Once per second ?! Aren't they both low and high sampling rate channels (for the acceleration increments) within the FDR ?
Disclaimer, my expertise is engines, not FDR. But I've looked at lots and lots of FDR data over the years and I've never seen a parameter that was recorded at more than 'once per frame' within a one second frame (usually this has been for anomalies such as an engine shutdown or similar a anomaly, but also for a handful of crashes).
Heck, I consider myself lucky if most of the parameters I'm interested in are actually recorded, and at once/second (one incident, engine parameters were recorded every 64 seconds - for an event that lasted less than a minute
). Digital Flight Data Recorders have improved hugely over the last 30 years - way more parameters recorded at a higher update rate (during the investigation of the BA38 777 that landed short at Heathrow, I was pleasantly surprised to see the FDR recorded the fuel metering valve position - so we could definitely determine that the fuel metering valve opened but fuel flow failed to increase). Leaps and bounds better than the original FDRs that recorded a couple dozen parameters on foil
All that being said, a 200+ ton aircraft doesn't change direction that fast. Looking at 1/second data normally produces very good indications of what the aircraft was experiencing. I've seen indications of jet wake and wing vortex encounters that were pretty obvious from 1/second data.
When it's all said on done - while FDRs are hugely more capable than they were 20 years ago, there is still a limit to how much data and at what update rate. 99.9% of the time, 1/second is adequate provided the necessary parameters are recorded - so the emphasis has been on more parameters rather then fewer parameters at a high update rate. Further, to get meaningful data for something like a MH17 missile strike, you need 20 or more updates per second - something we only get with dedicated flight test data system.
If it was up to me, I'd be looking at a zero risk way to keep the FDR powered for a meaningful length of time after main bus power was lost - ideally something like capacitors that could power for another 10 minutes or so. Although it probably would make no difference in this case (the FDR on a 777 gets all it's data via AIMS, which was likely knocked out an instant after the missile strike), there have been a number if incidents where the FDR lost power well before the actual crash.