Author Topic: Control JFET using MCU  (Read 2320 times)

0 Members and 1 Guest are viewing this topic.

Offline gaminnTopic starter

  • Frequent Contributor
  • **
  • Posts: 343
  • Country: 00
Control JFET using MCU
« on: August 20, 2019, 08:13:31 am »
Hi,
I want to control MMBFJ111 JFET using microcontroller (power supply = +3.3 V). JFET needs at least -3 V voltage to turn it off. There is also +5 V power supply available on the board.

How would you generate enough negative voltage and how would you turn off JFET using the negative voltage and microcontroller?

Thanks

Martin
 

Offline jbb

  • Super Contributor
  • ***
  • Posts: 1265
  • Country: nz
Re: Control JFET using MCU
« Reply #1 on: August 20, 2019, 08:24:02 am »
You may not need a negative rail at all... in some cases it’s possible to use a cascode circuit with an N Channel MOSFET.

Maybe you could post a little sketch of what you want to do?
 

Offline Ian.M

  • Super Contributor
  • ***
  • Posts: 13216
Re: Control JFET using MCU
« Reply #2 on: August 20, 2019, 08:44:49 am »
The  MMBFJ111's worst case Vgs(off) is -10V, not -3V, so to get guaranteed hard cut-off you'll need far more negative bias than you are expecting if either its (interchangable) drain or source are near 0V.  If you want a JFET that will actually cut-off with -3V on its gate,  try  MMBFJ113.

I echo Jbb's request for more information on your actual application, and we need the 'big picture' overview, not just the detail of what you think you need to do to solve your immediate problem.
 

Offline gaminnTopic starter

  • Frequent Contributor
  • **
  • Posts: 343
  • Country: 00
Re: Control JFET using MCU
« Reply #3 on: August 20, 2019, 08:50:06 am »
I don't have any sketch at the moment. There is just an MCU (with 3V3 power supply) which turns off and on N-JFET. N-JFET's source is connected to ground and drain is connected to 5 V via some load.

I don't have to use MMBFJ111 - anything with approx 30 ohms on resistance is OK. But I guess the lower the on resistance I want the higher cut off voltage the transistor has.
 

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 20357
  • Country: gb
  • 0999
Re: Control JFET using MCU
« Reply #4 on: August 20, 2019, 08:52:03 am »
Daft question: why are you using a JFET?

What are you switching?

Why not simply use a MOSFET? If you need a series resistance of 30R, simply add a resistor.
 

Offline gaminnTopic starter

  • Frequent Contributor
  • **
  • Posts: 343
  • Country: 00
Re: Control JFET using MCU
« Reply #5 on: August 20, 2019, 08:53:05 am »
When there is no power I want the switch to be ON. That is why I want to use JFET.

The lower on resistance the better.

I'm switching some weak load (5V voltage, max several mA current).
« Last Edit: August 20, 2019, 09:31:27 am by gaminn »
 

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 20357
  • Country: gb
  • 0999
Re: Control JFET using MCU
« Reply #6 on: August 20, 2019, 10:04:42 am »
How about the NSVJ3910SB3?
https://www.onsemi.com/pub/Collateral/NSVJ3910SB3-D.PDF
https://www.digikey.com/product-detail/en/on-semiconductor/NSVJ3910SB3T1G/NSVJ3910SB3T1GOSCT-ND/8538776

Also available in a dual package.
https://www.mouser.co.uk/datasheet/2/308/NSVJ6904DSB6-D-1368295.pdf
https://www.mouser.co.uk/ProductDetail/ON-Semiconductor/NSVJ6904DSB6T1G?qs=sGAEpiMZZMv4z0HnGdrLjlAuB7fe22XlaC1trqQwZfVqIC0BFd8NLQ%3D%3D

It has a cut-off voltage ranging from -0.6V to 1.8V.

Is the load switched by the FET isolated from the MCU supply? Can it be floated on the 3.3V supply without causing any issues?

If the load power supply is isolated and floating is on the 3.3V rail won't cause any problems, simply connect the gate to 0V and source to +3.3V.

How fast does it have to switch? The MCU can be programmed to generate a square wave, which can be converted to a negative voltage using a diode based voltage doubler, but it will be slow.


Failing that you need a negative supply and a level shifter.
 

Offline gaminnTopic starter

  • Frequent Contributor
  • **
  • Posts: 343
  • Country: 00
Re: Control JFET using MCU
« Reply #7 on: August 20, 2019, 10:15:07 am »
Not sure, what is NSVJ3910SB3's on resistance. Even though I'm switching only several mA I need  JFET with very low on resistance (e.g. 30 ohms).

No, the load switched by JFET is not isolated. The load is on 5 V rail, the same 5 V is connected to LDO that generates 3V3 for MCU.

The switch can be slow. tens of miliseconds are OK.
 

Offline janoc

  • Super Contributor
  • ***
  • Posts: 3925
  • Country: de
Re: Control JFET using MCU
« Reply #8 on: August 20, 2019, 10:17:25 am »
When there is no power I want the switch to be ON. That is why I want to use JFET.

That sounds like you want a depletion-mode MOSFET and not JFET
 

Offline gaminnTopic starter

  • Frequent Contributor
  • **
  • Posts: 343
  • Country: 00
Re: Control JFET using MCU
« Reply #9 on: August 20, 2019, 10:22:24 am »
That sounds like you want a depletion-mode MOSFET and not JFET

Yes, it looks like that e.g. BSS139 H6327 would fit.
 

Online Kleinstein

  • Super Contributor
  • ***
  • Posts: 15151
  • Country: de
Re: Control JFET using MCU
« Reply #10 on: August 20, 2019, 10:33:47 am »
The BSS139 or NSVJ3910SB3 look both OK. There is no explicit spec for R_on, but a curve for I vs V_DS that suggests some 10-20 Ohms as a typical value.

It would still need a negative control or supply, e.g. through the charge pump circuit shown by Zero999.
 

Offline gaminnTopic starter

  • Frequent Contributor
  • **
  • Posts: 343
  • Country: 00
Re: Control JFET using MCU
« Reply #11 on: August 20, 2019, 10:52:12 am »
OK, I think I will implement BSS139 and the charge pump circuit mentioned above.
 

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 20357
  • Country: gb
  • 0999
Re: Control JFET using MCU
« Reply #12 on: August 20, 2019, 12:56:26 pm »
The NSVJ3910SB3 has an on resistance of 30R or better, when ID <10mA, but the BSS139 is cheaper and has a lower on resistance, so go with that.

When there is no power I want the switch to be ON. That is why I want to use JFET.
That sounds like you want a depletion-mode MOSFET and not JFET
It just needs to be a depletion mode device, be it a MOSFET or JFET.
 

Offline janoc

  • Super Contributor
  • ***
  • Posts: 3925
  • Country: de
Re: Control JFET using MCU
« Reply #13 on: August 20, 2019, 06:44:09 pm »
It just needs to be a depletion mode device, be it a MOSFET or JFET.

Given that it is for a switching application, the MOSFET is likely a more logical choice.
 

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 20357
  • Country: gb
  • 0999
Re: Control JFET using MCU
« Reply #14 on: August 20, 2019, 07:36:09 pm »
It just needs to be a depletion mode device, be it a MOSFET or JFET.

Given that it is for a switching application, the MOSFET is likely a more logical choice.
Why?
 

Offline janoc

  • Super Contributor
  • ***
  • Posts: 3925
  • Country: de
Re: Control JFET using MCU
« Reply #15 on: August 21, 2019, 02:23:50 pm »
Why?

Smaller ON resistance and easier to control thanks to the isolated gate?
 

Online Kleinstein

  • Super Contributor
  • ***
  • Posts: 15151
  • Country: de
Re: Control JFET using MCU
« Reply #16 on: August 21, 2019, 02:48:50 pm »
The BSS139 depletion mode MOSFET and  NSVJ3910SB3 JFET behave quite similar if used in the normal drain to source polarity.
The advantage of the BSS139 is the likely lower price and usually tighter specs on the threshold voltage. The maximum DS voltage is also much higher. The JFET as the advantage of also working in the reverse direction (no body diode) - the input is nearly as high in impedance and the gate capacitance is considerably smaller.

Just for switching for one direction they should both work and could use the same circuit.
 

Offline SiliconWizard

  • Super Contributor
  • ***
  • Posts: 15797
  • Country: fr
Re: Control JFET using MCU
« Reply #17 on: August 21, 2019, 03:37:17 pm »
Of course it's much more common to use MOSFETs especially for power switching. I don't think there are many (if any?) JFETs around that can handle several Amps of current. Now for low-power stuff, they can have their uses. The absence of body diode Kleinstein mentioned is a good point and could be a distinct advantage depending on your particular application.

Now if it's for a switch controlled by an MCU's IO, and if the only reason you considered a JFET is that you needed a "normally ON" device, a depletion mode MOSFET seems more adequate here.

Also consider cost and availability to make a choice.
 

Online Kleinstein

  • Super Contributor
  • ***
  • Posts: 15151
  • Country: de
Re: Control JFET using MCU
« Reply #18 on: August 21, 2019, 04:18:57 pm »
There are powerful JFETs available, especially SiC and GaN versions. There are / were even a few enhancement mode JFETs with a high bandgap material.
 

Offline Ian.M

  • Super Contributor
  • ***
  • Posts: 13216
Re: Control JFET using MCU
« Reply #19 on: August 21, 2019, 07:01:12 pm »
Why is it a requirement that it be on with no power applied?
Unless you are driving something like a high sensitivity moving coil galvanometer (which absolutely do require the coil to be shorted to damp their motion for transport),  why  should it matter whether the FET is on or off when the 5V rail is off?

IMHO an ordinary low Vgs threshold enhancement N-MOSFET, with its gate pulled up to +5V can be used.   *IF* its a requirement that it be on when the MCU is off, drive it from another MOSFET (with a gate pulldown) or a NPN BJT,  to get a true open drain or collector drive that wont be pulled low via the I/O pin ESD diodes when the MCU's Vcc is not present.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf