Silver brazing porosity issues.

[rcs9250]

I have attached a sketch of a part I silver braze together.
Both componenents are 303 SS. They simply press fit together so there is no capillary action going on in the area of the press fit.
I heat and add silver braze to the chamfer area. Once the part is cooled and flux is removed I have to file and sand the top surface so it is completely flat.
I must be doing something wrong sporadically because about 25% of the time when I file the silver braze there can be porosity showing which is not permitted. The area being filled is relatively large. I say that with the belief that the intention of silver brazing is usually capillary in nature. I don't think the intentioon of silver brazing is to put a bunch on and then file a bunch of it away.
Being the porosity does not happen consistently it makes me wonder if I'm boiling the silver braze filler ( trapped bubbles). I do my darndest to apply minimal heat and even take a piece of pointed tungsten and stir around the chamfer to try and free any bubbles.
I'm wondering if porosity in a joint that large is typical.

I'm using StaySilv 45% with white paste flux.
I have used oxy-acetylene , acetylene-air and propane-air.
I prefer the propane or air-acetylene because it helps prevent overheating of the part.

Any help would be appreciated. I have 400 of them to do and would like to hit it the first time around.
And no, I am not permitted to use a different method to construct or join the part together.

Randy

[rcs9250]

I tried to open the attachment and was not successful. If you can't either here is a description of the joint.
It is a 3/16 diameter shaft with a 1/32 x 45 degree chamfer that is pressed into a 1/4'' thick tab with a 3/16 diameter hole in it. The hole has a 1/32 chamfer that when the two parts are pressed together form a vee groove that will be filled with silver braze.

Randy