It's "Big Wi-Fi Myth Number 2" that client devices are always "hunting for the best signal." They do not and some clients need it to get pretty grotty before they initiate a roaming assessment.
With many clients, there's nothing you can do about their roaming behaviour and it's in the gift of the client device designer. (If/when to roam is not prescribed in standards.)
To stand the best chance of roaming, SSID names and passphrases need to be identical, ideally all AP's offering the same protocols (A/B/G/N/AC,) though that is not absolutely essential and I'd ensure the AP's are on differing radio channels to ensure there's less "near-far" masking occurring and/or mitigating the chance of AP's & client transmissions in different cells "interfering" with each other.
I prefer not to use "auto" tuning functions on small (unmanaged) deployments. In the 5GHz waveband, I just ensure the AP's are tuned to different channels. In the 2.4GHz waveband, the channels need to be different and at least "5 apart," for example, choose from the set [1,6,11] If you wanted to go the whole hog, you could assess any neighbouring Wi-Fi using tools like InSSIDer and try to avoid the neighbours if you have any.
"N" Wi-Fi in the 2.4GHz can also be problematic if one uses "fat" (40MHz) channels as there is insufficient frequency spectrum to deploy multiple calls with "fat" channels without them interfering with each other. In big deployments, it's not unknown to restrict N in 2.4GHz to "thin" (20MHz) channels, accepting that so doing effectively halves the maximum link rate (speed) but could yield a more stable infrastructure and more usable radio channels. You could have a play with that if your routers offer such choices.
Cheers mate, but have to say before i done the reset and that it was working ok and not like this