5
\$\begingroup\$

The idea struck my mind while I was looking at this typical application for the PCF8563.

alt text

This RTC requires typically 250nA (interface inactive, like in circuit in power-down), but a general-purpose diode like the 1N4148 already leaks 10% of that. The gate reverse current of a JFET is only a fraction of that; max 1nA for the MMBF4391.
The whole circuit is very low voltage and dito power. Are there reasons not to use a JFET as a blocking diode here?

\$\endgroup\$
  • 1
    \$\begingroup\$ They do make low-leakage diodes - on Digikey I found a diode with 1nA leakage at 125V, and if you want to spend $16, you can get one that has only 3pA leakage. If powered-down operation takes 250nA, then one of the low-leakage diodes will lose less than 0.4%. \$\endgroup\$ – W5VO Dec 2 '10 at 15:52
6
\$\begingroup\$

1N4148 is a general-purpose, legacy diode that you shouldn't use if you're looking for some level of performance.

From a very cursory glance at Digi-Key, the BAS116 (in stock, 0.25/10, 0.035/reel) has a typical ~400 pA, max 5 nA leakage at 75 V, 25 °C.

Not sure if it's an actual "series", but BASx16 seems to go to BAS416, a 75 V, 200 mA diode with 3 pA typical, 5 nA max reverse leakage @ 75 V, 25 °C. In-stock at Mouser.

The JFET you linked (MMBF4391) doesn't shut off until -10 V (though the MMBF4393 in that datasheet shuts off at -3 V), and the quoted 1 nA leakage is at -12 V (for either -1 or -3), which you can obtain? "Shutoff" is defined as where the JFET conducts 10 nA.

The diode's leakage is specified at 75 V, so if you're running it at 3-5 V the leakage will be much lower.

\$\endgroup\$
  • \$\begingroup\$ 3 pA vs. 5 nA, what the heck?! That's a factor 1000! But I admit, I should/could have had a look at low-leakage diodes first. That said, the JFET doesn't do bad at all, even better than the BAS416 (1nA vs. 5nA, max. values, which are the only ones that count) \$\endgroup\$ – stevenvh Dec 3 '10 at 8:09
  • \$\begingroup\$ @steven: Added more to answer about the JFET. It may be an interesting use, but it seems somewhat odd. Using maximum/minimum parameters when characterizing the power use of your device is extremely conservative and is more of an engineering-side thing than sales-side. \$\endgroup\$ – Nick T Dec 3 '10 at 15:48
  • \$\begingroup\$ Well, I'm not a sales person! If you design with typical values, the Law of Conservation of Misery says that the actual value will be closer to the maximum specified. And your circuit may not work, or at least not to its specifications. Worst thing is that you can't even blame the manufacturer. \$\endgroup\$ – stevenvh Jun 16 '11 at 13:41
  • 1
    \$\begingroup\$ This should not be the accepted answer, Richard Thiessen's answer is better. Nick, you're misreading the datasheet for the JFET. You say the MMBF4391 doesn't "shut off" until -10V, but you're referring to Vgs(off) max, which is for Vds=20V. When used as a diode, a JFET has a Vds of 0v since the source and drain are connected together as the N junction. At room temperature, the MMBF4391 has a superior leakage of only 1nA. Using a MMBF4117 instead would be the best solution, with a leakage of only 10pA max at room temperature, and 100fA typical as noted in Richard's link. \$\endgroup\$ – JDW Mar 2 '16 at 22:54
  • 1
    \$\begingroup\$ To clarify, Vgs(off) max refers to the Vgs at which Ids drops below a certain level (1nA in the MMBF4391 datasheet), given a certain Vds bias (20V in the datasheet), and says nothing about Idg. The value of Vgs(off) in the datasheet is completely irrelevant to the leakage current. What matters is Igss, the gate reverse current (i.e. leakage current) when the drain and source are connected together (and used as the N junction of the "diode" for an N-channel mosfet). Note that this leakage current decreases with a lower applied bias. \$\endgroup\$ – JDW Mar 2 '16 at 23:16
6
\$\begingroup\$

Jfets can be used as diodes. TI has an app note on this where a diode connected JFET is used to protect an op amp input from overvoltage. Leakage current was in the sub 100 femtoampere range (http://www.ti.com/lit/an/sboa058/sboa058.pdf). See also: Current Sources and Voltage References: A Design Reference for Electronics Engineer p180-183

The 2N4117A referred to is no longer available, but the MMBF4117 is a good replacement, and has the same guaranteed maximum leakage of 10pA at room temperature and a reverse bias of 20V. Note that the leakage current is less at lower voltages, so since Vdd is at most 5.5V for the PCF8563, the leakage current is even less.

For best results connect the source and drain together as one terminal of your diode and the gate as the other.

A note regarding temperature and this applies to all PN junctions, diodes (including low-leakage diodes), JFETs, BJTs MOSFETs (etc):

The reverse-biased leakage of a P-N junction has a strong positive temperature coefficient, approximately doubling for each 10°C increment in temperature. This exponential increase racks up quickly as shown in the normalized graph of figure 1. At 125°C leakage climbs to approximately 1000-times the room temperature value.

source:https://e2e.ti.com/blogs_/archives/b/thesignal/archive/2012/11/14/temperature-effects-on-input-bias-current-plus-a-random-quiz

\$\endgroup\$

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.