Electronics > Beginners
JFET mysteries
<< < (4/5) > >>
nick_d:
Regarding the complaint that power MOSFETs are hard to drive (high gate capacitance, etc), if you only want to switch light loads, the 2N7000 is an N-channel MOSFET that can drive 300mA and has a tiny gate capacitance (maybe 10s or 100s of pF). Apparently a BS250P is a reasonable P-channel equivalent of a 2N7000. If you want high=on then the simple solution is to drive it through an inverter such as a 74HC04. Of course this setup requires everything to be at 5V, otherwise the high side driving is rather tricky.

If the voltage being switched is up to about 15V then an appropriate MOSFET can be found such that a gate voltage of 0V (or essentially -15V relative to your supply that's being switched) won't kill it. If it is more, you will have to use either bootstrapping or isolated drive (transformer or opto coupled). If you are forced to do it this way, it is often just as easy to have the gate drive sitting above the rail, for instance suppose your supply is 30V, then you could use the 30V as the ground of an isolated gate driver chip and provide another voltage say 10V higher as the supply of the gate driver chip. If you do that you can use N-channel MOSFET as the high side switch and save money and stocking costs. And interestingly you have high=on as requested, provided the isolated gate driver chip doesn't invert the sense.

As to JFETs I do not know, but I learned a lot from this thread. Thanks!

cheers, Nick
David Hess:
They used to make higher voltage and higher current "power" JFETs but they were never competitive with bipolar power transistors or power MOSFETs because of high channel resistance.

4 pin MOSFETs with a separate substrate connection can be used as bidirectional switches and choppers but they were never very common.  The same parts made good RF amplifiers.  Linear Systems and Calogic make some small signal ones.  Micrel (Now Microchip) has a couple of 4 pin p-channel power MOSFETs intended for high side power applications; it is too bad they are not higher voltage and available as n-channel parts also.  I have no idea how to find their selection guide on Microchip's atrocious web site but the highest voltage part number is MIC94030.

The cheapest solution is to use a pair of back-to-back power MOSFETs but if you want good small signal performance, then JFETs and the above 4 pin MOSFETs from Linear Systems and Calogic are the way to go.


--- Quote from: nick_d on December 31, 2018, 11:11:13 am ---Regarding the complaint that power MOSFETs are hard to drive (high gate capacitance, etc), if you only want to switch light loads, the 2N7000 is an N-channel MOSFET that can drive 300mA and has a tiny gate capacitance (maybe 10s or 100s of pF).
--- End quote ---

The 2N7000 is 60pF input, 25pF output, and 5pF reverse maximum capacitance which is actually very high for a small signal application.  A 3N part or JFET will have 1/5th to 1/10th that much capacitance and can be 10 times faster.

For switching LEDs this does not matter but even the smallest vertical power MOSFET is useless for something like a sample and hold application.
Kleinstein:
Those 4 PIN MOSFETs with separate substrate are difficult to get now. But the same element is commonly used in switch chips.  For some experiment is might be even worth getting the old CD/HEF 4007 chip (old CMOS Logic) - it includes such FETs with no too many connections.

There are also depletion mode MOSFETs, that in many respects behave similar to a JFETs (turn on at 0 gate voltage, and turn off with a negative gate voltage). However they still have the reverse diode.
T3sl4co1l:

--- Quote from: David Hess on December 31, 2018, 11:28:59 am ---The 2N7000 is 60pF input, 25pF output, and 5pF reverse maximum capacitance which is actually very high for a small signal application.  A 3N part or JFET will have 1/5th to 1/10th that much capacitance and can be 10 times faster.

For switching LEDs this does not matter but even the smallest vertical power MOSFET is useless for something like a sample and hold application.

--- End quote ---

Small, usually-intended-for-RF MOSFETs are both damned hard to find, and also hard to use with respect to ESD -- even if you're quite good about ESD normally!

They're also in a bit of a death-valley market segment: too small to handle any real power (like a jellybean 2N7000), too large and too slow to do any serious RF (where PHEMTs, GaNFETs and such dominate).  Everything else is ASICs, even in modest quantities (100k's?) so there's no one buying discretes in quantity.

The smallest I know of is the RUM001L02 and brethren.  It's about 1/4th the size of a 2N7000, electrically speaking (for that matter, mechanically too -- it's a tiny bugger, hope your soldering iron tip is sharp).  This comes at the price of Vds(max) and Pd.  It's definitely a newer process, with Rds(on) being quite comparable, while Qg being so much smaller.

Any smaller (electrically speaking), and you should be looking at arrays (e.g., CD4007?), analog switches, oh or those "precision zero offset" MOSFETs by ALD ($$$), or doing it in an IC (which may involve a boring ADC-DSP-DAC scheme, that still wins on power consumption -- it's dumb, it works, get over it?..).

Tim
Kleinstein:
For RF use there are dual gate MOSFETs (e.g. BF 998).  In principle they should also work for logic switching if needed. One does not really need to actively use the 2 nd gate - it's quite common to have it at a fixed potential and than behave quite normal, like a small low voltage MOSFET with low input capacitance.

A slight problem is that they are so fast that they tend to oscillate quite easy at frequencies the entry level scope may no even show.
Navigation
Message Index
Next page
Previous page
There was an error while thanking
Thanking...

Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod