I'd go with a Jecklin Disk. You could cobble up a quick 'n dirty JD by glueing a stack of old LP's - I urge you to use up any Barry Manilow's, for optimum sound :-) - and covering that stack with some rubber carpet underlay or maybe a few old pc mouse mats.
With performers evenly around a circle, the Blumlein array does have two disadvantages:
1. Sounds from the two extreme L and R sectors fall in the "ambiophonic" region. That is, they deliver opposite polarity voltages to the two coincident mics (one mic getting a +ve signal while the other mic gets a -ve value). This results in a woozy, vague imaging of those sound sources.
2. Sounds from the rear sector will - besides generating signals of a polarity inverse to those from the front - also give a 'laterally-inverted' imaging, e.g. sources located at, say, "4-o'clock" and "5-o'clock" will image as if they had been located at 8- and 7-o'clock, respectively.
To illustrate this further, a consequence would be that, for a perfectly-centred array and 6 performers A...F stationed at direction angles of A 10-30, B 11-30, C 12-30, D 1-30, E 5- and F 7-o'clock, the resultant imaging from a two-speaker playback would likely be in the following sequence:
- from Jecklin Disk: L -> A - F - B - C - E - D -> R
- but from Blumlein: L -> A - E - B - C - F - D -> R