Low side FET switch

Go To Last Post
25 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I'd like to use a N FET as a low side switch to switch < 20mA from a battery(coin cell, actually, so around the 3V mark). The switching voltage will only be around 1.6V, which is why I ask the question, as I am no expert on this analogue stuff.
I realize that a high side switch is often preferred, but that, I believe, would require another transistor, and I'd like to keep component count and cost to the bare minimum, i.e. one FET one resistor.

Any help appreciated.

John

Four legs good, two legs bad, three legs stable.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

John,
Take a look at using a 74HC05, HCT05 or similar. The output stage can handle the 20 ma load current and possibly the 1.6 "switching" voltage. What supply voltage does your system run on?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

So were is 3V and where is 1.6V ?

If you just look for a MOSFET with Vgs around 1.6V, then search at Digikey and put the filter accordingly. There are plenty of MODFETs with Vgs lower than 1.6V.

George.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

@Chuck-Rowst and George. 74XXX all take power. I have a device that runs from a 3V coin cell, but I wish to effectively switch it off to save power unless it detects 1.6V(which comes from another device).
I will search at Digikey, I just wondered if there were any "gotchas" I should be aware of.

John

Four legs good, two legs bad, three legs stable.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Once it is switched off, does it ever need to go back on?

Jim

Jim Wagner Oregon Research Electronics, Consulting Div. Tangent, OR, USA http://www.orelectronics.net

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

ka7ehk wrote:
Once it is switched off, does it ever need to go back on?

Jim


Yes, the next time it sees the 1.6V from the other device!

Four legs good, two legs bad, three legs stable.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Bear in mind that once you have disconnected the low side you lose the common control signal path (GND). To avoid this you can power your device from an output of a a single gate noninverting buffer like 74VHC1GT50 - it draws less than 1 uA.

Warning: Grumpy Old Chuff. Reading this post may severely damage your mental health.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

MBedder wrote:
Bear in mind that once you have disconnected the low side you lose the common control signal path (GND). To avoid this you can power your device from an output of a a single gate noninverting buffer like 74VHC1GT50 - it draws less than 1 uA.

I don't believe that is a problem.
Let me be more specific:
I have a device that plugs into the headphone socket of a phone. It communicates with the phone via the microphone input on the headphone jack(this audio is capacitively coupled in both legs). The device draws between 5 and 10mA when active from an internal coin cell, and there is no off switch. I hope to use the electret mic. bias voltage to switch the low side N channel FET so that the gadget is only drawing power when it is plugged in to the phone. The problem I have is that most FETs seem to want a higher Vgs than this.

I could use a standard transistor and a P channel FET to switch the high side, but that would involve more parts and more cost.
I could probably use a standard transistor as a low side switch, but I think it would have a higher on resistance, and the base current would be higher than for a FET.

Four legs good, two legs bad, three legs stable.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

if you put a 1Meg Ohm resistor between the source of the N-fet and the gate and then a series resistor of 10K to an AVR pin you can switch High side with just 2 resistors that is not to much overhead I assume.
change the AVR pin from input without pull-up(off state) and the output low (on state)
you might need to play a bit with the component values, but this should work.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

meslomp wrote:
if you put a 1Meg Ohm resistor between the source of the N-fet and the gate and then a series resistor of 10K to an AVR pin you can switch High side with just 2 resistors that is not to much overhead I assume.
change the AVR pin from input without pull-up(off state) and the output low (on state)
you might need to play a bit with the component values, but this should work.

I'm not trying to use an AVR to switch anything, I want the AVR(yes, there is one, although I don't know how you guessed - I haven't mentioned it before) and its associated circuitry to be isolated from the coin cell that powers it unless the low-side FET switch is turned on by the Electret mic bias voltage.

[Edit} Sorry, forgot to draw a resistor between gate and 0V[/Edit]

Attachment(s): 

Four legs good, two legs bad, three legs stable.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I think this low voltage is quite a problem, even for very low threshold MOSFETs. A 1.6V threshold is just that, a threshold, it starts to conduct slightly then, not 'snap' on into full conduction. How much current do you need?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Look at BSS138. It will get you a minimum of 1ma at Vgs = 1.5V (worst case condiition, typically 1ma at 1.3V). These are Faichild specs. They may vary slightly with supplier. Sources include NXP, On, and others.

JayJay is correct, however, that nothing has a hard on/off. FETs or bipolars are pretty soft at low voltages.

Jim

Jim Wagner Oregon Research Electronics, Consulting Div. Tangent, OR, USA http://www.orelectronics.net

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

jayjay1974 wrote:
I think this low voltage is quite a problem, even for very low threshold MOSFETs. A 1.6V threshold is just that, a threshold, it starts to conduct slightly then, not 'snap' on into full conduction. How much current do you need?

As I said, between 5 and 10 mA.
Maybe I should go for the transistor/FET high side switch option.

Four legs good, two legs bad, three legs stable.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

NTZD3154N has a bit lower threshold voltage than BSS138 and adds ESD protection for the gate; but, it's a dual.
If willing to use some current, the OnSemi NSS series high current gain bipolar transistors will have greater ESD capability.

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

John,
sorry, missed that, thought that you wanted an avr to interface with the fet.

What is the output signal of the mic?
you say it is 1,6V but note that this is probably modulated so what are the absolute min max voltages?
In your drawing I would suggest taking a transistor instead of a FET. These are full open when you give it 1V on its base.
put a resistor network in to minimize the distortion from the transistor and to protect the transistor if the voltage goes way over what you think it will.
The minus with this low side switching is that you no longer have a common ground. ESD might give you strange failures on the FET(not the smokey type, but sudden end of life)
you could use 2 transistors. 1NPN that drives a PNP that switches the voltage. IN that case it is easier to ensure that your device supply will not be modulated with the mic signal. cost 2 resistors and a transistor. savings probably filtering to keep the signal and the power supply clean.
take 2 bc8x7c parts (4&5) you need 20ma from the battery(what is a lot for a cr2032( check datasheet of buttoncell that you use if it likes it or will give you trouble over lifetime as such). with a gain of lets say 100 you need 200uA from the base of the BC857c. this current needs to go into the bc847c that also having a gain of lets say 100 tells that you need to be able to run 2uA into the base of the bc847c for it to be far enough open to drive the other transistor. In real life with those low currents the gain is higher, but taking a low number ensures that the transistor acts as a switch. you could easily take a gain number of 200 if you like, then the base current of the npn should only be 500nA.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
What is the output signal of the mic?

It is the bias for the Electret mic that I want to use, not the audio.
Yes, it will be modulated at times(my device is modulating the phone's mic input - that's its raison d'etre) but the modulation is only 10 or 20 mV pp.

I think the easiest thing to do is to try using a common or garden NPN BJT as a low side switch and see what happens. My concern is that the collector/emitter voltage drop will be too high.
Maybe I'll model it with LTSpice, if I get a moment.

Thanks to those who have replied with suggestions.

John

Four legs good, two legs bad, three legs stable.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

from the head a bc857 does only 150mV, so that should not be the end of the world having that as a drop.

if the modulation is only 20mvpp then you should be OK building the schematic I described. Note that for the BC557 you can also use a P fet like the NSD352 or similar, but then you have to take a look at the threshold as it may become an important player.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

@meslomp
Where does the 20mA you mention in your earlier post go?

Quote:
) you need 20ma from the battery(what is a lot for a cr2032( check datasheet of buttoncell that you use if it likes it or will give you trouble over lifetime as such).

Four legs good, two legs bad, three legs stable.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
I'd like to use a N FET as a low side switch to switch < 20mA from a battery(coin cell, actually, so around the 3V mark).

I asume it is the current that your 'device' is consuming as stated in your first post. I took the worst case number being the 20mA.
So using a BC557C in the high side then the current will run from emiter to collector and then into your 'device'

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

meslomp wrote:
Quote:
I'd like to use a N FET as a low side switch to switch < 20mA from a battery(coin cell, actually, so around the 3V mark).

I asume it is the current that your 'device' is consuming as stated in your first post. I took the worst case number being the 20mA.
So using a BC557C in the high side then the current will run from emiter to collector and then into your 'device'

OK, sorry, I misunderstood you. I thought you were saying that the switching circuitry would consume 20mA.
To be accurate, my device consumes nextMy device consumes < 1mA most of the time, and the peak current I have measured is still < 6mA, and that's for fairly short bursts(100mS or so). When I stated <20mA I didn't have my notes to hand.

John

Four legs good, two legs bad, three legs stable.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

After much searching, it seems that I am unlikely to find a solution.
BJTs have too much collector/emitter voltage drop, and FETs require too much gate source voltage.
Even using an NPN BJT and a P FET is a problem, since I can only guarantee 2V6 to 3V? supply.

I thought about have the AVR go into power down mode, and be woken up by the mic bias voltage, but I can't be sure that, with a new coin cell, the mic bias will be high enough.

So I think what I will try is to feed the mic bias voltage into a BJT, take the collector to an AVR pin with a pullup(probably the internal one) and use that to wake the AVR. I suppose I could disable pullup after I'd woken up, and then re-enable at intervals to check if I'm still plugged in, but it's probably not worth the bother.

Now I need to check what sort of leakage there might be through the BJT collector/emitter.

Would that work, or am I missing something?

John

Four legs good, two legs bad, three legs stable.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hmmmm. Just found this:http://search.digikey.com/uk/en/products/ZUMT618TA/ZUMT618TR-ND/272394

Seems to have a really low saturation voltage.

Not entirely sure how to read these characteristics, though. I'm guessing you need to pour base current in to achieve such a low saturation voltage.

Four legs good, two legs bad, three legs stable.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I always use a low side FET. One like this one:
http://www.nxp.com/documents/data_sheet/PMV30UN.pdf

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

John_A_Brown wrote:
Hmmmm. Just found this: ... ZUMT618 ...
Thanks for the Zetex ZUMT618; I was not aware of these Zetex/Diodes parts. Simulates better than OnSemi NSS12201. The advantage of the high current gain (hfe) AND low Vce-sat bipolar transistors is the reduced base current for a given collector current and Vce.

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

John_A_Brown wrote:
Now I need to check what sort of leakage there might be through the BJT collector/emitter.
Almost always very low IF base current is correctly controlled; FET's d-s leakage can be a problem especially at elevated temperatures.

"Dare to be naïve." - Buckminster Fuller