current sensors

[Simon]

Hi guys,

I'm using the MAX4172 current sensing chip, but this has a max temperature of "only" 85C, does anyone know of a similar chip that does 125 or 150 C ?

[NiHaoMike]

http://www.maxim-ic.com/datasheet/index.mvp/id/3597

[Simon]

very nice, thank you, I'll have to give that a read, I'm trying to figure out if by SO case they mean SOIC ? my guess and hope is that it is the same pin-out as the MAX4172 as it is a very similar looking device but with more pins used, I hate to think what it costs, the MAX4172 is bad enough at up to £3 (2 if bought in bulk)

[Simon]

well I'll be damned, it's 1/2 the price of the 4172, I'll better get reading that data sheet

[RayJones]

Yeah, the max4080 is what I'm going to try soon myself.

The main reason I looked further than the MAX4172 is it would appear by Farnell and RS's pages it is becoming un-obtainium in Oz....

And as you say, the 4080 is cheaper 

[Simon]

yea at the time i cleared farnell's stock of just 12

[Simon]

bugger they have completely changed the pins around what a bunch of wollies same amount of pins and functions and they could have made a drop in replacement, now how stupid is that for design

[Simon]

right, a question: as I'm now trying to "drop" this same size chip but sillily not pin compatible into my project, can I connect a pin to a NC pin ? basically I could make connection by routing my trace straight through an NC pin, I presume that as it is not connected this is not a problem ?

[RayJones]

If the pin is truly NC, and you can get either device to fit the footprint then yeah, why not.

But I would advise a note on the schematic to explain why this was done to avoid confusion later.

They sometimes physically drop the centre pin on TO-220's for instance, but you are meant to use the tab, so it's not quite the same.
Dropping unused pins on a SMD device would probably weaken the mechanical aspect, and simply look odd, so they keep the unused pins on the package. (at least that sounds good to me). 

[Simon]

yea thats what I figured so i did it, at worst I'll have to rip up that one pin from the chip

[RayJones]

Hmmm, out of interest I checked out the pinouts of the MAX4172 vs MAX4080.

I think you are "shit out of luck" for a simple solution.

You would probably get away with the nominal swapping of Vcc / RS- BUT, the OUT pin of the MAX480 pin moves to what was the GND pin on the MAX4172.

Close, but no cigar.

However, a possible way out though would be a 3 pin link/jumper next to pins 5/6:

Pin4 SO8 = GND (NC on MAX4172)
pin5 SO8 = centre pin of 3 pin link
pin6 SO8 = top pin of 3 pin link - and this is also your output you can use
lower pin of 3 pin link = GND

To use a MAX4080, install link in position top-centre
To use a MAX4172, install ink in position bottom-centre

If you dont want links, you could bridge solder over a similarly closely laid out pads on the board (not the SO8 pins)

Horrid ASCII art version:

Code: [Select]

  Rs+     (1)     (8) Vcc/Rs-
   Vcc/Rs- (2)     (7) NC
       NC  (3)     (6) --------- LK.3 ----- to circuitry
      GND  (4)     (5) ----------LK.2
                                 LK.1 --- GND


LK3.2-LK3.1 = MAX4172
LK3.2-LK3.3 = MAX4080

                    

[Simon]

oh it's ok I rerouted the board, I've not made anything physical yet but having laid the whole thing out already and had the 4172's anyway it would have been nice to seamlessly upgrade to the 4080's when I ran out of 4172's. Due to Maxims dick-head way of doing it I'll have to have one or the other from the start or make a few of one and then the other but why bother - anyone want some MAX4172's I'll do you a good price 

[Simon]

looks like the MAX4080SASA (60 gain) is real hard to come by, no European stock, only TASA versions (20 gain) I seem to remember dave saying that maxim stuff was hard to come by, well I can cope on the TASA to start but if I'm going to have to go with that might as well use up the MAX4172's i have

[alm]

Aren't there any semiconductor manufacturers that can actually deliver silicon (i.e. most except Maxim and Atmel) that make these? Not sure about the scale of this project, but have a critical single-sourced part that's not available might force you into another redesign. I would try the Digikey/Mouser parametric searches to find other parts.

[Zad]

Design your own with an op-amp and 4 * 0.1% resistors?

I'm afraid Maxim aren't generating a very good reputation with me for the longevity of their components. A friend of mine has experienced them End Of Lifing components without much notice several times

[Simon]

Aren't there any semiconductor manufacturers that can actually deliver silicon (i.e. most except Maxim and Atmel) that make these? Not sure about the scale of this project, but have a critical single-sourced part that's not available might force you into another redesign. I would try the Digikey/Mouser parametric searches to find other parts.

well if this project ever goes anywhere I don't know how many chips I may need, possibly 200 maybe 1000, not the sort of thing I want to pre order to guarantee my supply, it should not have to work like that.

[Simon]

Design your own with an op-amp and 4 * 0.1% resistors?

I'm afraid Maxim aren't generating a very good reputation with me for the longevity of their components. A friend of mine has experienced them End Of Lifing components without much notice several times

yea I tried that in the first place with a differential op amp configuration, the problem is that I'm measuring the supply voltage, I'm sure I could find way's around it but they would be very inconvenient and possibly more costly than the maxim all in one chip. I'm also stuck for space, I have room for the maxim part and that's it. Maybe I should look elsewhere, I don't want to be in a position where I have to keep changing the layout to suit currently available parts, it will making support for the project impossible.

[RayJones]

LT6100 can give you x50, it all depends on how you strap the inbuilt resistors.
Only gotcha is the Vcc to the device *must* be below the rail you are sensing, unlike the Maxim devices.

These work fine, and I've used them in the past with great success, but thought I'd see how the Maxim devices go. (bigger for a start = easier to deal with)

[Simon]

are you saying that the measured voltages must be above the supply to the chip ? NO PROBLEM thats what I have

[RayJones]

Yes, it is actually quite well hidden amongst the parameters, certainly not up on the front page.

I simply used the 5V for my PIC to run the LT6100, sensing a nominal 12V level and it works fine.

The other thing I took care with was to run the GND of the HS sensor directly back to the PIC's GND pin, ala Kelvin fashion. ie none of the large currents should travel this path otherwise the ADC readings become distorted by ground shifts - which was the reason why I shifted to the high side in the first place.

[Simon]

well yea that would not be a problem for me as I'm running the chip off 5 V (more for it's own protection behind the voltage reg and other circuitry) and I'm sensing the current coming into the care battery. That chip is actually not quite as good a spec as the MAX4080 which has 100 uV offset, the LT one has 300 uV as I'm pulling the most from these chips by using the smallest Rsense i can because of large currents I want the least input error as it will impact more greatly.

[RayJones]

That's a bit harsh, comparing the max spec for the LT6100 against the typical spec of the max4080.

Using that basis I could claim the LT6100 has an input offset of 80uV vs the MAX4080's 600uV.

In the context of typical ratings and max ratings the LT6100 actually performs better on both counts +-80uV vs +-100uv, and +-300uV vs +-600uV.
If you go over the full temperature range, the same applies.

[Simon]

oh right well I've not fully read the datasheets yet, I've resolved to use up my MAX4172's while waiting for the MAX4080's to become available or in the mean time choose another chip.