current monitor

I might want to add a power supply current monitor to a product. The supply is 48 volts and I expect a huge amount of EMI. There are lots of dedicated chips to do this, but just a few that work at 48 volts.

I can do this from parts that we have in stock:

formatting link

I wonder if that's the best place for C1.

I can put my power LED in series with R2 and save one part.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics
Reply to
John Larkin
Loading thread data ...

INA168 good for 60V in a sot23-5, gain set with a resistor, bandwidth with a cap

Reply to
Lasse Langwadt Christensen

Huh, I only 1/2 get it. You need matched transistors? Bob Pease had a current sense circuit with an opamp that looked nice. (I never tried it.)

George H.

Reply to
George Herold

INA253? 80V Integrated resistor.

Reply to
krw

th a cap

TIs page says preproduction and no stock at distributors, so that might be bit young

Reply to
Lasse Langwadt Christensen

On a sunny day (Sat, 03 Nov 2018 17:00:41 -0700) it happened John Larkin wrote in :

Interesting circuit. I have somewhere one of those HALL sensors... to measure current. Nice little box, expensive, never used it.

formatting link

Payed 4 Euro for 2, in army dump...

:-)

Reply to
<698839253X6D445TD

Temperature tracking is better if you put an NPN current mirror transistor + diode-connected transistor flipped the other direction at the bottom and then put R3 and R2 at the emitter terminals of those instead.

Reply to
bitrex

If you don't mind 0% PSRR.

I used the same circuit in my limiter project,

formatting link
ground referenced of course. It has good bandwidth-current consumption tradeoff but requires a lot of support circuitry to take advantage of that (killing Miller effect), of course.

AD8210 comes to mind. It's pricey but you've never complained about that in your work, that shouldn't be a problem. Good precision, and very good bandwidth (both as signal and PSRR), compared to jellybean current sense amps.

Tim

-- Seven Transistor Labs, LLC Electrical Engineering Consultation and Design Website:

formatting link

Reply to
Tim Williams

My go to part for low amps measurement is the INA270.271. front end is good for 80v and it has an added amplifier with breakouts to add a filter network to get rid of the hash.

Reply to
alan.yeager.2013

With a big shunt drop, hundreds of mV, and modest accuracy requirements, static matching won't be a big problem. At 48 volts, self-heating of Q1 could be. I'd need to keep the signal current low or add a zener or a cascode... more parts.

I suppose I could buy an IC. Adding parts to the PADS library and to the parts database and physical inventory is a mild nuisance. We already have 107 different opamps in stock.

The EMI environment is ghastly (it lights up neon bulbs) so I'd be concerned about an IC. I can't measure temperatures on this board with thermocouples; I have to use an IR imager.

Hey, I'm a circuit designer.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics
Reply to
John Larkin

You should have one or two chips of that type in your set. But one transistor plus an opamp would be a good choice. A sot-23 or sc70 opamp takes no more space than a BJT.

--
 Thanks, 
    - Win
Reply to
Winfield Hill

We have several opamps good for 44 volts, but only one higher, OPA552, and it's big and expensive. Creating an opamp v- rail somewhere below

+48 would be a nuisance.

I might just buy a current monitor chip, but I thought I'd play with circuits first.

How about this? I need the LED anyhow.

formatting link

Of course, it needs some caps.

A quad r-pack would help.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics
Reply to
John Larkin

Yes, of course. That's much better than an HV opamp.

--
 Thanks, 
    - Win
Reply to
Winfield Hill

Cute; I'd like to cascode that PNP transistor, though; if you put three LEDs in series, the '40V' rail could be the negative supply and the base bias for a second PNP. Beta for the top PNP is more stable if the self-heating is kept low, and extra bias voltage could open the op amp options to other jellybeans (even the old LM307).

Reply to
whit3rd

A volt is more than enough bias for a cascode. Anyway, feedback takes out the Early effect (which is big at 48V), and beta is a second order effect. A Sziklai pair would get rid of that effect pretty well with only a bit more than one V_BE drop (~750 mV), which would fit neatly within the V_F of a LED.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC / Hobbs ElectroOptics 
Optics, Electro-optics, Photonics, Analog Electronics 
Briarcliff Manor NY 10510 

http://electrooptical.net 
http://hobbs-eo.com
Reply to
Phil Hobbs

Huh, fun. I didn't know about such low voltage opamps. Can I run it off a single AA battery?

George H.

Reply to
George Herold

It's spec'd down to 1.8, but the data sheet suggests that it still sorta works at 1.5.

It also makes a nice slow RRIO comparator.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics
Reply to
John Larkin

I have found they're useful for all kinds of things, and especially good for building in diagnostic info.

Recommended.

--
 Thanks, 
    - Win
Reply to
Winfield Hill

a long time ago I made this

formatting link
just needs a resistor and a connector on the board

Reply to
Lasse Langwadt Christensen

What part did you use? What part you'd use now? I see the SMA connector, how fast was it? And now?

--
 Thanks, 
    - Win
Reply to
Winfield Hill

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.