Author Topic: DIY DC Current Probe?  (Read 20706 times)

0 Members and 1 Guest are viewing this topic.

Offline Jeroen3Topic starter

  • Super Contributor
  • ***
  • Posts: 4067
  • Country: nl
  • Embedded Engineer
    • jeroen3.nl
DIY DC Current Probe?
« on: September 23, 2017, 04:02:43 pm »
While I was looking into method to get DC current on my oscilliscope I discovered this is an expensive hobby.
You can get DC current probes, for high currents. But what if you want to measure milliamps?
So I thought, why not make one?  :-//

Now, the immediate problem that arises when you're using a shunt is that the scope ground lead will be mains earth referenced.
This would mean that high side, or in-circuit measurements are a problem. I could think of three possible solutions to this.
1. Isolation amplifiers (AMC1100).
2. Isolated Sigma Delta modulators and an analog filters. However, I'm not sure if this would work reliably.
3. High common mode voltage amplifiers. (INA200, up to 80V)

I found one device on Tindie that looked like what I would want. Except not quite suitable for high side.

Simply put, I want to make a uCurrent with a BNC plug on it that can have some common mode voltage.


Would this be a suitable project? Or is it a rabbithole?

-- Update --
Well, rabbithole it is. My efforts so far are on github, and I did some 3D modeling to see if it would fit in an enclosure.
« Last Edit: November 15, 2017, 07:37:37 am by Jeroen3 »
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14076
  • Country: de
Re: DIY DC Current Probe?
« Reply #1 on: September 23, 2017, 04:21:52 pm »
Usually a scope is made for measuring AC. An AC current transformer is relatively easy if you don't need higher accuracy - a scope is usually also not that accurate on the amplitude. The current sensing transformers are usually used for higher current, but could also be used for smaller current, if more primary windings are used (thus however higher inductance).

There are special isolated DC (and LF AC) transducers. On the low end it could be something like this:
https://www.pollin.de/p/differenz-stromsensor-vacuumschmelze-t6040n4646x95081-300-ma-5-v-180071

It is possible to build some kind of battery powered input amplifier (for small currents and optional voltages too). It can be rather simple at lower frequencies and get tricky if it needs to be more accurate / fast. Something like a 10 or 100 kHz BW should be reasonably simple, e.g. with an analog optocoupler or isolation amplifier.
 

Offline ajb

  • Super Contributor
  • ***
  • Posts: 2582
  • Country: us
Re: DIY DC Current Probe?
« Reply #2 on: September 23, 2017, 05:06:14 pm »
Sensitec's CMS3000 sensors would be interesting for this purpose.  They're magnetoresistive, with the sensor isolated from the sensed conductor.  The range of DC-2MHz is pretty impressive for such an inexpensive sensor, and it would require minimal external components.  Unfortunately, last I looked they weren't easy to get in the US, which is a bummer since they look pretty perfect for a couple of my projects.
 

Offline Jeroen3Topic starter

  • Super Contributor
  • ***
  • Posts: 4067
  • Country: nl
  • Embedded Engineer
    • jeroen3.nl
Re: DIY DC Current Probe?
« Reply #3 on: September 23, 2017, 07:22:14 pm »
I am familiar with hall sensor current transformers. They are nice, but also challenging in usage because they are very easily offset or disturbed by nearby fields.
Such as units from LEM. Obviously I could use something from ACS, those little chips, but they have limited resolution due to being digital.

The CMS3000 looks promising though. Basically the hard part of this project is done for me already with that thing :-+.
I'd just have to make a little PGA and power supply from three AAA's. (it will suck them dry fast though)
« Last Edit: September 23, 2017, 07:24:41 pm by Jeroen3 »
 

Offline ocset

  • Super Contributor
  • ***
  • Posts: 1516
  • Country: 00
Re: DIY DC Current Probe?
« Reply #4 on: September 23, 2017, 07:40:06 pm »
What about using an isolated power supply and reference it to a current sense resistor in your hi side.....then connect up to it with a battery scope and measure the voltage across the sense resistor...put it through a IN amp if you want good accuracy.
 

Offline Jeroen3Topic starter

  • Super Contributor
  • ***
  • Posts: 4067
  • Country: nl
  • Embedded Engineer
    • jeroen3.nl
Re: DIY DC Current Probe?
« Reply #5 on: September 23, 2017, 08:15:22 pm »
It wasn't that hard if battery operated scopes were common and cheap.
 

Offline hlavac

  • Frequent Contributor
  • **
  • Posts: 536
  • Country: cz
Re: DIY DC Current Probe?
« Reply #6 on: September 23, 2017, 11:04:23 pm »
I would say shunt and differential probe, but these are expensive hobby as well :( Dave has some in the store

Good enough is the enemy of the best.
 

Offline hlavac

  • Frequent Contributor
  • **
  • Posts: 536
  • Country: cz
Re: DIY DC Current Probe?
« Reply #7 on: September 23, 2017, 11:07:14 pm »
Another option is to use shunt and two channels on the scope, and math function to subtract the two channels. But it is wasteful and noisy
Good enough is the enemy of the best.
 

Offline Jeroen3Topic starter

  • Super Contributor
  • ***
  • Posts: 4067
  • Country: nl
  • Embedded Engineer
    • jeroen3.nl
Re: DIY DC Current Probe?
« Reply #8 on: September 24, 2017, 11:59:19 am »
I've decided to go an build a test board with the CMS3005.
Powering it from batteries will be a challenge, it uses 0.6 Watts!

At work I've made a device like this on a friday afternoon for +-10V signal with a LEM CASR. They are nice to have with selectable primary current multiplier, but you have to use them in a metal box for any accuracy.
 

Offline Neganur

  • Supporter
  • ****
  • Posts: 1138
  • Country: fi
Re: DIY DC Current Probe?
« Reply #9 on: September 24, 2017, 04:58:35 pm »
Personally, I'd try to use the scope's USB port as power source. If I understand correctly, the CMS sensor is already isolating and blowing batteries on this kind of application is quickly getting frustrating.

of course assuming there is a USB port. 0.6W is a challenge.
 

Offline Jeroen3Topic starter

  • Super Contributor
  • ***
  • Posts: 4067
  • Country: nl
  • Embedded Engineer
    • jeroen3.nl
Re: DIY DC Current Probe?
« Reply #10 on: September 24, 2017, 06:03:16 pm »
Yes, I've thought about the USB port. The other method I considered was using 18650 cells.
Both methods do require a boost/inverter +-12V regulator for dual rail. The LT3471 is on the list for later.
 

Offline joeqsmith

  • Super Contributor
  • ***
  • Posts: 11630
  • Country: us
Re: DIY DC Current Probe?
« Reply #11 on: September 24, 2017, 07:45:46 pm »
For DC, I have an old Tektronix current probe along with a few LEM sensors.  The Tektronix is good into the MHz.  I also modified a cheap UNI-T clamp (UT210E) which can do a couple of KHz without changing the front end.   

Depending what you needed, you could do some simple digital modulation across some optos or maybe some magnetics off an old Ethernet board.   

Rare I would use a meter to read current but even the UT61E has the link if you don't need much BW.   

Offline Jeroen3Topic starter

  • Super Contributor
  • ***
  • Posts: 4067
  • Country: nl
  • Embedded Engineer
    • jeroen3.nl
Re: DIY DC Current Probe?
« Reply #12 on: September 26, 2017, 01:09:29 pm »
The CMS3005 is here. So, whats on it?
- AD829A
- 8L08A
- OL S4N
- AL S48
- A7W 44
- Some custom silicon.
On a ceramic pcb.

I get the price now.
 
The following users thanked this post: ajb

Offline Leo Bodnar

  • Frequent Contributor
  • **
  • Posts: 801
  • Country: gb
Re: DIY DC Current Probe?
« Reply #13 on: September 26, 2017, 09:37:40 pm »
While I was looking into method to get DC current on my oscilliscope I discovered this is an expensive hobby.
You can get DC current probes, for high currents. But what if you want to measure milliamps?
So I thought, why not make one?  :-//

Here is completely wild and pointless idea - if your measurement range is reasonably small and you are not looking for lab grade accuracy, you can use [a pair of] VOM1271 either directly into the high impedance scope input or via a buffer amp.

At least it wins by requiring no power at all.

1271 is my favourite solution when I can't be bothered doing a proper driver for high side or across isolation barrier.

Somebody characterised its transfer performance and it is reasonably good: http://www.edn.com/design/power-management/4425510/Isolated-AC-linear-power-control

Cheers
Leo
 
The following users thanked this post: chickenHeadKnob, PointyOintment

Offline Jeroen3Topic starter

  • Super Contributor
  • ***
  • Posts: 4067
  • Country: nl
  • Embedded Engineer
    • jeroen3.nl
Re: DIY DC Current Probe?
« Reply #14 on: September 28, 2017, 07:34:52 pm »
A linear optocoupler could be useful to add to the parts list. Yet, I would still need additional circuitry, such as isolated power and an shunt amplifier for that way.
How would such linearity hold when exposed to non-room-temperature situations?

I've played around with the CMS3005 today. Cool thing, it's spot on, and it detects 1 mA without issues. (The noise on my breadboard made it somewhat inaccurate here)
It also gets remarkably hot when doing nothing, so that might be one of the reasons why ceramic was chosen.

When probing on it, I did not find any voltage regulators. So those sot-223 are doing something else.
It isn't greatly affected by unbalanced supply (eg: 12 and -15), only a few mV. So it does contain a reference. No super precise supply required, but it should be a stable one. No switching noise, since running it from an RP-1215D directly didn't work very well.
 

Offline Brutte

  • Frequent Contributor
  • **
  • Posts: 614
Re: DIY DC Current Probe?
« Reply #15 on: September 28, 2017, 08:53:05 pm »
Now, the immediate problem that arises when you're using a shunt is that the scope ground lead will be mains earth referenced.
You did not mention the bandwidth and the voltage bareer needed I think.
If that is for in-circuit measurement of something that is mains powered then I would rely on some isolation.
OTOH if that shunt is just floating around common mode +-1V (earth lead voltage drop) then I think you could use a raw op-amp with 5-resistor arrangement.

In this post I mention one with +-300V common voltage /100 and 1V diff voltage so a bit of overkill in your case.
I suggest you could start from +-10V common voltage /5 and 10mV diff shunt voltage amplified to 100mV for scope. For 1MHz bandwidth you need a 50MHz GBW opamp then. And some decent resistors.





 

Offline ajb

  • Super Contributor
  • ***
  • Posts: 2582
  • Country: us
Re: DIY DC Current Probe?
« Reply #16 on: September 28, 2017, 10:51:20 pm »
Interesting!  Looks like some of those resistors were laser trimmed?  What's the back of the sensor look like?

They actually sell an SOIC version of their sensor that is intended to sit over a PCB track that carries the sensed current, but it's only spec'ed for DC to 500kHz.  I'm not sure if the lack of a well-controlled substrate is part of that limitation, maybe they're just unable to package sufficient analog circuitry.

https://www.sensitec.com/products-solutions/current-measurement/cfs1000
 

Offline Jeroen3Topic starter

  • Super Contributor
  • ***
  • Posts: 4067
  • Country: nl
  • Embedded Engineer
    • jeroen3.nl
Re: DIY DC Current Probe?
« Reply #17 on: September 29, 2017, 07:23:22 am »
I think you really need die better parameters of the ceramic for this to work. The laser trimmed resistors are probably the only way to get the precision they advertise. Because I see no programmable parts.

Now, does anyone know of a -12V and +12V SMD LDO pair?
 

Offline Assafl

  • Frequent Contributor
  • **
  • Posts: 600
Re: DIY DC Current Probe?
« Reply #18 on: September 29, 2017, 07:46:55 am »
Perhaps a belated reply but I needed a similar sort of setup for detecting motor current. Ended up using an Allegro ACS712 (which have been discontinued and now are called ACS723).

Similar thing albeit cheaper. Also seems to be more finicky (but doesn't need the dual polarity power supply).


Why not use a 12V LDO and a voltage inverter like http://www.ti.com/product/lt1054/datasheet ?

I like the TPS60400DBVR voltage inverter which is in an SOT23 package but it is only up to 5V...
 
 

Offline capt bullshot

  • Super Contributor
  • ***
  • Posts: 3033
  • Country: de
    • Mostly useless stuff, but nice to have: wunderkis.de
Re: DIY DC Current Probe?
« Reply #19 on: September 29, 2017, 07:47:47 am »
The CMS3005 is here. So, whats on it?

Looks quite the same as their predecessors (CMS20xx), except for the faster OP-Amp.
I should ask them if they use a different sensor chip (that's where their knowledge is).
The CMS2000 series is somewhat comparable to the LEM closed loop sensors, some aspects were better with the CMS, others with the LEM LA-100 (e.g.)
In the early days of the CMS2000, they were quite useful as thermometers :-)

Safety devices hinder evolution
 

Offline capt bullshot

  • Super Contributor
  • ***
  • Posts: 3033
  • Country: de
    • Mostly useless stuff, but nice to have: wunderkis.de
Re: DIY DC Current Probe?
« Reply #20 on: September 29, 2017, 07:58:00 am »
When probing on it, I did not find any voltage regulators. So those sot-223 are doing something else.
It isn't greatly affected by unbalanced supply (eg: 12 and -15), only a few mV. So it does contain a reference. No super precise supply required, but it should be a stable one. No switching noise, since running it from an RP-1215D directly didn't work very well.

The other SO-8 chip is the voltage regulator (an 78L08 as far as I remember). The SOT-223 are output current booster transistors for the OP-AMP (your standard NPN / PNP emitter follower pair). The absolute value of the supply voltage isn't important, since this thing operates closed loop - the magnetic field of the conductor (on bottom side) is compensated by a current loop on the sensor chip, the GMR sensor itself always sees "zero" (just like the Hall elements in the LEM sensors). So basically, the accuracy of these sensors depends on mechanical precision. I don't know if there's trimming required or possible. Maybe zero (offset) trimming is necessary.
The small blocks left and right to the sensor chip are permanent magnets to bias the GMR sensor. If you remove or demagnetize them, the sensor won't work anymore.

Safety devices hinder evolution
 

Offline 0xdeadbeef

  • Super Contributor
  • ***
  • Posts: 1570
  • Country: de
Re: DIY DC Current Probe?
« Reply #21 on: September 29, 2017, 08:47:10 am »
I worked on a simple current measurement box with the CMS3025 lately.
It's supplied with a ~15V single supply and uses a LT1054 to create +/-13.4V @100mA (according to LTSpice) and two LDOs (79M12/78M12) to create the +/-12V.
I also added a measurement rectifier and a voltage ladder with comparators to get some LED display of the current. Still untested as I'm waiting for some parts.
The PCB is supposed to be held by the connectors which are fixed in an acrylic front plate which again is screwed to a 3D printed case.
Unfortunately, the CMS3025 is slightly higher than the BNC- and Banana Plug connectors, so I'll have to mount it hanging upside down on the lower side of the PCB.
Posted something before here regarding design decisions etc.

[EDIT]
Added new screenshot of the two assembled (well, SMD parts only) boards I built today.
« Last Edit: September 29, 2017, 06:59:25 pm by 0xdeadbeef »
Trying is the first step towards failure - Homer J. Simpson
 
The following users thanked this post: Jeroen3

Offline Jeroen3Topic starter

  • Super Contributor
  • ***
  • Posts: 4067
  • Country: nl
  • Embedded Engineer
    • jeroen3.nl
Re: DIY DC Current Probe?
« Reply #22 on: September 30, 2017, 08:59:15 pm »
Looks great. But why don't you flip the parts to the other side and use a uCurrent type box?
 

Offline 0xdeadbeef

  • Super Contributor
  • ***
  • Posts: 1570
  • Country: de
Re: DIY DC Current Probe?
« Reply #23 on: September 30, 2017, 11:07:33 pm »
Well, I considered several mechanical solutions but this was the best I could come up with in the limited time I was willing to spend. And well, I'm not a mechanical engineer ;)
Compared to the µCurrent, the main problem is the BNC connector as it can't just sit on top of the PCB (if it acts as front panel).
This is supposed to be used in a development environment where it it has to be mechanical robust to some degree and doesn't allow to touch parts which could have high voltages (like >= 50V).
So the banana plugs need to be fully isolated and specified for high currents (the beefier ones for the current are specified for 32A).
Plus my goal was to use affordable parts which could be soldered directly to the PCB (to avoid additional wiring and the according inductance and resistance).
This is the high current safety banana plug for the current jacks:
https://www.voelkner.de/products/89893/Seb-2620-F6-3-Ge-Einpressbuchse.html

Actually, I would really like my design if there wasn't the problem that the CMS sensor is slightly too high and has to be mounted on the bottom.
The main issue isn't that it looks a bit weird but that this increases the necessary case size by another 3cm or so.
The only other issue with this design is that once the four banana jacks are soldered in, it's impossible to access the PCB's top side without desoldering them.

BTW: I appended a front panel mock-up using the cardboard prototype I used to test the matching.
I also tested the SMD part today. The symmetric voltages look OK (albeit not very exact without load like 1-2% or so) and the LED stuff (measurement rectifier + comparators) works as expected.
« Last Edit: September 30, 2017, 11:12:37 pm by 0xdeadbeef »
Trying is the first step towards failure - Homer J. Simpson
 
The following users thanked this post: lem_ix

Offline Jeroen3Topic starter

  • Super Contributor
  • ***
  • Posts: 4067
  • Country: nl
  • Embedded Engineer
    • jeroen3.nl
Re: DIY DC Current Probe?
« Reply #24 on: October 01, 2017, 04:49:38 pm »
I think it's going to topple over when you attach a BNC. It'll be a pita.

But, today I think made a power supply circuit with a TPS6513x.
Now on to sourcing all the components, which isn't easy if you are not a business. Then create test board, since I never get these right first try.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf