EEVblog Electronics Community Forum
Electronics => Open Source Hardware => Topic started by: jram112 on July 22, 2013, 05:57:07 pm
-
So I love my uCurrent ... thank's Dave. I've been searching for uCurrent mods from users, and thought that we could start a topic to collect any mods that other users are willing to share to the community. So please mention any mods that you have made in hopes that someone will find them useful.
Now, I also have a request on how best to mod the uCurrent for high-side current measurement. I find myself needing to do high-side current measurements often. My current project will also require me to do a high-side current measurements. Any suggestions on how I could mod the uCurrent?
-
uCurrent...... mods? :o It's a pretty simple device...
I find myself needing to do high-side current measurements often. ... Any suggestions on how I could mod the uCurrent?
I don't have one - but unless I'm sorely mistaken on how it works, the answer is "connect it high-side".
-
I don't have one - but unless I'm sorely mistaken on how it works, the answer is "connect it high-side".
Yes, unless you use a grounded meter or scope on the output.
-
Maybe I'm just confused about virtual grounds, but the uCurrent schematic shows that the -In input terminal is connected to the VGND. VGND is mid-supply for the uCurrent or 1.5V. Wouldn't the 1.5V potential on the terminal cause trouble?
Specifically I want to measure the current on the mirror side of a n-channel cascoded current mirror. See attached image.
-
It's only a 1.5V potential relative to what the uCurrent calls "0V". Remember that potential is relative. The uCurrent is 100% floating.
-
OK...makes sense. What about using an oscilloscope probe connected to the uCurrent output. I assume that most scope probes are grounded?
If that is the case then would I need to mod the uCurrent for a dual supply setup (+/- supply)?
-
Yes, the vast majority of oscilloscopes are grounded. But there's nothing else in your circuit that needs to be grounded, so just make sure it's not. Build the entire circuit so that nothing, including the power supply and the uCurrent, is grounded. Then you can stick your scope probe wherever you want, as long as you keep in mind that once you do, the circuit is ground-referenced, so you can't go sticking channel 2's ground somewhere else.
Most DC test equipment is not ground referenced, so you can just make sure everything floats, and you can stick your scope anywhere. You have to pay attention when you start attaching function generators, spectrum analyzers and such, because AC equipment is often grounded. (A lot of function generators are not, but many - especially cheaper - ones are.)
-
Thanks. I learned something new.
However, I will more than likely have to use more than one probe and a function generator. Thus, my floating circuit ground will no longer be floating.
I've played with the INA282-INA286EVM from TI, but the output from this doesn't match the uCurrent nor my multimeter. I trust the uCurrent more. They are described as a voltage output, high-side measurement, bi-directional, zero-drift current shunt monitors.
-
You can always measure differentially by connecting Ch 1 to the high side of something and Ch 2 to the low side, then setting the scope to add/invert, but unless you have a four-channel scope, that doesn't leave much, does it?
Of course, you can use two probes, I didn't mean to imply you can't, you just have to connect their ground clips to the same potential.
Check whether your function generator is grounded or not. Many aren't.
Those INA current monitors aren't bad, but they're hardly instrumentation-grade. uCurrent is.
-
I've done the two-probe differential method, but as you mentioned...I don't have any probes left (2ch scope). Any other suggestions?
-
Short of getting lucky and finding a working differential probe cheap on fleabay, I'd say try to figure out a way to test your circuit without having to probe multiple signals at once. If your circuit is being stimulated by a function generator, one thing you can try is to connect the external trigger input to the generator and sync to that. Then, when you move from signal to signal, they will line up on the graticule phase-accurate to how it would look if you probed them simultaneously, and you can measure the time difference between signals that way. It's a bit of a pain - I have to say, I really love my 4ch scope...
-
Thanks for all the useful tips c4757p, but I think I would prefer to have a permanent high-side measurement solution that I could grab off the shelf. I've been unsuccessful at finding any "cheap" differential probes. Someone needs to make an opensource probe design.
This is a post about uCurrent mods; thus, I'll take us back to the original question about modding the uCurrent for high-side current measurement. Anyone have solutions?
-
Yes, unless you use a grounded meter or scope on the output.
Out of interest what is the AC bandwidth of the uCurrent?
-
Yes, unless you use a grounded meter or scope on the output.
Out of interest what is the AC bandwidth of the uCurrent?
There is a lot of info on bandwidth here:
http://alternatezone.com/electronics/ucurrent/uCurrentArticle.pdf (http://alternatezone.com/electronics/ucurrent/uCurrentArticle.pdf)
Basically, it is 2kHz. I think you could get much faster by replacing the op-amp with something better, but I haven't checked if the footprints are standard (they usually are for dual and quad op-amps, I have no idea about single chopper op-amps).
By the way, I am interested in building my own low-spec differential probe designed to be used in low voltage digital circuits (gets everything easier because we are not dealing with 1200V isolation or something like that). It's sole function would be measuring current in these systems with a scope. 100Mhz bandwidth would be nice, even if it needs compensation to be done at that bandwidth. Remember it's all low voltage, you could just use some shunts like Dave does in the uCurrent and connect that directly to a fast differential amp with some gain. Since we are dealing with scopes, 1% accuracy is not a deal breaker, and things like offset voltage can be compensated the same way you compensate passive probes.
I have an Elektor article on something similar, but the bandwidth is not quite there (the isolation is high, though) and it's missing essential things to recreate the project like gerber files and the model number of the box it was put in. If anyone knows about something like this, please tell us all. :-+
-
Basically, it is 2kHz. I think you could get much faster by replacing the op-amp with something better, but I haven't checked if the footprints are standard (they usually are for dual and quad op-amps, I have no idea about single chopper op-amps).
Just a small point before anyone gets the wrong idea. I think you mean faster. I haven't seen anything lately that'll give you better DC accuracy than the op-amp that's in it. You'd be trading off bandwidth for other specs which would almost certainly lower it's DC performance.
-
Just a small point before anyone gets the wrong idea. I think you mean faster. I haven't seen anything lately that'll give you better DC accuracy than the op-amp that's in it. You'd be trading off bandwidth for other specs which would almost certainly lower it's DC performance.
Yes. The chip in the new uCurrent is technically trading off better AC performance (30KHz + no visible chop noise), for some DC offset performance.
-
That is exactly what I meant. I kinda like the idea of having a box containing a differential amp, current sense resistor and grounded output for current measurement.
-
By the way, I am interested in building my own low-spec differential probe designed to be used in low voltage digital circuits (gets everything easier because we are not dealing with 1200V isolation or something like that). It's sole function would be measuring current in these systems with a scope. 100Mhz bandwidth would be nice, even if it needs compensation to be done at that bandwidth. Remember it's all low voltage, you could just use some shunts like Dave does in the uCurrent and connect that directly to a fast differential amp with some gain. Since we are dealing with scopes, 1% accuracy is not a deal breaker, and things like offset voltage can be compensated the same way you compensate passive probes.
How resistive would that shunt + differential amp be at 100 MHz? Especially for the higher value shunt resistors (lower current ranges) the input capacitance of the amp may become an issue. For low values trace/shunt inductance may become an issue.
-
How resistive would that shunt + differential amp be at 100 MHz? Especially for the higher value shunt resistors (lower current ranges) the input capacitance of the amp may become an issue. For low values trace/shunt inductance may become an issue.
That's probably why it hasn't been done ::)
Maybe I could just get this one to work at 20MHz and then have another probe designed to work on the higher frequencies as a separate device, using a hall effect sensor formed by a toroid coil (that one I would definitely not build).
-
Agilent has a current probe that works like that (http://www.home.agilent.com/en/pd-2229629-pn-N2820A/3-mhz-50ua-high-sensitivity-ac-dc-current-probe-2-ch). Look at the bandwidth figures.
For higher frequency AC only you can use the a current transformer. Achieving 20 MHz bandwidth is probably not easy, but definitely (http://www.tek.com/datasheet/current-probe-p6021a-ac-current-probes-datasheet) possible (http://www.tek.com/datasheet/current-probe-ac-current-probes).
-
Agilent has a current probe that works like that (http://www.home.agilent.com/en/pd-2229629-pn-N2820A/3-mhz-50ua-high-sensitivity-ac-dc-current-probe-2-ch). Look at the bandwidth figures.
For higher frequency AC only you can use the a current transformer. Achieving 20 MHz bandwidth is probably not easy, but definitely (http://www.tek.com/datasheet/current-probe-p6021a-ac-current-probes-datasheet) possible (http://www.tek.com/datasheet/current-probe-ac-current-probes).
Seems like I haven't realised measuring current at high frequencies is hard :palm:
$4000 bucks for that Agilent probe makes the 750kHz differential probe I found on Elektor look like a real bargain! ;)
To bring this back on topic, the uCurrent could be modded with a faster op amp, as I said, then it could have a differential amplifier module after that (in the same style of box) that allows up to ±30V differential voltage. Sounds like a good mod and the diff. amp. box can be reused as a tool all by itself.
-
I'm hoping to create a mod for the uCurrent for low-voltage differential current measurements.
In the mean time here's a mod I found for the uCurrent's enclosure. It's a 3d printed low profile enclosure with easy battery access. http://www.thingiverse.com/thing:26878 (http://www.thingiverse.com/thing:26878)
-
Agilent has a current probe that works like that (http://www.home.agilent.com/en/pd-2229629-pn-N2820A/3-mhz-50ua-high-sensitivity-ac-dc-current-probe-2-ch). Look at the bandwidth figures.
The make before break test connector they describe, is that just two pinheaders with a jumper? (With the jumper removed after the test head is connected.)
-
Does anyone know of a drop-in replacement for the MAX4239 that would give higher bandwidth? I've been unsuccessful at finding one with as good DC characteristics.
-
Does anyone know of a drop-in replacement for the MAX4239 that would give higher bandwidth? I've been unsuccessful at finding one with as good DC characteristics.
You won't find anything that is both high bandwidth and very high accuracy. Maybe you can install both op-amps and switch them. SO you have a high bandwidth bode and a high accuracy mode. That's a cool mod :-+. Just be careful so that anyway you do the switching doesn't limit the bandwidth or adds noise. You could try making a separate board with the switching circuitry and solder some wires where the original op amp used to be.
-
Could you not add some voltage input to the negative feedback with a POT and trim out the DC offset in a higher BW opamp or does the DC offset change?
It would be very helpful for me to have something that is close but does not need to be perfect. If I could be within 1 mA on a 10mA scale that would be great but I would really like to be able to see transients on an oscilloscope. I have an HP current clamp which can do I think 1 mA per division but it is sooo huge and the clamp always gets in the way. I know on my HP current clamp I have to trim it to be DC accurate before I apply power. Can we not do the same with this?
Not sure if I am being WAY over simplistic with this.
-
The mod's seem nice. I will also try some of them out for myselfe. Good luck!