[AllyCat]

Those versions have an special advantage for the WeatherDuino users. Due to the par of FETs included, they can be connected directly to extended I2C bus port on the TX boards, while the more recent versions can't (at least without using another bus extender chip on the sensor end).

Going slightly Off-Topic.  I believe the same applies to the VEML6075 (UV) sensor (i.e. 3.6v max rating).  I'm particularly interested in using (or testing) that device because I believe it offers the best possibility of obtaining a "real" UV Index value (which is heavily weighted towards the UVB end of the spectrum), because it has separate UVA and UVB channels.  Also, it has two more channels weighted towards the Visible and Infra Red parts of the spectrum, so I wonder if it can also provide the "normal" Solar Watts/m2 value.  That gives the possibility of doing all the Solar measurements via a single (extended) I2C bus and a tiny optical "window".  But of course it would require a "special" Arduino Library, which I'm certainly not competent to devise (yet).

So your comment got me thinking (and searching) and I found the following; several Pounds more expensive than the normal 3.6 volt VEML 6075 breakouts (like the one I already have), but perhaps an "easy" solution. 

https://www.ebay.co.uk/itm/VEML6075-UV-A...:rk:9:pf:0

Cheers,  Alan