The Pass Pub: The High-End Off Topic Thread

[Gyuri]

This is perfect according to my opinion!

http://ugowlw.bay.livefilestore.com/...C_00980001.JPG

Gyuri

[luvdunhill]

I have a quick question for the masters of off-topic. My question is about the second graph on this Datasheet

I have 30V bench supply and a resistor to source 150mA of current. Since Vds = Vgs when testing (according to the Zenv3 documentation) and this value ranges from 3.20V to 3.55V depending on the sample. Ambient temperature is 60C for all samples and the devices are not heatsinked, but enclosed in a box with a probe to verify when 60C is reached to take the Vgs reading. I don't know if anyone has done things this way, but it seems to be very repeatable. Anyways, back to my quesiton... the second graph is at 25C and it looks like if we take the 3.50V for Vds = Vgs there should be 233mA of current. Would this value drop this much because of the case temperature? I just wanted to see if I was in the right ball park.

The compliments are about 1V less Vds on average (same voltage and current) and I can almost convince myself of this looking at the Datasheet At Vds = Vgs = 2V for example, the Nch says ~180mA @ 25C and the Pch says 100mA @ 25C. I'm not sure how to do some calculations to see if this difference would result in approximately 1V lower Vds I see for the Nch, is this possible?

[Nelson Pass]

Graph #4 looks to show about a 30% or so drop in Vgs over that
temperature range, so it appears reasonable.

[luvdunhill]

Originally posted by Nelson Pass
Graph #4 looks to show about a 30% or so drop in Vgs over that
temperature range, so it appears reasonable.

excellent! This makes since.

One more question. When paralleling MOSFETs, how do you determine the best value of the ballast resistor? I have four of each polarity of these paralleled with 2 ohm ballast resistors (sorta like F4 output stage), with the MOSFETs chosen at random. Now I've matched them to Vds 0.01V or so. Is this a case where one needs to be able to measure distortion? Could I just lower the values until the voltage drop across the resistors isn't equal? Is there somewhere I could read some more on this perhaps?

[luvdunhill]

ok, so let me try again

Starting with the F5, the resistors in the output stage are 0.47R and the voltage drop is 0.6V. 3W resistors are speced and this is roughly 4 times the power at quiescent current.

Now, with the F4, the resistors are 0.47R and voltage drop is 0.25V and 3W resistors are used.

So, in the case of F4, why not cut the 0.47R value in half, assuming well-matched devices, and you could parallel two 0.47R resistors together to get the lower value? At what point does it make sense to stop lowering this value, at the point where the voltage drop across this resistor varies by a certain amount, due to matching becoming increasingly important as this value decreases?

Is 4x the power at quiescent current used in the F5 a good rule of thumb to size output resistors?

[Nelson Pass]

Usually it's a matter of what class the OS will operate in, what
damping factor you want, and how much thermal stability you need.

As a simple rule, you want the lowest values that are consistent with
good current sharing (matching) and thermal stability. 0.47 ohms
is a convenient figure, but it can easily range from 0.1 to 1.0 per device.

[R-K Rønningstad]

Seems it is N.P.'s birthday. Congratulations Nelson!

All the best

RK

[steenoe]

Originally posted by R-K Rønningstad
Seems it is N.P.'s birthday. Congratulations Nelson!

All the best

RK

Seems like the most exciting fact, that this forum can come up with

[Nelson Pass]

Works for me.

[Zen Mod]

Originally posted by Zen Mod
re-post from Texas ......... wait - not from there ...... from Introductions :

http://en.wikipedia.org/wiki/Diyaudio_forum