High side high voltage current sensing with low common mode current sns amp.

[AJrocks]

I am trying to implement a high side current sensing circuit which has low common mode voltage. I found this application circuit in LTC2946 datasheet.

Here they have shifted the amplifier's GND to 57v using the circuit highlighted in the attached image below.

I was wondering that, if same thing can be implemented using just a zener and a resister. Then what do they have those extra mosfets and transistors for.

regards
Abhijeet

[T3sl4co1l]

Bizarre, 2N3904 (TO-92, THT) beside SOT-23 and SOT-223 (SMT)?

I'd just do an emitter follower (PNP) boosted zener, a very old regulator topology.

Has to handle a few mA for the LEDs.  Beware of voltage and SOA limitations, and current or power limiting would be nice to protect it.  These are counter-indications for BJTs, but there isn't much selection for P-ch MOSFETs, let alone depletion mode types, which may be behind their choice (the '6520 handles well under 1mA here).  At some point, a simple transformer isolator (DC-DC) would be more desirable, etc.

Tim

[Renate]

Well, they might be using 57V. What are you using?
If it's under 36V I'd use TI parts designed for this, INA226.

[Wimberleytech]

Why all the extra transistors? 
That chip has an internal shunt regulator which is limited to 35mA.  This application schematic supports high side from 6V to 300V.  I suppose the resistor chosen to give 35mA at the 300V end would be too high to operate properly at 6V.  So, another solution is needed.

An external zener-resistor and PNP as has been suggested might work but then you will still have a wide range of currents through the zener as the supply goes from 6 to 300 V.  The solution proposed in the application schematic builds a current source to supply the zener current using the depletion MOSFET.  If the zener supplied ground via the MOSFET current source, it might not be low enough impedance (ac).  So, they beefed it up with a little feedback using the other transistors.  Not sure why the other MOSFET is needed in series with the 3904.

[David Hess]

It would have been nice of them to use the depletion mode MOSFET symbol.

The circuit is straightforward.  M3 is a high voltage constant current sink to bias the zener diode.  The Vbe of Q1 compensates for the Vbe of Q2.

Now normally a higher power version of Q2 without R11 would seem to be all that is needed however driving that low ESR 0.1 microfarad decoupling capacitor directly with an emitter follower could be problematical and adding the emitter series resistor compromises load regulation.  So instead, Q2 acts as an error amplifier to drive Q3 which has M1 in a cascode to provide high voltage capability.  I suspect you could do away with M1 if Q3 was a higher voltage and power transistor but using M1 means one less transistor part number and 2N3904s are dirt cheap.

[Wimberleytech]

It would have been nice of them to use the depletion mode MOSFET symbol.

They do show a different symbol for that MOSFET. 
The way it is configured is also a giveaway.

M1 in a cascode to provide high voltage capability. 

Yup, the purpose of M1 is obvious on second look

[Jwillis]

Bizarre, 2N3904 (TO-92, THT) beside SOT-23 and SOT-223 (SMT)?

You can get the 2n3904 in a SOT-23 package. Normally designated as MMBT3904 but I have seen it just designated as 2N3904.

[T3sl4co1l]

JEDEC designations specify package -- that's why for example there is 2N7000 (TO-92 MOSFET) and 2N7002 (SOT-23 MOSFET).  I would at the very least beware of anyone trying to sell you a "2N3904" in "SOT-23"!

Tim

[Jwillis]

JEDEC designations specify package -- that's why for example there is 2N7000 (TO-92 MOSFET) and 2N7002 (SOT-23 MOSFET).  I would at the very least beware of anyone trying to sell you a "2N3904" in "SOT-23"!

Tim

I guess I should have been a bit more clear . I wasn't specifically talking about sellers . I was talking about schematics having a different reference than  the material list  .Whether that be a mistake or for the sake of labelling space on the schematic it's self I can't be certain.  In this particular case I don't know what package the OP is using.  But one should still refer to a material list to get the correct package type.     

[AJrocks]

I made a mistake. The amplifier GND is at 57V When the high side is at 100V and a 6.2V zener is used.

In the simulation of this circuit with 300V as high side voltage, our GND will be at 170V.

This was my first post on this forum and your answers were truly insightfull.

Thanks all
Abhijeet

[T3sl4co1l]

Ah yeah, LT doesn't exactly give a complete BOM in most app notes, could be they replace it with something else on assembly.  In that case then I guess the schematic is just lazy.

Tim