Yes i think it could be a red herring. "Erp version confirmed"
If you are using a brand new head assembly then i would say the ball pcb is faulty, it is possible as stu says to have a faulty new head and that is what it could point to.
The non erp version pcb output with no head connected would give you 320+vdc ish and would discharge in a few minutes as there are 3x 68k resistors in series over the dc output (204k) as a discharge load. This non erp version is a very basic full wave bridge rectifier,couple of inductors, reseviour cap and discharge R. nothing complicated with that.
To have 300v is correct i believe from an erp version, and looking at the circuitry on the pcb i would say that if you measure voltage then turn machine off that 300vdc will vanish within a couple of seconds but with a load on it will not produce anything. This is due to control circuitry and possibly current sensing ?
Either you have been very unlucky buying a faulty new head, very unlikely but not unheard of, or the original head motor failed - destroying the pcb. not uncommon.
The seller will probably refuse a return of the head, and will have had this experience many many times with buyers.
I would strip the new head so you can measure the voltage over the motor whilst connected to the machine, and my bet is it will fire up for 0.5s then drop to 0v when switching the machine on.
I assume that the yoke cable is good,? another red herring thrown in,,, i have seen similar problems with one of the wires almost severed, and arcing inside the insulation, the carbon created by the arcing is enough to pass a voltage to measure with a meter but obviously not passive for current. seen this in many industrial applications at vauxhall motors production line robotics looms years ago, we were building and supplying add on monitor units for them.
Ideally you could do with a dummy load instead of the head, (i should make one tbh) a resistor with a value of 600ohms should draw 500ma for testing, but needs to be high wattage (150watts ?) and then measure the voltage.
The erp pcb is VERY different to a non erp pcb, not really done much with these as a repair to an erp one would probably cost more than a new one.
Here are some pics of the pcb's.
Careful if buying a used erp pcb, as some sellers do not know the difference and cannot confirm it to be working, they scrap and strip faulty vacs, assumming the head has failed and sell the pcb not knowlingly faulty !
This subject repeatedly occurs, and covers most models with powered motorized brush heads, people think it's the brush head at fault, order a new one then fall in to the trap. most common is the dc25/dc40 for me, and the yoke cable is more often to blame, it like playing darts in the dark with a blind man trying to do the marking.
(sometimes it is just the brush head)
Many get to this point of diagnosis and give up, and the forum has many related posts.
However i could be wrong.....