7
\$\begingroup\$

I have a circuit which does something not important, but it results in an output being normally low, and then on activation climbs to a little below Vcc. At the moment, it appears my Vcc is 1.8 V and the signal gets up to between 1.2 V and 1.4 V. I want to use this signal to control the gate of an n-channel MOSFET which is grounded on the source and has the drain tied to Vcc through a large resistive load.

The problem is the threshold voltage of the standard MOSFETs I have appears to be too high, and I have to raise Vcc to about 2.2 V in order to get the circuit to trigger the MOSFET correctly. Are there better MOSFETs for this situation or some other design idea that would work well and solve my problem?

BTW, the switching speed is not really a consideration... we're talking 1-200 Hz for frequency.

Application Note, see page 3. I was hoping I could just use their circuit exactly and just mess with the tuning component, but it doesn't seem to work at 1.8 V.

\$\endgroup\$
  • 1
    \$\begingroup\$ How much current? \$\endgroup\$ – stevenvh Jul 25 '12 at 14:06
  • \$\begingroup\$ Woah, quick response. Again, very little... Let's say microamps? \$\endgroup\$ – NickHalden Jul 25 '12 at 14:09
  • 1
    \$\begingroup\$ How about using an NPN bipolar transistor? \$\endgroup\$ – Oli Glaser Jul 25 '12 at 14:14
  • \$\begingroup\$ Why does it have to be a FET? A bipolar can be turned on easily with 1.2 V. \$\endgroup\$ – Olin Lathrop Jul 25 '12 at 14:14
  • 2
    \$\begingroup\$ I was going to comment on "quick comment" and "simple question" but didn't. I should have. The question is NOT simple without the circuit diagram and component values. You should have included the app note reference in the question. As asked it made the question uncertain and led people astray. | The MOSFETS that I have suggested should work. The bipolar solution probably won't. \$\endgroup\$ – Russell McMahon Jul 25 '12 at 14:40
3
\$\begingroup\$

Go to eg Digikey
(1) CETSEMI make some superb parts. Available in NZ via agents. May be hard to find in US.

Example CETSEMI CEM8208
20V, 7A dual MOSFET. See graph below. At Vgs = 1V and 25C it will handle about 500 mA at OK Vds. Warmer is better :-).

enter image description here

(2) Search favorite sources for MOSFET N Channel logic
eg Digikeys MOSFET N Channel logic search results.
Select single MOSFETS.
Select Vgth range desired - say up to 400 mV.

The following are all under $1/1 at Digikey (last getting a bit dearer).

The BSH103 may do although a bit unhappy at low low Vgs

NTJS3157 - better

Better - Si4836Dy http://www.onsemi.com/pub_link/Collateral/NTJS3157N-D.PDF


(3) Bipolar NPN will do this with ease. For super low Vsat drive base at much higher current than collector. eg if you have 10 uA collector current then eg 100 uA or 1 mA base current will give much lower than usual Vce_sat.

\$\endgroup\$
  • \$\begingroup\$ Russell, he needs only a few uA, and VGSth for the CETSemi is specified at already 250 uA... \$\endgroup\$ – stevenvh Jul 25 '12 at 14:39
  • \$\begingroup\$ @Steven - that's a net advantage. Vgsth is specified for a given Vds at a give Ids - and at lower Ids than specified Vds will be lower. \$\endgroup\$ – Russell McMahon Jul 25 '12 at 14:44
  • \$\begingroup\$ But I was thinking that a gate voltage of less than the 500 mV VGSth may already saturate the output. I don't know how low you have to go to guarantee less than, say, 10 uA out. Which may be the maximum the load resistor will allow. \$\endgroup\$ – stevenvh Jul 25 '12 at 14:47
  • \$\begingroup\$ If you could explain which aspects of those MOSFETs make it good for my application, and why the BJT won't work it would be very helpful and I'd mark your answer correct. Thanks for your consistently helpful answers, Russell. \$\endgroup\$ – NickHalden Jul 25 '12 at 16:02
  • \$\begingroup\$ MOSFETS I suggested all have low gate threshold = turn on voltage - about 400 mV at 10 uA to 250 uA or so depending on device. MOSFET has zero DC gate current so essentially zero load on high z point. Bipolar draws current and is inherently low impedance. Can use large series base resistor and design accordingly but not at simple as with a MOSFET. \$\endgroup\$ – Russell McMahon Jul 25 '12 at 17:51
5
\$\begingroup\$

It isn't hard to find a FET with a \$V_{GS(th)}\$ of 1 V or less, but then a current in the microampere range may be a problem. \$V_{GS(th)}\$ is often specified at 10 or even 100 times higher current, so it may conduct at 100s of mV lower than that.

The solution may indeed be an NPN transistor. At near zero volt base voltage, you'll have a very low leakage current, but anything above 0.6 V will give you an output current in the µA range. A 100 kΩ base resistor will do.

\$\endgroup\$
  • \$\begingroup\$ And that base resistor is an ok load to put on the V_HIGHZ node? \$\endgroup\$ – NickHalden Jul 25 '12 at 14:25
  • \$\begingroup\$ Your R2 will be in that order of magnitude, and the base resistor replaces that. You can even increase the value to 1 M\$\Omega\$ if you want; the base current will then be around 1 \$\mu\$A, , so an HFE of a few tens will suffice to get 10 \$\mu\$A out. Nice device, by the way. \$\endgroup\$ – stevenvh Jul 25 '12 at 14:34

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.