uCurrent Gold mods

[Robaroni]

Hi,
I just received my ucurrent Gold. Dave did a nice job on it but I'm sure some of you had an issue with the 2032, anyway I did, I popped a new 2032 in and checked the current, it's 3.8 mA but it looks like the LED is taking a good percentage of that current. After not too much time, maybe an hour, the 2032 started to drop and the LED started to blink. I figured at 4 mA's a 2032 rated at 240mAh's should have lasted longer but either my 2032 was weak or the circuit under load upped the current draw considerably - the 2032 didn't last very long. I decided to save my relatively expensive 2032's for my welding helmet so I ran in a couple rechargeable AA's 'cause I'm one of those people who walks away from the bench and forgets to shut things off which I'm sure no one else on the planet ever does!
I don't plan on mounting the Gold in a box so I just ran standoffs under the PCB put another PC with a battery holder on it under the standoffs then mounted more standoffs to keep the batts from hitting the bench. Worked fine and sits about 30mm off the bench.

Anyway, Dave, it's a neat little unit and I found an anomaly in my super duper Agilent (or whatever they call themselves today!) 34410a, 6-1/2 digit meter on the 1mA scale when I was checking out the Gold.Hard to beat for the money and worth not having to dicker finding parts and building my own schematic.

Rob

[uski]

Hi

In mine I've put 3 AAA alkaline cells (yes, 3) so it gives me a slightly higher dynamic range.
I've also changed the resistor of the LED to reduce the current consumption.

uski

[con-f-use]

Did you see MLXXXp's uCurrent Gold Mod? Its rather nice as well.

• Change on-short to off-short [no need for it to be powered]
• Lower the LED brightness [less annoying plus less battery consumption]
• Add some input protection when of [diodes across the input when powered off]
• Increase the positive measurement range [virtual ground shifted because negative or ac not measured that often]
• Lower the case [too high for coin cell usage]

I like the idea to shield the case but I'd do it from the inside (looks less botched) and use aluminium tape, because copper looks nasty when it corrodes. Also I'd add reverse polarity protection either by low rds_on mosfet or a Schotky diode.

[LA7SJA]

After changing the LED for a High Bright Red and changed the resistor R4 I managed to reduce the Power consumption to ~2mA With more light. I then removed the 2032 holder and mounted a 3 x AA battery holder in the box. With 3 x Lithium AA cells it works for about 1500h (3000mAh / 2mA=1500h = 187, work days).

Johan-Fredrik

[Robaroni]

Good ideas!
I was thinking of raising the LED resistors value too but I didn't think of higher batt voltage. I'll look at the chips and see what VCC max is.

Rob

[Robaroni]

The reason I stuck with two rechargeable AA's is because the indicator chip runs at about 2.6 volts or so. You don't want to run your rechargeable batts to low or they won't recover. Using three batts would bring the 1.1 volt per batt to 3.3 volts which is above the low batt indicator volts. I could change out the indicator chip but then it starts to get crazy. I just don't want to go that far.

Another note is this thing is more accurate then my $1,200 Agilent 34410a current section on low current! Nice!

Rob

[LA7SJA]

I'm trying to find a low capacitance diode and put two of them in opposite directions over the input to limit the burden voltage (and even protect the circuit). It should idealy withstand 1-2 amps and have a maximum capacitance of 2,5pF (106KOhm @ 300KHz) 1N4148 has a low enough capacitance but can not withstand that much current. So if anyone has a fitting diode I would be very happy. The problem I have is that when you messure the sleep current in a avr/pic and the device vake-up and the current goes from 100nA to 1mA the cpu brown out from the burden voltage.

R1 0.01 Ohm 1,000W  Destruktiv current  I=SQR(P/R)=10A        (Burden voltage 0.10V)
R9 10 Ohm  0,125W    Destruktiv current  I=SQR(P/R)=112mA   (Burden voltage 1.12V)
R2 10 KOhm 0,100W   Destruktiv current  I=SQR(P/R)=3.16mA  (Burden voltage 36.1V)

Johan-Fredrik

[TomWS]

Did you see MLXXXp's uCurrent Gold Mod? Its rather nice as well.
<snip...>

Thanks for the lead.  I looked at the referenced post and was surprised by the comment that the Common Mode input voltage for the LMV321 was too low to shift the Virtual Ground (something I want to do).  Looking at the datasheet, I then realized that MLXXXp didn't increase the power supply voltage so his comment made sense in that context.  However, to me, shifting the virtual ground significantly increases the range if you only measure current from a unipolar source.

My need is to use uCurrent Gold for embedded controllers that typically draw 10-20mA when operating and single uA values when sleeping.  To achieve this, I think what I need to do is:

• Increase the power supply to ~5V to increase the measurement range.
• Change the shunt resistor on the mA setting to 1.0Ohm to change the range to 0.1V/mA.
• Shift the Virtual Ground so that V+=~4V, V-=~-1V

#1 can be accomplished very nicely with a 3xAAA Lithium battery holder.  This will give me nominally 5.1V.

#2 Will give me 20mA full scale with 1uA resolution with a 20000 count DMM.

#3 Shifts the upper range to be close to 40mA, which is perfect to handle the surges on most of my embedded controllers.   The Virtual Ground can be shifted easily with a combination of 120K/30K combination for R6/R7 respectively.

Anyone see any pitfalls with these changes?

Tom

[MLXXXp]

Anyone see any pitfalls with these changes?

You may wish to increase the value of the BATT OK LED dropping resistor R4, to decrease the brightness that will result from operating at a higher voltage, but mainly to reduce the risk of damaging the TPS3809 voltage monitor IC, by having it drive too much current through the LED.

Also note that the low voltage monitor will still keep the LED lit down to around 2.6V, so you may need some other means to determine if your batteries are providing a high enough voltage to give you the full scale readings you require without the op amps clipping.

[TomWS]

You may wish to increase the value of the BATT OK LED dropping resistor R4, to decrease the brightness that will result from operating at a higher voltage, but mainly to reduce the risk of damaging the TPS3809 voltage monitor IC, by having it drive too much current through the LED.

Also note that the low voltage monitor will still keep the LED lit down to around 2.6V, so you may need some other means to determine if your batteries are providing a high enough voltage to give you the full scale readings you require without the op amps clipping.

Thanks.  I realized that after I posted - to me the LED is sort of superfluous anyway so I might just remove it (and the TPS3809) altogether.  Since I have a meter attached to the device, it seems pretty trivial to add a couple of test points to check the battery...

Tom

[manbert]

Hi all!
I had a look at the µCurrent schematic and the MAX4239 datasheet, because i was thinking of how to adjust the positive range (i probably don't need the negative one).
In the MAX4239 datasheet you'll find a "common-mode input voltage range":  VGND-0.1V  to  VCC-1.3V
If you set the virtual ground to VBAT/2, you'll have +-1.5V rail2rail (best case with full battery).
So, is there a positive margin of only 0.2V left? Then it would not work very well!?
Would it not be better to set the virtual ground to 0.8V? (to have similar pos. and neg. ranges)
Did i misunderstood something in the way this OP works?
Can't find any statement of this little beauty having a rail2rail input...
 
Regards,
Michael

[MLXXXp]

Would it not be better to set the virtual ground to 0.8V? (to have similar pos. and neg. ranges)

Did you see modification number 4 of the ones that I made?

[manbert]

Did you see modification number 4 of the ones that I made?

Yes, i did. You said:

...
Position 1 leaves the circuit unaltered, with the original range of +/-1250mV
...

The idea of shifting the range just led me to the concern, if the original range is +/-1250mV at all !?
With a "common-mode input voltage range" of VGND-0.1V  to  VCC-1.3V, i would expect an original range of -1250mV to about +200mV.
But with hundreds of devices sold around the world and being used and tested, i guess it works all right...
I just don't get why. Maybe i'm missing some knowledge regarding op amps...
I would gladly find out the reason, why the voltage can go up to +2.75V (virt.GND 1500mV + 1250mV), although the limit (by common-mode input range) is about 1.70V ? (VCC[~3V] - 1.3V)

Regards,
Michael

[manbert]

Think, i got it:
At the input of the first op amp we have just the burden voltage.
At the second input, there is max 1250mV / 10 = 125mV.
The margin is 200mV, so it should work...