Keeping an amine unit from foaming without a smart skid

I'm the mech guy babysitting an MDEA sweetening unit that foams at the worst times. When contactor DP climbs and the regenerator level chatters, our compressors start singing. Vendor wants to sell a cloud-tuned antifoam skid. I’d rather keep control in the rack room.

What’s worked: clamp hydrocarbon slip, swap carbon when lean color darkens, drain coalescers daily, and a very low continuous antifoam trickle through a rotameter. I use hot-lean shaker-bottle tests and tie dose to contactor DP and gas rate, not a calendar. If DP ramps, we hand-boost then taper. Silicone helped, but I’ve seen it foul downstream and murder carbon life.

For MDEA with 10 - 15% H2S and some condensate carryover, what antifoam chemistry is the least ugly long-term (silicone vs polyether vs amine-friendly blends)? Any field tests you trust beyond the bottle shake to set dose and catch hydrocarbon early? Rules of thumb for carbon changeout and coalescer spec so I’m not masking problems with antifoam?