Attenuating Unbalanced to Balanced Signal. Instrument level to Mic Level

[abbey road d enfer]

Why not simply buffer the guitar signal (Op amp stage / 1M input Z) then a resistive divider to give you the chosen attenuation, or use a pot to make it variable if desired, - then another unity gain buffer op amp stage that is the active element in an impedance balanced output. Sorted !
Or you could use that buffer output feed a THAT Outsmarts balanced driver device.

In both cases the noise level may not be acceptable. Typical noise level in either case is about -105dBu, which is about 20dB higher than any half-decent mic preamp. The nice thing about a passive attenuator is that it doesn't add significant noise.

 

[ccaudle]

The circuit I have been using is actually this one. where the "-" block is driven by the output of U1A. This one from the musique.com site which is similar topology to the correct one I meant to reference on the sound-au site.

You say you want to attenuate from guitar to mic level, but you are starting out with almost 6dB of gain, and adding another 6dB with the second op-amp stage. Starting out with nearly 12dB of gain is a weird way to build an attenuator. There could be good reasons to do that (e.g. optimizing noise contribution from each stage), but I don't think anyone has articulated any yet.

What gain range does your microphone amp have? A guitar may only need about 10dB of gain, but if your microphone amp can go to that gain range, you may only need a buffer, not an attenuator.
This article has 'scope shots of two different guitars and a bass, which gives some idea of the range of input amplitudes you might have to deal with:
Guitar voltage article at Elliot Sound


If you do need an attenuator to get into the right range for your preamp, you may only need around 6dB -12dB at most if you didn't add gain in the first stages.

You have also started with a circuit which has symmetrical outputs, which is related but not exactly the same as balanced output. I mentioned using impedance balanced output in post #5, if you are no familiar with the term I would recommend searching for info, because it simplifies your circuit design task considerably, and also starts out with 6dB less gain than the circuits you have drawn. Also makes it very easy to have an unbalanced output, just use a TRS jack for the output and you can have balanced or unbalanced without having to worry about what that does to the cold side output driver.

So you might be able to get away with just a unity-gain follower and a handful of resistors. Some capacitors if you have a single supply instead of bipolar.
I would recommend leaving in a series input resistor to protect the op-amp input. At some point you are going to put a big ESD pulse into the input from picking up a cable, and with no current limiting you will blow out the input stage of your op-amp.

Let me know if you aren't following my suggestion, I can try to whip up something quickly in Kicad tomorrow.

 

[Colortone]

You say you want to attenuate from guitar to mic level, but you are starting out with almost 6dB of gain, and adding another 6dB with the second op-amp stage. Starting out with nearly 12dB of gain is a weird way to build an attenuator. There could be good reasons to do that (e.g. optimizing noise contribution from each stage), but I don't think anyone has articulated any yet.

Thanks ccaudle,
Some more info is required for this conversation to go further.
The circuit is part stage box that takes a dynamic mic input ('58) and has a preamp to send recieve at guitar level /impedances for running through guitar FX units. It outputs a balanced signal at the levels/impedance of the mic source. It contains a bypass / kill switch so the balanced output level needs to be 'transparent'.
With a U-Pad on the output of the circuit shown in #9 I am achieving this well however I went down the rabbit hole of fine tuning and looking for problems that perhaps dont exist. However I feel that it seems very counterintuitive to use a circuit that has gain for me to then have to attenuate it further.
I thought perhaps I could swap U1A and U1B around so we start with an inverting input (Gain<1) and then invert back to the correct phase with another inverting stage, however im a little-lot technically challenged when it comes to Balanced signals.
Im assuming an 'averaged' guitar signal level amplitude of up to around 70-150mV. The output wants to assume that the resultant mic balanced output needs to come down to around the levels of a Dynamic Stage mic like the '58 @ 10mV. Unbalanced In High Impedance @1M Source -> Balanced out Low Impedance @150 Ohm Output.

 

[Newmarket]

In both cases the noise level may not be acceptable. Typical noise level in either case is about -105dBu, which is about 20dB higher than any half-decent mic preamp. The nice thing about a passive attenuator is that it doesn't add significant noise.

Yeah - I see that point. Trying to keep it simple. Avoiding "Serious Pads" - that tbh have never been my thing in design terms. And if for standard electric guitar then I wouldn't think it s problem. But yes "Better is Better".
Overall I veer away from attenuation to match the input of a mic pre. Whilst acknowledging that it is the preferred route for some. On a technical design level the engineer of the pre on the Sony Oxford console comes to mind on that front. Last Century though so don't quote me.

 

[ccaudle]

has a preamp to send recieve at guitar level /impedances for running through guitar FX units. It outputs a balanced signal at the levels/impedance of the mic source. It contains a bypass / kill switch so the balanced output level needs to be 'transparent'.

That is a lot of important information which was left out of the original post, which also simplifies things considerably. What is the gain of the first stage? That tells you how much attenuation you need for the final stage. I assume by "transparent" you mean that the levels in bypass and the levels with effects active should be as close to the same as possible. Since effects pedals can change the level, that implies you will probably need some level adjustment on the return input.

This is a case where input and output transformers would possibly (probably?) simplify things, but you can certainly do it all with op-amps.
Either way your original circuit post left out all the input stages, which gave a misleading impression of what you were trying to do, and is overly complicated for what you need.

Guitar pedals (typically) don't have high inductance outputs, and usually don't have output impedance quite as high as a guitar pickup, so the input impedance of the return stage isn't quite as important as it would be if you were really plugging in a guitar. Something like a typical line input stage for consumer unbalanced equipment would probably work, then an attenuator pad on the output that gets the level and impedance you want, and a matching pad for the cold output to create an impedance balanced output.

 

[Colortone]

What is the gain of the first stage?

The Input gain is variable so the user can can create a 'unity' output with their chosen flavour of dynamic mic and bypassed chained FX.
My initial design was with a transformer which solved a lot of simple problems but it was prohibitive in terms of costs :/

 

[Newmarket]

Guitar pedals (typically) don't have high inductance outputs, and usually don't have output impedance quite as high as a guitar pickup, so the input impedance of the return stage isn't quite as important as it would be if you were really plugging in a guitar. Something like a typical line input stage for consumer unbalanced equipment would probably work, then an attenuator pad on the output that gets the level and impedance you want, and a matching pad for the cold output to create an impedance balanced output.

Yes - lots of info' added that alters the basic requirements. No actual guitar pickup so the previously stated "1M0 Input Z" requirement doesn't apply.

 

[Colortone]

Yes - lots of info' added that alters the basic requirements. No actual guitar pickup so the previously stated "1M0 Input Z" requirement doesn't apply.

I would think the 1M input Z does apply, as a lot of new designed guitar fx units abide to that very rough standard of +1M input impedance to Sub 1-10k output impedance. Its entering a varying world of unknowns but im sure thats the same when ever a professional starts patching/chaining hardware audio devices in the mix world, guitar world or synth world no? Happy accidents....

I apologise for being loose (and incorrect) on the initial info, im a newcomer to audio hardware design and I am in that newbie place of feeling like you have to hide some of your IP ideas/information when in actual fact you are talking to people here that have heard it all before Its all been done before but how to do it well and efficiently is the right goal.
I completely appreciate the responses on this forum and give a million thanks to people like you who offer guidance and their time.
Im learning a lot here.

 

[Newmarket]

I would think the 1M input Z does apply, as a lot of new designed guitar fx units abide to that very rough standard of +1M input impedance to Sub 1-10k output impedance. Its entering a varying world of unknowns but im sure thats the same when ever a professional starts patching/chaining hardware audio devices in the mix world, guitar world or synth world no? Happy accidents....

I apologise for being loose (and incorrect) on the initial info, im a newcomer to audio hardware design and I am in that newbie place of feeling like you have to hide some of your IP ideas/information when in actual fact you are talking to people here that have heard it all before Its all been done before but how to do it well and efficiently is the right goal.
I completely appreciate the responses on this forum and give a million thanks to people like you who offer guidance and their time.
Im learning a lot here.

Nice post.
On the 1M Z. It (or some other 'High' value) is chosen as the input Z for fx units because it may have a guitar plugged into it.
In your case it won't have a guitar plugged in with a high Z pickup as the source so it's not a requirement.
There's no actual harm in having 1M0 Zin but the typical combination of this with a FET opamp (non-inverting configuration) for electric guitar may not be optimum for use with a much lower impedance source.
As you indicate guitar fx pedals have various output circuits - sometimes direct from a pot wiper and build out resistor such that the output impedance varies with level - and some via a transistor or opamp buffer with a far more tightly defined value.

 

[RuudNL]

Wouldn't it be easier to use a standard DI box, instead of reinventing the wheel?
(After all, that is what DI boxes are made for...)

 

[Newmarket]

Wouldn't it be easier to use a standard DI box, instead of reinventing the wheel?
(After all, that is what DI boxes are made for...)

Ahh...cue discussion of what constitutes a "standard DI box"
Back in the thread the OP does state that he had a transformer based solution but that cost was an issue.
I don't know if this is intended to be a one off or built in quantity.

 

[Colortone]

Haha Yes, Ive found a "standard DI box" mostly assumes a "line level" out and I want a Mic Level out hence the padding/attenuation required. Ive seen and tried a transformer based Mic Level solution as Newmarket suggested but it was cost prohibitive for this first incarnation of a small run of products. It was good though

 

[JohnRoberts]

DI stands for direct injection... historically they involved a step down transformer to better match mic preamp input levels.

JR

 

[abbey road d enfer]

Haha Yes, Ive found a "standard DI box" mostly assumes a "line level" out

Certainly not. Actually, the subject of level was subsidiary to taht of balancing, because wiring in studios was always balanced.
When the concept of bypassing the amp, speaker and air came about, three main concepts emerged:

• Direct connection of the instrument output to a mic input, which resulted in very poor results, because of the large impedance mismatch, and the alteration of the potentiometer taper. There are some examples of that, particularly in the 70's country music. Because of the mismatch, the guitar's volume control did not work correctly so it had to be cranked up, and the signal was so squelched the gain on the mixer had to be cranked up too, which resulted in the inadvertent invention/discovery of fuzz.
• Modification of an existing mic preamp, bypassing the input transformer, which resulted in a line-level output. The paternity can probably be attributed to Les Paul, with his seminal 1950's recording, where he probably used the mic inputs of his Ampex recorder with the input xfmr bypassed.
  A nice example is the Guitars Unlimited, on Barclay Records, using various modified tube preamps
• Use of a step-down transformer, that provides better (but not very good) impedance matching and reduced level, suitable for a mic input. Actually Les Paul also used that a lot, with the transformer in his guitar. He later had his guitar fitted with low-impedance pickups and wiring, with a direct balanced low-Z output. It did not really catch on.

Actually DI'ed guitar has seldom been a favourite for flamboyant solos or gritty rhythm. Now it has been given a new lease of life with the advent of pedal/amp/cab simulators, which any half-decent DAW includes today.

and I want a Mic Level out hence the padding/attenuation required.

That's what most current DI's provide. It's mainly due to the creation of phantom-powered solid-state DI boxes in the early 70's.
It's undoubtedly the most economical solution, and the one that can potentially deliver the best performance.

 

[Whoops]

convert a instrument level signal to a balanced mic level

That's a DI box.
That's what you need.
There's a great DI box project here in the forum it's called Bo Hansen DI

Haha Yes, Ive found a "standard DI box" mostly assumes a "line level" out and I want a Mic Level

You are making some confusion, a DI box is a device that converts unbalanced Hi-Z "Instrument Level" into Balanced Low-Z "Microphone Level.

Like JR said, DI means Direct Injection and the goal was and is to direct inject an Instrument signal into a Microphone Preamp (in the old days a Mic input in the console)

Just search here or google "DIY DI box" and you will find many DI projects to achieve what you want, but I trully advise the Bo Hansen DI project

 

[Colortone]

** Edited to add my own take on the circuit.

Ive made the BoHansen DI Box and I loved it. However as previously mentioned the cost of good transformers was prohibitive to this small run product.
Interestingly yesterday I came across this which appears to do in theory what Im looking for, without the need for pre or post padding. (image from site attached below for reference and my take on it below that)

The idea would be to use an inverting opamp first to achieve a "less than 1" attenuation which would create the (-) signal with its output through another inverting opamp to invert and achieve the (+) signal.

Is this worth me following up on or am I missing a crucial gotcha?





My take on it :
* 1M input impedance with 220nf Cap to reject DC (forming a high pass Filter @ 1Hz)
* First inverting section has a -26db attenuation and forms the - Signal.
* Second Inverting section maintains that gain and forms the + Signal.
* Output DC blocking (another highpass filter @ 1Hz) and ZOut set to @ 47 Ohms. Expecting to feed a 1.5k to 2k Pre on a mix desk.

 

[abbey road d enfer]

The idea would be to use an inverting opamp first to achieve a "less than 1" attenuation which would create the (-) signal with its output through another inverting opamp to invert and achieve the (+) signal.

Is this worth me following up on or am I missing a crucial gotcha?

Noise.

View attachment 100345


My take on it :
* 1M input impedance with 220nf Cap to reject DC

The actual input impedance is 9 900 ohms.

(forming a high pass Filter @ 1Hz)

72 Hz

* First inverting section has a -26db attenuation and forms the - Signal.

Noise generated by the opamp is about 0.7uV (-121dBu). Assuming an input level of -20dBu, hence an operating level of -46dBu, the best ideal S/N ratio is limited to 75dB, much less in practice.
As much as someone suggested it was a bit silly to amplify signal for attenuating it later, I see no reason against operating at unity gain at the input and inserting a potentiometer before the balancing stage, with a passive attenuator right at the end.

 

[Whoops]

Ive made the BoHansen DI Box and I loved it. However as previously mentioned the cost of good transformers was prohibitive to this small run product.

I can share with you 2 good options:

1) There's an active DI Box project, with no transformer, in ESP projects, all the details and schematic are there, it's exactly what you describe you are looking for,
check it out in the link below

https://sound-au.com/project35.htm




2) the second option is that you can buy for very cheap the Orchid Micro DI box, it's very affordable, small, transformerless and also does exactly what you're after. It will be cheaper for sure than to build a DIY unit

http://orchid-electronics.co.uk/micro.htm




Hope this helps

 

[Bo Deadly]

Here's a circuit that I just whipped up that I think would satisfy most environments and be a lot more effective over a passive transformer circuit:


Bo Deadly's Simple DI JFET / Impedance Balanced / Dual 9V

 

[abbey road d enfer]

Here's a circuit that I just whipped up that I think would satisfy most environments and be a lot more effective over a passive transformer circuit:

View attachment 100368

It doesn't answer the desire of the OP for attenuation, whether it's still necessary or not? The design brief seems to be a moving target.