My '84 (LH2.0), B23F/automatic, flunked emissions for the first time in its life. Two years ago (about 8,000 miles ago), it passed easily on the dynamometer road-speed tests: NOx 248 (req. 1227), HC 36 (req. 162), CO 0.15 (req. 0.91), and CO2 14.4 (no req.).

But New Jersey has a new test protocol: no more dynamometer rollers, and they don't even look at NOx. Instead, they have what they call an "idle" and "high idle" mode.

After I first failed ("1st" column in the table below) last week, I changed the air filter, plugs (they were gapped very wide, over time), checked the AMM (adjusted the pot to exactly 330 ohms). After all this, today I still failed ("2nd" in the table below). Here are the results:

gas ---- req. --- 1st --- 2nd

IDLE:
HC ppm. 220 ---- 648 --- 693
CO % -- 1.20 --- 0.03 --- 0.04
CO2 % --- na --- 6.40 --- 6.98
O2 % ---- na -- 11.56 -- 10.86

HIGH IDLE:
gas ---- req. --- 1st --- 2nd
HC ppm. 220 ---- 666 --- 454
CO % -- 1.20 --- 0.04 --- 0.03
CO2 % --- na -- 11.42 --- 12.26
O2 % ---- na --- 5.00 --- 4.55

You'll notice that my cursory R&R increased my idle HC and lowered my high idle HC (albeit only slightly); and increased my %CO2 a lot!

I think that my AMM is OK -- I've blocked the heated air many years ago, so it always gets only cool air, and its pot easily adjusted to what I think it should be (330 ohms); likewise, the heating circuit checks out to specs, too.

I also found that I couldn't get a dwell reading (hooking up the dwell meter to the pink wire (on the blue plug near the coil, shared with the blue-white wire) -- my dwell meter read full before attaching the wire, and 0 after attaching the wire -- no reading in between.

My O2 sensor reading from a parallel hookup to its connector gave fluctuating voltages under 1.0 volt, but not steady swings. Do I need to change the O2 sensor? It's a bear to do (really tight).

So that's the story -- any suggestions? Anything will be appreciated.
In advance, thanks to all.