Is this match ok--controlled group, two plans

Controlled group.  Two different plans.  Must be tested together due to coverage.

Plan 1:  Discretionary match:  100% up to 4%

Plan 2:  Discretionary match:  0.00%

Do I have to test this for coverage? 

BRF?  will zero discretionary match fail one or both current/effective availability?

Yes and maybe, if you don't have high enough percentage of NHCEs in Plan 1.

New idea:

Plan 1:  100% to 3% of pay

Plan 2:  100% to 0.75% of pay

Coverage passes.  What about BRF?  Can I use Nondiscriminatory classification test?

On the new idea, yes, the aggregated Plan 1+2 has two different match rates.  Only Plan 1's match rate must pass BRF testing (the participants who have a match at least as favorable as 100% to 0.75% of pay available to them is the same as the coverage test). 

I don't know whether you can use the nondiscriminatory classification test because you haven't told us whether the criteria used to determine who gets the better match rate (i.e. who is eligible for Plan 1) are reasonable and based on objective business criteria.  See Treas. Reg. Sections 1.401(a)(4)-4(b)(1) and 1.410(b)-4(b).

Objective business criterion (I'm my opinion):  Company A gets one match.  Company B gets the other.  Can't get more objective than that.

Can someone check my math?

6/9 HCE get higher match (66.667%)

19/77 NHCE get higher match (24.675%)

Coverage percent:  37.0130%. FAIL Ratio Test

Concentration %. 77/86 = 28.25%

PASSES Nondiscriminatory classification test

(I believe it is a valid business reason)

Math looks fine to me.