Rather than wasting your time to create a variable pad, Shure already makes an XLR barrel Pad. In fact they make numerous ones that have switchable characteristics. These are balanced pads. You don't need continuously variable, trust me. 10, 20, 30, DB switching should be more than adequate. That's why you have gain trim. Hey, this is the way all professional consoles used to be built. So what's the point for this? What is the real application intended for?

That guitarfreak I know has something up his... something?
Mx. Remy Ann David

I didn't quite follow the pot and shunt resistor arrangement you were suggesting. It's necessary to be a bit careful when constructing a balanced pad for a mic input that you use a configuration that works with phantom power. You have to pot across the balanced inputs and not have any reference to ground.

I've never had the need to pad an FF800 microphone input, even with lively mics on loud sources. A pad switch on a microphone body itself is a different matter, as this can adjust the dynamic range of the mic for the sound source it's exposed to. If you really need a pre-amp input pad, I would go for a fixed attenuation one, as Remy suggested.

I don't actually need the pad myself, I am trying to help out a friend who has need for one before his FF800. I saw those in line pads, and I think that they are cool and all, but I think it would be great to have some sort of control on the level of pad in an instantaneous manner. Not to mention designing one which does not alter the sound of whatever passes through it for brighter or darker by using the correctly calculated shunt impedance for the given application. You are familiar with the quest I am sure :D Here is a picture which displays the idea from the original post. I know these sort of things are tough to explain verbally. Words are shallow and audio is a deep science.

(Expired Link Removed)

GF:
What sort of source input are you using?
The 2k input impedance you listed indicates the mic preamp inputs (7-10)?
If this is for a guitar or synth, then I would think using an outboard DI and then running that balanced lowZ into the FF preamps would be all that's needed. The input trim controls should act as your pad unless your trying to drive a really hot line level into the preamps...but then that's why the external balanced DI would be useful.
Also the built in instrument DI (input 1) on the FF has those switchable instrument options (limiter, speaker cab sim etc.) if using a guitar that can be selected.
Then there are the line level inputs either from the front (7-10) or the ones on the rear that could also be used.
Any of those inputs can be trimmed with TotalMix set to adjust your levels going in.
If this is all experimental (i.e. building breadboard pad configurations), then I guess you just need to do your calculations and build what you are thinking and then try it out...without some idea of what your trying to accomplish it's difficult for anyone to give you advice.
Those inline pad connectors are all fixed value pads and good for balancing varying multiple input sources in order to match levels to the input of a mixer, but that's about all they do.
Resistor based pads don't impart any sort of "tone" on anything either if that's what you were thinking...
Hope that helps

[quote=Guitarfreak, post: 377255]I don't actually need the pad myself, I am trying to help out a friend who has need for one before his FF800. I saw those in line pads, and I think that they are cool and all, but I think it would be great to have some sort of control on the level of pad in an instantaneous manner. Not to mention designing one which does not alter the sound of whatever passes through it for brighter or darker by using the correctly calculated shunt impedance for the given application. You are familiar with the quest I am sure :D Here is a picture which displays the idea from the original post. I know these sort of things are tough to explain verbally. Words are shallow and audio is a deep science.

(Expired Link Removed)
Ah, the picture explains everything! The "everything" includes how this is not the best way of achieving the required result.

The working rule in connecting up unterminated electronic gear is that the output impedance of one section should be much less than the input impedance of the next. There are exceptions, of course, but this case is an example of where the rule should apply.

You are driving an FF800 XLR input, which has a nominal input impedance of 2K Ohm, so the output impedance of the attenuator should be of the order of 200 Ohm. As an aside, the noise specifications for the FF800 input will assume a source impedance of this sort of value, and will worsen if a relatively high-impedance source is used.

If you want variable attenuation, there are three ways of connecting the variable element: series, parallel and series/parallel. Your scheme is for series attenuation, which would be fine with the shunt resistor reduced to (say) 220 Ohms, but would need an additional pair of fixed resistors to set the minimum attenuation. If you added 330 Ohm resistors in series with the variable resistor elements, you would get a minimum attenuation of around 12dB while avoiding drive problems that would otherwise happen when the variable resistors were set to zero resistance. A pot value of 10K (with a log law) would give you an attenuation range of 12 - 27.5 dB. With these resistance values, you would not be able to hear any tone change due to the attenuator over its operational range.

It's great that you are tackling this sort of often-overlooked problem and that you are sharing your design thoughts for others to find through searching for something similar. I still raise an eyebrow at the need for an attenuator in front of a FF800 input, though.

Boswell, post: 377308 wrote: Ah, the picture explains everything! The "everything" includes how this is not the best way of achieving the required result.

The working rule in connecting up unterminated electronic gear is that the output impedance of one section should be much less than the input impedance of the next. There are exceptions, of course, but this case is an example of where the rule should apply.

You are driving an FF800 XLR input, which has a nominal input impedance of 2K Ohm, so the output impedance of the attenuator should be of the order of 200 Ohm. As an aside, the noise specifications for the FF800 input will assume a source impedance of this sort of value, and will worsen if a relatively high-impedance source is used.

If you want variable attenuation, there are three ways of connecting the variable element: series, parallel and series/parallel. Your scheme is for series attenuation, which would be fine with the shunt resistor reduced to (say) 220 Ohms, but would need an additional pair of fixed resistors to set the minimum attenuation. If you added 330 Ohm resistors in series with the variable resistor elements, you would get a minimum attenuation of around 12dB while avoiding drive problems that would otherwise happen when the variable resistors were set to zero resistance. A pot value of 10K (with a log law) would give you an attenuation range of 12 - 27.5 dB. With these resistance values, you would not be able to hear any tone change due to the attenuator over its operational range.

I did neglect the idea of the pad impedance summing with the output impedance of the mic itself. It really seems that without creating a balanced amplifier stage in and of itself, the entire concept of an inline mic pad is flawed in and of itself... Seems sort of like a pick your poison situation. You either do like I did and put too much series resistance between the mic and pre, while keeping the load network as close to unloaded as possible. Or on the other hand, you use a more appropriate series resistance and in order to get a good enough reduction ratio, your shunt resistor is going to end up pulling the impedance of the mic pre low. This solution seems almost like we are treating the mic pre as a current operated device... Hmm.

Boswell, post: 377308 wrote: I still raise an eyebrow at the need for an attenuator in front of a FF800 input, though.

A valid point. Just trying to help out a friend. He's not using an external preamp, just the Fireface, and he says that when recording drum set, in order to get the close mics to a reasonable level he has to have the mic channel gains very low, almost off. I told him it might be better to pad the mics a bit and then run the channel gains a bit higher. He asked how much pad he needed and here we are.

Guitarfreak, post: 377326 wrote: ...You either do like I did and put too much series resistance between the mic and pre, while keeping the load network as close to unloaded as possible. Or on the other hand, you use a more appropriate series resistance and in order to get a good enough reduction ratio, your shunt resistor is going to end up pulling the impedance of the mic pre low. This solution seems almost like we are treating the mic pre as a current operated device... Hmm.

No, not really. A simplified way of looking at connecting up this type of equipment (including in-line attenuators) is that the input impedance is relatively high and the output impedance is realtively low of each piece in the chain. The relative defininition of "low" and "high" here could be a factor of 10 difference between them, which would be enough not to make much of a dB change when loading an output with the next input.

So a mic has an output impedance of a couple of hundred Ohms, and would normally plug into the FF800 2KOhm input. This fits with what I said in the previous paragraph. If you make the output impedance of your attenuator also around 200 Ohms by using a 220 Ohm shunt resistor, the FF800 input thinks it's seeing the mic directly and is happy in noise terms. It's then a matter of increasing the series resistors in the attenuator to give you the required attenuation value, given that anything above about 1K Ohm is not going to trouble the mic that is driving it. I'm assuming that the microphones in use are not the type that change their tonal qualities with load resistance.

By the way, the max input level of the FF800 mic inputs is +11dBu. I somehow don't think that even close-miking a drum kit with high-output mics will cause pre-amp overload, although it is exactly the situation where I would be looking for pad switches on the mics themselves to give the mic's internal circuitry a bit more headroom to work with.