Sorry mate, but I did warn you about going the small SSD route and recommended the larger ones. I too thought that it would be able to do more than that or I would have put it much more strongly. Damn surprising actually as the PERC was a good raid board back in the day and it should have been able to give more than that!
TBH, your numbers are decent but if you want MORE go with lowfats recommendation. Onboard, hmm havent run 4 on fake raid (never seen the need, let alone want for that level of headache). I know with TWO SF drives you can hit about 500+mb/s so it might be a CHEAP way to go for the short term (once again though the numbers are all over the place with fake raid and SSDs on a SB890. Some tests will show about a 80-85% performance boost by adding a second SSD into an array...others show a mere 20% boost), BUT I really, really, really loathe fake raid. Spend the 3 bills and get a very good card you will be happier in the long run.
Oh and I would have answered earlier but I am on nights for Jan (and probably part of Feb) and this is about as early in the day as I am up. :)~
EDIT:
Actually now that I think about it. Some of your problem might actually be with the test you ran. Like on fake raid you might get wonky numbers all over the place with your PERC. SSDs and RAID do make for some "odd" results sometimes. Give AS-SSD and CDM a whirl. IF you are interested in seq speed, try HDtach and take a look at the graph it makes...maybe it can shed some light on to what is going on.