Hi Alan,
Well those are some interesting results, and I wouldn't say my mind can't be changed I think we're seeing that the Kemper performs a lot of the functions of an active DI, but not all.
1 - Impedance - ...both are 220ohm on my multi meter across two pins.... Am I measuring wrong?
I doubt it - it sounds like it is me that is plain wrong here. I'll take that one on the chin; anything under 600 ohms would be fine. I'll get the multimeter out for my Stage and Radial DI boxes later as I'm curious how they compare. I've got an Art dpdB lurking somewhere as well.
2 - Balanced signal - the Kemper XLR Out is a balanced XLR signal which is suitable for long cable runs.
Sure, the XLR out is balanced, and this is one of the primary functions a DI performs, but balanced line outputs on professional gear is par for course.
3 - Level - while the KPA is capable of sending up to +16db (Head/Toaster or +15db Stage) this can be reduced to - infinity by using the Output Volume. A DI Box pads the output with a fixed attenuation (usually 20 - 40db). Why is simply reducing the volume being sent from the KPA different to engaging a pad switch on the DI Box?
It's this last bit that, for me at least, that means the Kemper output does not replace a DI box. The output of the Toaster/Rack is only +16dBu on the unbalanced TS outs. The balanced XLR outs max out at +22dBu. The additional 6dBu equates to twice the signal intensity (peak to trough being measured by a differential amp rather than just peak to 0V reference).
The Stage must have an attenuator for the XLR outs so that there is no signal level change between XLR and TS (I actually rather like this approach, to prevent 'blatting' during setup - I think this was the reason cited by Kemper for doing this).
But mic level is down in the -60dBu to -40dBu range, which equates to a few millivolts, so even with the -12dB attenuation switch on the Kemper output, you've got a lot of gain reduction needed before it becomes suitable for a mic input that could well have a minimum gain value of 20dB to consider (most live engineer workflows I've witnessed don't start with all mic inputs padded or set to Line/unity gain, for example).
I don't think using it in this way would be a sustainable workflow, as the potential and penalties for it going wrong are high. An analogue mic pot will be log, and if you're having to use it on minimum gain it will be difficult to adjust by a small amount if you subsequently need to, plus the ends of the pot travel are where the pot is noisiest, as that's where all the crud from the wipers gets pushed to.
I guess you could argue that any low impedance line level output with a level control could be reduced enough to feed a mic input, but there is an inherent noise floor in any system, and trimming down the output gain this much will surely take you closer to it, with the noise floor being amplified up with the audio signal at the mic input with this reduced ratio. The Stage might be easier to manage, as it has the same -12dB output pad as far , but is already lower in signal level with its max +15dBu output. With the output pad, you'll still be at up to +3dBu, which is over 1V. I'll have to have a play with this to be sure- I'm interested to know what the noise floor performance is like at minimum output. Could do with finding someone with a D-Scope at home, as I'm unable to go to the office for the foreseeable future.
By comparison, engaging a PAD switch on a DI would result in the noise floor also being attenuated, preserving the signal to noise ratio of the Kemper output. It's possible Kemper could be working some voodoo magic that means signal to noise is better than expected even when the gain value is at its lowest, but I will reserve judgement until someone can scope it out. I'm also not sure if the Kemper output PAD is an analogue function or not. If so, it will indeed be switching in/out circuit components so that the gain range remains optimised. It might be that this is just moving the digital trim range before the DAC. If it was a -30dB or even -40dB analogue PAD function, it would serve much better as a DI output.
Still, after all this, the functions of a passive DI that still remain are the transformer isolation that can be reassuring in certain live situations. I mentioned line isolating transformers in my first reply, so perhaps they are worth mentioning again, as one of these, with a PAD function, might be closer to what the OP actually wants than a DI box, given that the Kemper is partway there.
There is a lot of context dependancy here, for sure; playing level, the rig levels, the output trim, the Kemper model, the input stage of the mixer. If someone can bend it to their will and is happy with the audio results and the workflow without additional hardware, then ok