MOSFET datasheet confusion - negative values and threshold questions

[luxfx]

I bought a handful of IRF5305PBF p-channel FETs to try out. There's a couple of things that have confused me on the datasheet though.

1 - What's with the voltage values all being negative? Is this characteristic of p-channel FETs and just indicates the polarity of the flow through the device, or does it mean something else?

2 - I thought I'd chosen a good logic-level option (V_GS(th) is listed as -2 to -4V). However it won't turn off until a full 5.7V and has a fading ramp between about 5.2 and 5.7V. I tested three and they were identical. Am I completely misunderstanding the spec here?

I'm powering the circuit with a 6AA pack, a 1k resistor and a 3mm yellow LED, and using my power supply to trigger the gate. Both grounds are tied together. Is the threshold voltage an absolute or related to the other voltage?

It's just occurred to me as I'm typing this that the -2 to -4V could mean 2 to 4 volts below Source voltage, which would mean the 5.7V being 3.3-ish volts below the nominal 9V I could expect from 6AAs is about right. Am I on the right track?

[Skimask]

+V(ds) = -V(sd)
-V(ds) = +V(sd)
It's all about the point of reference.

[luxfx]

Very good to know, thanks. Nice to get something confirmed

[forrestc]

1 - What's with the voltage values all being negative? Is this characteristic of p-channel FETs and just indicates the polarity of the flow through the device, or does it mean something else?

Sounds like you figured it out, but just to add one more note:

Let's assume you're using a pchannel fet to turn on a very large voltage.  Say 50V.

The source gets attached to the voltage you are switching in this case, 50V.

On the gate:
Off is the voltage you're switching (50V in this case).
On is the addition of Vgs(th) to the voltage.  this is also negative.  Assuming a -5V threshold,  The fet would be on at 45V.
Gate destruction is at the voltage your switching added to the Vgs(max) which is negative.  Assuming this is -15V, you'd blow the fet at 35V.

The last piece is the one you have to worry about if you're switching high voltages.

[luxfx]

Gate destruction is at the voltage your switching added to the Vgs(max) which is negative.  Assuming this is -15V, you'd blow the fet at 35V.

Very interesting, somewhat counter intuitive. My first thought was "how on earth would 45V be safe but a lower V, 35, would blow the fet??" But let me see if I've got this one figured out. It's because the issue isn't the discrete value, but due to the differential / potential between gate and source, right?

[Zero999]

Some schematics would make this easier to understand.

[luxfx]

Some schematics would make this easier to understand.

Here, the attached sums up what I've got.

My ultimate goal is to use a pair of mosfets to switch between two available power sources, so as I'm experimenting I'm using two different voltages from different sources.