You don't really mention hat you're setting dialnorm to.
Both are ways of measuring something. Dialnorm was king for a long time (LM100), and it's used often for DVD authoring, and yes, it should be read and set to the correct value. Then dynamic range playback settings will alter mix based on this value.
LUFS/LKFS/BS1770.1.2.3 are the current trend for adherence to CALM/R128.
I suspect you know this though.
I'm not aware of a spec that utilizes both. It's hard enough being creative in one of them, much less hitting two targets.
T.C. and also the EBU officially recommend to put the integrated LUFS measurement into the dialnorm field, so that's what we do here. Right now listening to the TV broadcast of a concert mix I did (in 5.1) that was mixed to -23 LUFS and dynamics seem to be right. Although I don't know if that means anything since I have no idea what metadata is set in the AC-3.
Found out that this station doesn't transfer the metadata from the delivered Dolby-E to the EAC-3 that is broadcast.
The ac3 broadcast is probably a single stream using a single setting that's being made realtime on the broadcast side. You don't want your decoder/receiver to 're-lock' to a new ac3 every commercial.
Nitpicking here; dynamic range is actually not altered by the dialnorm setting. The dialnorm is an overall gain. Dynamics can however be adjusted by applying the compression profile, which in turn used the dialnorm value as a reference point. So if the dialnorm is set way off, but no compression profile is applied at the end user (DRC setting in a home receiver), the dynamics are not altered, but the overall volume might be too low/loud compared to others.
In your case I'd stick to R128. If you deliver a R128 compliant 5.1 mix, the mixdown will probably be 1-2 LU louder (in my experience).