Low leakage FET switch

[fala]

Hello board, I want to use solid state to switch 10A with leakage of less than 100pA, so probably I have to use MOSFET. Is there any chart that shows how IDSS and IGSS change when VDS and VGS change?
Thanks alot.

[A.Anand Srinivasan]

those curves can be found only in the datasheet of the device you use.... so when you decide upon a MOSFET just get the datasheet and have a go through....

[fala]

Anand, In all power MOSFET datasheets eg. IRF540 (as I said I want to switch 10A), there are such charts but in all mosfets that I've seen those charts resolution is not appropriate for my application as I said I want less than 100pA leakage. Those charts show currents at best at 100 mA resolutions so how should I know in what VDS and VGS I have IDSS+IGSS less than 100 pA???
Thanks

[A.Anand Srinivasan]

whenever you talk about power MOSFETs and power transistors you gotta be ready for the large leakage that is inherent in them.... i dont think power MOSFETs with that low leakage is easy to find.....

[fala]

Anand, I know they have large leakage. I want to lower leakage with bootstrapping but I need to know their IDSS/VDS and IGSS/VGS at low currents

[A.Anand Srinivasan]

give me a clear picture of your circuit because as far as know i think you would surely be needing a capacitor to carry out bootstrapping and capacitors are even worse when it comes to leakage.... and i hope you are using a discrete one....

[rvionics]

which leakage is important to you - in your solid state switch
1. When the switch is off then between the contacts or
2. Between your control signal to your power control lines?.

[fala]

hi rvionics, both is important for me. If I had the charts of IDSS vs. VDS and IGSS vs. VGS of power mosfets when VDS and VGS are in mV range then I could solve the problem but unfortunately all power mosfets that I've studied their datasheet only show VDS and VGS for relatively high voltages so I have no idea about appropriate design.

[rowokii]

If you really need data that is not in the datasheet, only thing to do is either contact the manufacturer's applications group (and hope they actually respond) or measure the data yourself. If you are potentially a big customer, they tend to respond better....

Measuring the data doesn't sound too difficult either. Get a variable power supply and measure current while you vary the voltage. Since you are only interested in low currents, set the current limiter on the power supply low so you don't blow up the MOSFET.

[A.Anand Srinivasan]

i dont think measuring by yourself is a good idea because the gate current is gonna be very very low and it needs special instruments and probes to measure them......
our DVM wont even detect them....

[rowokii]

Anand, I believe you were the one who mentioned the power fets are leaky. Random pick : irl520ns 10 A hexfet has max gate leakage of 100nA @ 16V. A good hand held DMM has something like 10nA resolution (Fluke 83)
http://us.fluke.com/usen/products/specifications.htm?cs_id=34473(FlukeProducts)&category=HMA(FlukeProducts)

And a good benchtop DMM has 100pA resolution
http://us.fluke.com/usen/products/88...keUnitedStates

So yes, it is low current and there might be gotchas but it sounds possible to do. Also, even if you don't get down that low, fala might be able to take data that is at least better then what the datasheet has and maybe able to extrapolate down.

Oh, also, I would image you would have to do the measurement with the device heated (worst case temp since it needs to turn off after dissipating large power) leakage of the device should be rather high and make it easier to measure. (Sucks for design though....)

Also, if fala is going to verify his design, he will have to be able to measure down to 100pA level, so might as well start thinking about and working out the measurement solution.

[A.Anand Srinivasan]

rowokii i referred the fluke 80series link posted by you
http://us.fluke.com/usen/products/specifications.htm?cs_id=34473(FlukeProducts)&category=HMA(FlukeProducts
and it has nothing specified in terms of minimum measurable current and i'm supposing that you are talking about the resolution of the multimeter....

if you see the second link that you posted for the benchtop DMM
http://us.fluke.com/usen/products/88...keUnitedStates
it can be seen that even a precision and costly instrument like that can only measure from 10µA to 10A only and its resolution is 100pA.....

you are confusing resolution for minimum current measureable by the instrument....

measuring a current in nA itself would require costly instruments as far as i know....

[rowokii]

I take resolution to mean just that, minimum value the instrument can resolve... mostly. If it's a cheap no-name company I might wonder if the spec'd resolution is just display resolution and not instrument resolution. But, if fluke told me it has a 10nA resolution or 100pA resolution, I would tend to believe them unless you have prof otherwise.

[A.Anand Srinivasan]

please check it for yourself.... the range has been given for the desktop DMM to be 10µA to 10A clearly and only the display resolution is 100pA and it is not the instruments resolution....

the display resolution has been given right below the range....

[rowokii]

I don't see where you get the 10uA.

The range is given as 100uA to 10A. This does not mean the unit can only measure 100uA to 10A. It means, it has range settings that allows measurement from 0 to X.

Here's the table from extended spec:

Range Full Scale Resolution
(6½ Digits) 4½ Digits 5½ Digits 6½ Digits
100 μA 100.0000 μA 10 nA 1 nA 100 pA
1 mA 1.000000 mA 100 nA 10 nA 1 nA
10 mA 10.00000 mA 1 μA 100 nA 10 nA
100 mA 100.0000 mA 10 μA 1 μA 100 nA
1 A 1.000000 A 100 μA 10 μA 1 μA
3 A[1] 3.00000A 1 mA 100 μA 10 μA
10 A 10.00000 A 1 mA 100 μA 10 μA

As you can see, it has 7 range settings and at 100uA range setting, the FULL SCALE (MAX READING) is 100uA.

EDIT: Yes, I admit reading the extended spec, minimum accuracy in 100uA range setting is 20nA, but it still does not give what the minimum resolution is and it is still a lot better then 10uA you seems to suggest.

[A.Anand Srinivasan]

oh that is cool.... i heard range of a instrument is something which denotes the interval of values that a instrument can measure.... so if 100pA can be measured(you said 100pA is the instrument resolution) then why haven't they given the range from 100pA to 10A.... if 100pA is measurable then it should be the lower limit of the range and why have they put 100µA...

can you please define range for me....

[rowokii]

Just because we can measure 100pA does not mean it will be accurate. In fact, the data sheet gives lowest accuracy at 20nA + % of reading.

I thought I did explain what range is in above post.

Apologies to the other members.... we seemed to have gotten way off topic. Anand, maybe we can take this off-line to email?

[artem]

fala what are you going to switch ? Power ? If it is , is the switching time important to you ? Please describe application in details .

[fala]

Hello artem, Yes I need 100KHz switching speed, relays are out of question.
More Detail, The instrument is used in electrochemical lab and it applies 1-10A current with different frequencies(up to 100KHz) and with different duty cycles.