panel patchbay.

Ideally, though, everything should work correctly with no cords plugged in. So if you expect your direct outs to feed interface inputs 99% of the time, arrange them in normalled vertical pairs with outputs above inputs (ie. Direct outs on the top row, interface inputs on the bottom row).

On the other hand, if the direct outs are only going to be used occasionally with no preferred destinations, you could bunch them together on a bunch of isolated sockets out of the way somewhere.

And careful planning is the key. Draw the patchbay on paper, allocate the sockets and normalling as you think best, and then work through as many scenarios and practi8cal applications as you can imagine to make sure you can overplug and patch to achieve what you want in every situation (and with the patch cord lengths you have!). If it doesn't work, modify and try again....

And don't forget relevant handy functions such as tie-lines to other areas of the room or studio, parallels (mults), polarity inverters, attenuators, tone sources etc... (depending on your specific requirements, of course).

The guidelines that you gave, have been of immense help. Thank you.

I have been able to figure out most of the basic configuration aspects, and detailed out each and every analogue audio I/O in my system.

What I need to now do is to put the information in a programme or excel file, so any templates or programme that can help me quickly achieve the interconnection chart, would be helpful. This way I can straightaway start cabling.

[jimjazzdad]

Spreadsheets are also a good way to make labels for your patch bay - I use MS Excel and, by playing with the cell size, aspect ratio and printer scaling, I am able to print professional-looking labels that I cut out and affix with a bit of rubber cement. Save the spreadsheet and its dead easy to print new labels when you reconfigure your patch bay (as you inevitably will).

[Hugh Robjohns]

What I need to now do is to put the information in a programme or excel file, so any templates or programme that can help me quickly achieve the interconnection chart, would be helpful. This way I can straightaway start cabling.

Not entirely clear what you're asking for here.

If a patchbay is designed correctly, everything should work without anything plugged in, thanks to normalisation.

If you need to 'over-plug' to achieve some alternative connections for a specific project the patching should be obvious from your requirements and the patchbay labelling.

I've routinely created installation block wiring diagrams for studios, identifying how equipment is wired together and via which patch points in advance of construction... but I've never used an 'interconnection chart' for day-to-day patching as it's never been necessary.

[Jackie The Shepherd]

So, the idea is to have a chart in front of me that can help me quickly patch between points depending upon the required connection being Open, Normal, Half-normal. May be I am unable to explain it any better, but I have read something called Patchbay Worksheets, so I assume they would mean that all my studio connections are on a piece of paper for ready reference.

[Hugh Robjohns]

So, the idea is to have a chart in front of me that can help me quickly patch between points depending upon the required connection being Open, Normal, Half-normal.

The labels above (or in some designs, below) each socket should be all the guidance you need.
You just need to know what source you want to feed to which destination, identify those sockets from their labels, and plug a chord in.

Some patchbays have front panel switches which indicate which pairs are normalled (etc) but traditional patchbays are configured semi-permanently internally. In those cases it can be useful to mark normalled or half-normalled sockets in red — either with coloured gel or highlighter, or by drawing red stripes or borders (different organisation have different policies).

It's also helpful in navigating a big patchbay to colour-code useful groups of sockets — like all direct outs, or mixer line ins, or whatever.

I have read something called Patchbay Worksheets, so I assume they would mean that all my studio connections are on a piece of paper for ready reference.

Fair enough... I've never come across one or needed one. I just look at the patchbay and plug as needed.

[ajay_m]

If you don't need ground isolation then an unbalanced to balanced convertor is simply two unity gain opamps, one inverting and one non inverting as in this circuit https://www.radioworld.com/tech-and-gea ... ed-adaptor

I believe you can purchase boards already made but the cost per channel is peanuts compared with a transformer based solution.

Modern opamps are a marvel, they can do rail to rail off a single 5v rail with vanishingly low noise figures too.

[Hugh Robjohns]

That's an 'interesting' design. Disappointing that it promotes the very obsolete 600 Ohm matched impedance output format (albeit as an option).

Also, running the first opamp at -40dB for a mic level output isn't going to do the overall SNR any favours. A switchable pad on the output would have been much better.

But the most heinous cardinal sin is the variable gain arrangement. If that wiper lifts off the track due to dust, dirt, or general failure on movement the opamp will go to maximum open-loop gain, which is definitely not good and its going to hurt!

[ajay_m]

To be honest I didn't take a close look at that design. I would personally use a quad fet input opamp and waste one opamp, have one side through a unity gain non inverting stage which will present essentially an infinite input impedance to the source, and take a second identical non inverting stage and feed through a unity gain inverting stage (which will of course present a much lower impedance) using fairly small resistors, say 10k.
Off the standard split 15v rails that'll swing in excess of +18dbu without going rail to rail. Noise and distortion should be negligible and if you're not worried about going right down to DC you don't even need any coupling caps. Any jellybean opamp ought to have slew rates at unity gain way beyond the audio passband. Njm2164 at a quid each for example.
Just put a couple of resistors across the inputs to present some practical impedance and prevent spurious signal pickup on inputs disconnected, maybe 10k I guess.
I know you could potentially at 10k impedance just use two opamps per channel and omit the unity gain buffer on the inverting side but CMRR is probably gonna be a little better if both sides present identical inputs in terms of capacitance in particular.

[Hugh Robjohns]

To be honest, if I need unbalanced to balanced conversion I use either the Sparkfun modules (for single channels):

https://www.soundonsound.com/reviews/sp ... out-boards

Or one of the stereo Ebay offerings such as:
https://www.ebay.co.uk/itm/39514260312 ... media=COPY

Both designs use one or other of the dedicated transforner-like driver chips which I find very versatile and effective, and which have extremely good specs.

[Matt Houghton]

the Sparkfun modules (for single channels):

https://www.soundonsound.com/reviews/sp ... out-boards

These are no longer current products, Hugh. (I checked recently when I had need, having remembered your review!)

But there's lots of this stuff on eBay, as you say... including some Spark Fun modules, in fact

[Hugh Robjohns]

These are no longer current products, Hugh. (I checked recently when I had need, having remembered your review!)

That's a real shame as they were so easy to use.

But there's lots of this stuff on eBay, as you say...

Yes, always people offering built PCBs.

For unbalanced to balanced converters, I look for boards using THAT1646 Outsmarts chip which seem to be the best option.

There are similar (pin-compatible) equivalents, though including the Analogue Devices SSM2142 or the Texas Instruments DRV134 and DRV135.

(There's also a THAT1606 outsmarts chip which is the same as the 1646 but has a differential input, intended for direct connection to D-A converter chips with differential outputs).

For balanced to unbalanced conversion, THAT offer their excellent InGenius 1200/1203/1206 balanced line recievers which emulate transformer balanced inputs extremely well. Again, other manufacturers have equivalents...

[Jackie The Shepherd]

So, the idea is to have a chart in front of me that can help me quickly patch between points depending upon the required connection being Open, Normal, Half-normal.

The labels above (or in some designs, below) each socket should be all the guidance you need.
You just need to know what source you want to feed to which destination, identify those sockets from their labels, and plug a chord in.

Some patchbays have front panel switches which indicate which pairs are normalled (etc) but traditional patchbays are configured semi-permanently internally. In those cases it can be useful to mark normalled or half-normalled sockets in red — either with coloured gel or highlighter, or by drawing red stripes or borders (different organisation have different policies).

It's also helpful in navigating a big patchbay to colour-code useful groups of sockets — like all direct outs, or mixer line ins, or whatever.

I have read something called Patchbay Worksheets, so I assume they would mean that all my studio connections are on a piece of paper for ready reference.

Fair enough... I've never come across one or needed one. I just look at the patchbay and plug as needed.

Thanks Hugh.

I have a set of balanced patchbays so am wondering if these could be used with unbalanced mixer inserts or do I need to look for unbalanced patchbays like the Behringer PX2000. And if I were to stick to the balanced patchbay that I have, and configured as half normal, how would I handle the wiring, i.e.

from Insert Send: TS plug -> TRS plug of balanced patchbay Rear top row,

and

from rear bottom row of balanced patchbay -> Insert return TS plug.

Likewise for the compressor / limiter.

Any wiring diagram that I can refer to?

[Sam Spoons]

Assuming ¼" jack plugs on the back, as soon as you plug an unbalanced cable into it the R & S are shorted so if you use a TRS to 2 x TS cable it won't matter if your patchbay has TS or TRS jacks.

[Jackie The Shepherd]

Assuming ¼" jack plugs on the back, as soon as you plug an unbalanced cable into it the R & S are shorted so if you use a TRS to 2 x TS cable it won't matter if your patchbay has TS or TRS jacks.

Many thanks.

[Hugh Robjohns]

^ Wot he said. TRS to dual TS from the unbalanced console inserts, with the send on the top hole at the back and the return on the bottom hole.

It's obviously a cludge, but it usually works acceptably in most situations.

A balanced TRS patchbay can essentially be downgraded to work with unbalanced TS connections, but you can't go the other way with a TS unbalanced patchbay for balanced connections... so better to buy TRS in case you upgrade the console with balanced inserts in the future.