I did this ungrounded pot mod on my P/J bass using a 250K M/N. For my needs, it's great, as I tend to not do delicate blends, but run the bass really in two settings: full on P or center detent blend with the J pickup. I could have used a switch and been just as happy, but I just had to try this blend.
I have read that a 100K M/N allows finer blending. I'm sure it does, but I don't have first hand knowledge of it. However, I already had a 250K M/N and wanted to keep the loading impedance as high as possible, to retain clarity with the Wilde pickups on the bass. I use a 500K linear volume and 500K audio tone pot on the bass.
~
This leads me to an observation which I have wanted to make about how this blend/balance hookup functions.
The signal flow chart is like this :
Pickups > Blend Circuit > Tone Circuit > Volume Circuit > Output Jack
I wish to initially address the relationship between the blend pot and the volume pot, in the ungrounded version, and without cutting any traces of the blend pot.
What we essentially have is each pickup feeding into the volume pot through a rheostat [a variable series resistor], which is one or the other gang of the blend pot.
So, if the blend pot is 100K [per gang] and the volume pot is 250K, then it means that for each pickup we have a voltage divider between two resistances, with each upper leg of the divider representing anywhere between 0 K and 100K, depending upon the blend setting, and the lower leg is a constant 250K.
With those ratios, at first glance one would expect most of the signal of the unfavoured pickup to still be getting through when the other one is favoured, yet people have reported that hardly any of the signal of the unfavoured pickup gets through.
"Why is this ?", some may ask.
It is because the favoured pickup is connected directly across the volume pot when it is favoured, for HALF of the rotation of the MN tapered blend pot, and has a much lower resistance than the volume pot itself does, and consequently contributes in a large way to the attenuation/division of the signal of the unfavoured pickup.
If we speculated that each pickup has a 10K internal resistance [and some do], then instead of the unfavoured pickup's signal seeing simply a 250K resistance to ground when it reaches the volume pot, allowing most of the signal to go into the volume pot, it also sees the 10K resistance of the favoured pickup in parallel with the volume pot, which loads down the signal of the unfavoured pickup much much more than the volume pot alone would, sufficiently so that we hardly notice the unfavoured pickup's signal when the favoured pickup is set to maximum.
So when we are running tests of MN tapered blend pots of various resistances [100K, 250K, etc] effect upon the unfavoured pickup, we must consider the ratio between the resistance of each gang of the blend pot and the resistance of the favoured pickup, moreso than the ratio of resistance between the blend and volume pots.
~