A bit of background info for anyone who comes across this and doesn't already know it.
TM20 Manual wrote:Make sure the Red Box is set to the right level: Line is always the best choice due to the higher output level. If the mixing console’s XLR input can only handle microphone levels, set the Red Box to mic. If the mixing
console or soundcard/audio interface lacks XLR inputs, you will need an XLR to 6.3mm (1/4") jack adapter, which should be readily available in music stores. Please make sure the Red Box is set to line when it is connected to 6.3mm (1/4") jack inputs.
The TM20D schematic is not in public domain so I can't refer to that, and the TM18 which is available has a different Redbox setup so doesn't apply. The TM20D manual shows the Redbox output as "XLR, balanced, 1360Ω, max +0 dBV". If, as H&K recommend, you use the full XLR Balanced mode with +ve on pin 1 and -ve on pin 3, it is being fed from two 680R resistors, one in each branch. If you use it to an unbalanced input you should be using pins 1 and the ground on pin 2 and you will get half the voltage at 680R impedance. The impedances don't usually matter as long as the input impedance of the following unit is much higher which it generally is.
So the TM20 output is a maximum of 0dB compared to a reference of 1V and that is, errrm, exactly 1V as it happens. dB are very misunderstood by non-techys, and by many techys too. The dB is NOT a measure of anything at all. It has no units like Volts, Amps or Ohms and can be applied to anything you like. Technically depth, age, weight could all be expressed as dB if we knew what they were being compared to. It is a comparison of 'something' to a 'reference' level, or to put it another way 'something' is 'so many times as big (or small) as something else'. For it to have any meaning you need to know what the reference in the case under question is. It makes no sense to just say that a voltage is '20 times as big'. You need to know '20 times as big as what'? As I said, for us the reference was 1V which is what the 'V' after dB tells us. Another reference standard is that dBu, used a lot in studios, which is against a reference of 0.775V, (that's a bit of a magic number in electrical terms). If we simply say 'the voltage is +6dB' it means nothing at all. We need to say something like 'the voltage is +6dB referenced to the input' which shows the first voltage is being compared to a second one to which it will be 2x as large.
The Scarlett 2i2 manual says it has inputs at Mic - +4dBu or 1.228V, Line - +22dBu or 9.752V and Instrument - +13dBu or 3.46V. You should be able to match to those easily. Remember those are the maximum signals the unit can cope with without distortion and you have a volume control to be able to set it lower as you wish. I would try using the Line input first and see if the levels pan out to sensible settings at the controls of each unit.
Some useful hints when dealing with dB are that, for voltages, 6dB = 2x, 12dB=4x and 20dB = 10x, (approximately but very close). Also bear in mind that one reason to use dB is that they simply add up while 'so many times' must be multiplied to apply more than one. For example if we amplify a voltage by 2x then by 4x then by 10x we would get a voltage of 80x the initial level, that's 2 x 4 x 10 x Vref. Using dB we would be increasing the first voltage by 38dB, or 6dB + 12dB + 20dB. That's a simple calculation to demonstrate the point but it gets much easier when you are dealing with less helpful numbers including fractions.
Just as Pat says, it is better to use as high a signal level as your devices can sensibly work with in order to reduce the effect of noise. It is also better to use balanced mode connections where you can for exactly the same reason, they are much better at rejecting noise signals than single sided unbalanced ones.