EEVblog Electronics Community Forum
Products => Test Equipment => Topic started by: necessaryevil on February 10, 2015, 09:31:02 am
-
Hi folks, I have some n00b questions about multimeters and the µcurrent.
1. Is the current input o multimeter isolated from the circuit of the multimeter? I know there won't be high voltages across a shunt resistor, but I can imagine that it is desirable to have some sort of isolation between the input and the rest of the circuit.
2. Can I use the µA/mA ranges of a multimeter to measure currents on grid powered circuits?
3. Can I use the µcurrent for AC?
4. Can I use the µcurrent in grid powered/'high voltage' circuits? If not, is it a simple safety issue or wil it blow right away?
5. What (replacement) should I use in a multimeter? Should I only buy them from the manufacturer of the multimeter, or can I use other fuses (with proper ratings of course)? Besides speed and current, what other ratings and approvals should I look for?
When I say multimeter, I don't think of a specific brand or type, but of a multimeter in general. Let's assume a decent quality multimeter; intended for measurements for domestic grid measurements (CAT II, or are there any other approvals?)
-
1. no
2. depends on your cat rating of the multimeter. Every "decent" model should be able to do that
3. yes
4. hmm not sure about that. I would say no, looking at the design of the device
5. Use the same types as in the meter. Do not try to "improve" the meter by yourself.
-
Thank you. About answer 4, now that you say it, the switches have a metal casing and are nowhere near big enough to have enough clearance/creepage space.
About 5.
Well, I wasn't thinking of a better fuse, but I was trying to find out what fuses should be used. I'm not a cheapskate when it's about safety, but digikey charges €38 for a Littlefuse FLU series fuse (intended for multimeters). At the same time, a Littlefuse 508 series fuse costs about €3-8 euros and has similar ratings. Very tempting isn't it?
-
Well, I wasn't thinking of a better fuse, but I was trying to find out what fuses should be used. I'm not a cheapskate when it's about safety, but digikey charges €38 for a Littlefuse FLU series fuse (intended for multimeters). At the same time, a Littlefuse 508 series fuse costs about €3-8 euros and has similar ratings. Very tempting isn't it?
Murphy's law: a 3 cent fuse will blow a 300 dollar tube :-DD
What I am trying to say: fuses are so slow, when the circuity around it is not 100% perfect, the fuse will blow after de rest has gone dead.
-
Murphy's law: a 3 cent fuse will blow a 300 dollar tube :-DD
What I am trying to say: fuses are so slow, when the circuity around it is not 100% perfect, the fuse will blow after de rest has gone dead.
Thats a very sobering note!
-
Thank you. About answer 4, now that you say it, the switches have a metal casing and are nowhere near big enough to have enough clearance/creepage space.
That has nothing to do with it. The uCurrent will survive in a high voltage system just fine because it is completely isolated floating system (assuming the meter is also isolated). The maximum differential voltage applied to any part of the uCurrent will be the burden voltage (i.e. 10mV max).
It's the same way these guys can survive on 500kV power lines:
High Voltage Line maintained via Helicopter (https://www.youtube.com/watch?v=SkFH8lLvKZ0#)
But yes, the switches are metal, the connectors are exposed, as will be any multimeter you connect up, so it's a human safety issue not a device rating issue.
So you wouldn't want to go touching it after you hook it up. If you don't understand why, then I'd suggest that you shouldn't be playing with high voltage systems!
-
Well it is also a device safety issue as the µcurrent probably can't safely contain a mains short circuit which is a necessary condition to higher cat's.
-
The maximum differential voltage applied to any part of the uCurrent will be the burden voltage (i.e. 10mV max).
I knew, but I wanted to be sure, maybe I did forget about someting. Well, I won't use it for mains anyway, so, don't worry!
What I don't get now is this circuit: http://www.analog.com/static/imported-files/tech_articles/Simple-Circuit-Measures-the-RMS-Value-of-an-AC-Power-Line-MS-2405_Final.pdf (http://www.analog.com/static/imported-files/tech_articles/Simple-Circuit-Measures-the-RMS-Value-of-an-AC-Power-Line-MS-2405_Final.pdf)
Why not use a simple voltage devider and put the whole circuit in a plastic enclosure?
-
The maximum differential voltage applied to any part of the uCurrent will be the burden voltage (i.e. 10mV max).
I knew, but I wanted to be sure, maybe I did forget about someting. Well, I won't use it for mains anyway, so, don't worry!
What I don't get now is this circuit: http://www.analog.com/static/imported-files/tech_articles/Simple-Circuit-Measures-the-RMS-Value-of-an-AC-Power-Line-MS-2405_Final.pdf (http://www.analog.com/static/imported-files/tech_articles/Simple-Circuit-Measures-the-RMS-Value-of-an-AC-Power-Line-MS-2405_Final.pdf)
Why not use a simple voltage devider and put the whole circuit in a plastic enclosure?
Because the circuit you linked also does the RMS conversion and the AD8436 does have a somewhat low and imprecise input impedance of something like 8 k?, which would load a voltage divider and alter the dividers ratio.
The AD8436 has a built-in buffer with high-input impedance, too, but these are not safe to drive beyond the supply rails.
Using a dedicated diff-amp with very high, supply independent common mode range solves these issues.
-
That has nothing to do with it. The uCurrent will survive in a high voltage system just fine because it is completely isolated floating system (assuming the meter is also isolated). The maximum differential voltage applied to any part of the uCurrent will be the burden voltage (i.e. 10mV max).
It's the same way these guys can survive on 500kV power lines:
High Voltage Line maintained via Helicopter (https://www.youtube.com/watch?v=SkFH8lLvKZ0#)
But yes, the switches are metal, the connectors are exposed, as will be any multimeter you connect up, so it's a human safety issue not a device rating issue.
So you wouldn't want to go touching it after you hook it up. If you don't understand why, then I'd suggest that you shouldn't be playing with high voltage systems!
Thank you. This was of course already clear to me, but the previously discussed analog device app note caused some confusion.
Just to be clear: I like the µcurrent very much!
I like the device very much!
-
The AD8436 has a built-in buffer with high-input impedance, too, but these are not safe to drive beyond the supply rails.
Using a dedicated diff-amp with very high, supply independent common mode range solves these issues.
I have some other n00b questions:
What if I use a 'normal' opamp instead of a differential amplifier?
I'll make sure that the input won't get to close to the supply voltage, I'll also use a device with sufficient input impedance to avoid loading the resistive devider.
Will there still be the possibility of damage if there is a input voltage applied when the device is powered down?
-
Will there still be the possibility of damage if there is a input voltage applied when the device is powered down?
Well .. it depends 8)
In every (most) datasheet there is a section "absolute maximum ratings". If you exceed one or more of this ratings, it is likeley that the device gets damaged.
In this case you have to calculate what currents will flow, and what voltages are on the device when there is no power to the device. As long as you stay within "absolute maximum ratings" you will be ok.
-
From the max4239 datasheet:
Absolute maximum for all pins other than Vcc and GND: Vgnd-0,3 and Vcc +0,3. Does this also apply when the device is powered down? So, the input voltage has to be from -0,3 to +0,3 since the Vcc and GND are both 0?
-
From the max4239 datasheet:
Absolute maximum for all pins other than Vcc and GND: Vgnd-0,3 and Vcc +0,3. Does this also apply when the device is powered down? So, the input voltage has to be from -0,3 to +0,3 since the Vcc and GND are both 0?
Applying a voltage to a pin, while the device is not powered on can result in several things:
- if the input is a mosfet or simular and it is behaving like a small capacitance
- if there are input protection diodes (or simular constructs due to the die) current will flow through the diodes, lifting the power supply lines.
These are the most common, but there are more. You have to study the datasheet.
n.b. I often put a zener diode over the power lines, with a value a bit higher then the supply value. When the power lines get lifted, energy is absorbed in the zener diode, preventing a higher lift of the power lines then the zener voltage.
All to be with the "absolute maximum ratings" :P
-
From the max4239 datasheet:
Absolute maximum for all pins other than Vcc and GND: Vgnd-0,3 and Vcc +0,3. Does this also apply when the device is powered down? So, the input voltage has to be from -0,3 to +0,3 since the Vcc and GND are both 0?
Correct. If Vcc is zero and you apply, say, 5 V to an input current will flow into the input, over the ESD/substrate diode, to the positive supply rail. Large currents (>20 mA) may flow (e.g. charging decoupling and filter capacitors, powering devices), destroying these diodes and damaging the IC structure. So while technically most chip makers don't specify much about these diodes it is often (for hobby-use) okay to have a bit of current flowing there in some circumstances, if the current is limited to sane values (<1 mA). For a commercial product or something that should be reproducible and reliable this is not really acceptable IMHO. Adding discrete diodes and/or additional Zeners to limit voltages is a safe alternative. Or, using specialized ICs like the diff-amp in the Analog appnote.
-
I understand a lot more about it now! I can do some experiments now. Maybe I will just disconnect the inputs on power down ore something like that.
-
Maybe I will just disconnect the inputs on power down ore something like that.
You could also add protection diodes for when power is off, like I did:
https://www.eevblog.com/forum/crowd-funded-projects/current-gold-on-kickstarter/msg448208/#msg448208 (https://www.eevblog.com/forum/crowd-funded-projects/current-gold-on-kickstarter/msg448208/#msg448208)
-
The reason I ask is because differential amplifiers seem to have a higher offset voltage. The best I have found so far is the OP07C, well I'm not sure this thing is a differential amplifier, Texas Instruments calls it a opamp with true differential input.
http://www.ti.com/product/op07c (http://www.ti.com/product/op07c)
Input offset is 0,15mV
Now I'm wondering how multimeters deal with the input. I guess a good multimeter can handle input voltage without being powered on. But it is also common for a multimeter to have a 0,1 mV resolution. How do they deal with this ?
I guess it has a normal opamp which is only connected to the input if a low voltage is detected; combined with diode protection to prevent violation of the common mode voltage during the autorange delay.
By the way, does anyone has some advice on literature about (op/instrument/differential)amlifiers? And maybe even about multimeter input circuits? I start to feel like a moron right now...
-
Most DMMs have >10 M? input impedance. These 10 M? are usually always in series with the voltage input to the DMM.
If you use limiting diodes the maximum current flow through them will be ~100 µA at 1 kV input voltage.
High resolution DMMs use other approaches, which are more complicated. See for example the input circuitry of the HP 3478A multimeter (5.5 digits). It uses a multitude of measures against input overload, including two spark gaps, a NTC, voltage limiting diodes, active voltage protection (for the ohms current source) etc.
(Note: when looking at the 3478A input circuitry don't get confused by the front/rear panel switching. It's a bit involved, because there is now amps input on the rear. Also, the input hybrid is the same as for the 3457a, which has a few more features, connection-wise. So you can ignore the GUARD and precharge/guard amplifier. All the FETs in the hybrid are shown w/o gate ; the gate is controlled by a shift register internal to the hybrid. It's essentially a shift register + FETs + guard amp + trimmed resistors + some relay drivers.)
-
I have found the HP 3478A service manual, very interesting.
I have also found an appnote from Intersil, which describes a 10µV resolution autoranging multimeter: http://www.intersil.com/content/dam/Intersil/documents/an02/an028.pdf (http://www.intersil.com/content/dam/Intersil/documents/an02/an028.pdf)
-
I came up with some other questions.
What are the limiting parameters when selecting an opamp for low voltage measurements?
From the datasheet of the max4239 (used in µcurrent):
offset: 0.1µv typ/2 µV max
offset drift: 10nV/°C
long-term offset drift: 50nV/1 000hr
peak to peak input voltage noise: 1,5µV
The µcurrent (classic) has a resolution of 10 pA. That translates into 0,1 µV across the 10K shunt. Does this mean I can use a modified µcurrent for measuring those kind of voltages?