Vacuum fluctuates when brakes applied

Hey guys, got a weird one here, just installed new dual power booster with master cylinder. Brake only travels 1/2 down and stops rock solid with vacuum connected and when I disconnect the vacuum, brake pedal travels all the way down like it should. Also, my vacuum dramatically fluctuates when brakes are applied. I lose 3-5” of vacuum when I apply brakes and 8-10” of vacuum when brakes released. Any ideas? I drained out the air in the lines.
Hang on
by this description, the booster is working correctly. The booster always pushes the pedal up after first start. That is normal.
And only traveling half-way down is also normal.
The rock-solid indicates no air in the lines.
If the car stops nice with modest pressure, then everything is normal.
But if the degree of slowing is somewhat uncontrollable then the atmospheric valve is probably damaged.

Some vacuum fluctuation is to be expected. On a stock-type engine you just never notice it. But on a performance engine, which already has a low vacuum idle, then yes you can notice it. But I haven't seen as big a drop as you have nor trouble bringing it back up. But here's the thing; at a stop, when applying the brakes, atmosphere rushes in past the open atmospheric valve,and because the backside of the diaphragm has been evacuated, the leg-power is multiplied by some variable amount depending on how hard you are pushing and how much resistance is in the hydraulic system. But this whole system depends on the pushrod length being the right length.
If the pushrod is too short, the boost will be small and the pedal-pressure required will be great. Perhaps not a great as without a booster. The M/C size also comes into play, with large ones requiring a lot of leg-power and a small amount of pedal stroke; while the smaller ones require more stoke but less effort.
If the pushrod is too long, boost will start early, and it will not require much effort to over-brake. And it is possible to shatter the atmospheric valve if you stomp on the pedal. After that, depending on how it broke, as soon as you ask for brakes the dang thing will put your nose into the windshield.
The 8-10drop is hard to understand, on an engine that pulls 20; I don't understand that unless the idlespeed is very low or the ring-seal is poor. And the following statement is kindof pointing in that direction.
Yeah, idle goes down as with vacuum (2-5”) when I apply brakes, when I release brakes idle drops 8-10” then slowly comes back to 10-20” of vacuum and idle gets restored when vacuum goes back to normal.
But, I suppose idle-timing could have a large part in that; more specifically a too-small transfer slot exposure, due to too much idle-timing. But I'm just guessing.
So where are you picking up the vacuum signal? off a runner? off the valve? or off the plenum?