Driving a MOSFET power follower

[RichardJones]

I am trying to drive 1400pF worth of gate capacitance from +10v to -10v.

IRFP140N biased at 3A 12v. CCS underneath at 3A -12v.

For low distortion at high frequencies.. (10khz-20khz) I need a pre with low (<200ohms) output impedance.

Can I make a small JFET follower as a buffer before the MOSFET? i.e connect it to a set of +24 to -24 batteries... and run it at 50mA or so?

Thoughts? ideas? part #'s?

Thanks!!!!

 

[gromanswe]

See AKSA advice

at thread "Linearize..."

 

[tschrama]

Would a JFET give you that low output resistance? .. I don't know .. and are you shure aboput the 1400pf.. that looks a bit high ..

I once build a buffer for headphones with power MOSFET's and the bandwidth obtained with a 10KOhm seies resistance of the input pot was way more than expected.... i think the datasheet input capacitance is not that important

...maybe because since there's not very much voltage to be charged (it is a folower right? so all voltage relations are quite constant..)

gr,
Thijs

 

[Nelson Pass]

I say hook it up to your source through 200 ohms or
so and see what you get. Probably it will work
fine. My experience with followers is that around
1000 ohms you start seeing some real degradation.

Remember that the high Gate capacitance is mostly to
the Source pin, which is following, and you are not
really charging that capacitance nearly as much as you
might imagine.

 

[mlloyd1]

This is an excellent point I remember being told when I first started building MOSFET amps and was worrying about the capacitance!

...Remember that the high Gate capacitance is mostly to
the Source pin, which is following, and you are not
really charging that capacitance nearly as much as you
might imagine.

 

[gromanswe]

Todays MOS/HEXFET

Has much lower input capacitance,
then yesterdays.
And who would say he knew better than Nelson??
Only a fool....

gromanswe
the fool on the hill

 

[Circlotron]

Miller effect

As Nelson said, it is not so much the capacitance but the change in voltage across that capacitance. So if your gate goes up by 10v and your source follows it by 9v then G/S capacitance really only sees 1v change.

Now if you have a common source cct (not to be confused with no-name tomato sauce) as well as the driver seeing the full capacitance of G/S because the source stays put, the gate/drain capacitance is *multiplied* approximately by the voltage gain of the cct. If the gate goes up by 1v and it causes the drain to come down by 10v then the gate/drain came 11v closer for 1v drain voltage change, making the G/D capacitance look 11 times larger. Yoicks!

GP.

 

[RichardJones]

low Ciss MOSFETS

Then this begs the question:

Any low Ciss MOSFETS that can handle a bias of 3A at 12v with high S and low Rds ?

The IRFP140N is 1400pF...

Does it help by putting in a low value (200mOhms) resistor in the source of the MOSFET? sort of local negative feedback?

Thanks for this insight... I just assumed that this capacitance is for any difference between gate and source... but if the source follows the gate then the capacitance is not really an issue.

Richard

 

[gromanswe]

I prefer
IRF530 HEX 600pf N-channel 0.18 ohms
IRF540 have 1450pF 0.06 ohms

 

[HarryHaller]

Mosfet power follower

Mr. McGuire: I just want to say one word to you - just one word.
Ben: Yes sir.
Mr. McGuire: Are you listening?
Ben: Yes I am.
Mr. McGuire: 'Transconductance.'

 

[gromanswe]

Re: Mosfet power follower

HarryHaller: 'Transconductance.'

IRF530 IRF540 are still my favourites
of various other reasons
you did not have me your way

 

[dshortt9]

Did I miss something? Show me a JFET that runs at 50 ma.

 

[gromanswe]

JFET-technique <20mA

I haven't any knowledge about JFETs, that can handle more
than max 24 mA.
A good working point could be some 10 mA.

I have made stages mostly with 2-6 mA working current.
Also is highest voltage limited.
most wouldn't handle more than 40, usually 30
Drain to Source

Anyone used BF245B for audio?
It is a HF-JFET. High Frequencies, radio equipement etc

 

[yves]

low Ciss mosfets

Try RF power mosfets. They exhibit extremely low
capacitances, which is why they are great for RF.
However, you will definitely need an RF spectrum
analyser to check for oscillations. These things oscillate
at the slightest provocation and with the least
parasitics. But once you get them to work
stably, they are stable. And I *believe* they make
the best audio mosfets too.
See www.advancedpower.com. I would like to
purchase small quantities, but I can't find a source
that would sell them.

One other alternative with audio freq mosfets
is to operate them such that their capacitances
are minimised, or the change in capacitances
are reduced. If you look at the C-V curves
for any audio mosfet (or any mosfet for that
matter) they are anything but linear vs voltage.

A cascode works well to reduce the Miller effect,
but some people find that the sound is not
the best, YMMV.

yv

 

[gromanswe]

Yes, Power MOS for RF

MUST be good
what else would price them so high
I am no stranger for to use
"HF" "RF" devices for Audio.
On the contrary. Think it is way narrow
not to test that option ....

Speed can be good

 

[jcarr]

dshortt9:

>Did I miss something? Show me a JFET that runs at 50 ma.<

Among the devices that I have actually tried, I'd suggest the following.

N-ch: 2N4856. Minimum idss, 50mA. Maximum idss, 175mA.

The closest complementary that I have been able to find is the 2N5114. Minimum idss, 30mA. Maximum idss, 90mA.

regards, jonathan carr

 

[dshortt9]

Thanks, I didn't know these critters existed.

 

[RichardJones]

problem with amp

ok,

I hooked up my MOSFET power follower last night.. I have zener protection diodes on all the mosfets.

I connected my batteries and the lower 0.2ohm resistor in the CCS turned bright red!!!

It got so hot that some solder joints melted..

well this isn't right.

So I checked some resistances... the resistance across the mosfets drain/source is 220 ohms.... hmmm...

So I think I fried both MOSFETS on the left channel.

And I have no clue how I did it.

So I will replace the mosfets and the MPSA18 sense transistor on my CCS.

I think the 0.2 ohm is still ok.

very strange.

any ideas / clues?

 

[mrfeedback]

Re: problem with amp

..any ideas / clues?

Hi Richard,
Yup use a variac with a lamp in series with the output, or wire a lampholder across the amp fuseholder, and have a selection of lamps.
If you have a fault the lamp will do the dissipating.
This saves a lot of grief whilst doing the initial running up.

Regards, Eric.

 

[RichardJones]

grief with amp

Well the point is that the amp was working fine before. 0.7v across the 0.2ohm resistor... about 3.25A bias..

My heatsinks get nice and toasty... then I can trim output offset down to <10mV per channel... adjusting the upper mosfets bias.

well I will check everything tonight... very odd..

Unless I blew my MOSFETS before I turned it on.... very possible.

oh well.

I guess thats whats great about tubes... really cannot blow them up.