Multimeter overload protections

[Wytnucls]

There seem to be some confusion about what is required in a multimeter for adequate protection.
Correct me if I'm wrong:
Fuses are installed in series on the A and mA connectors to act as protection in case of a short, due to low impedance circuit, while measuring high voltage, with the test lead left in the A/mA connector and also to protect against possible high currents on the circuit under test.
PTCs are installed in series on the Volt/Ohm/Cap connector, to protect against overvoltage and spikes on those ranges.
MOVs are installed between connectors and COM to act as a short in case of overvoltage on all ranges.

[saturation]

Yes, that generally right. There are other protective elements used for high voltage, that are similar to MOVs: gas discharge tubes or TVS diodes that could be added/substituted.  PTC in series protect against much slower overvoltage, such as if you connect a 600V limited DMM to 1000V, but for transients, milli to micro second high voltage spikes, MOVs or TVS diodes respond faster and work by short circuiting that energy to commons.

[Wytnucls]

Gas discharge tubes are quire rare in meters, I noticed. Maybe they are on the expensive side.
I find it strange that national regulations ignore these specific protection devices, ony requiring meters to survive high transients on all ranges, in accordance with their claimed categories. Proper trace separation to avoid arc-overs seems to be the main concern.
I would have thought that HRC fuses would be required for all CAT III and IV devices, but that is not the case. Hence, some manufacturers are still using glass fuses in their equipment. Perhaps they are worried about burden voltage. 

CAT II 600V 4000V peak impulse transient 12 Ohm source
CAT II 1000V 6000V peak impulse transient 12 Ohm source
CAT III 600V 6000V peak impulse transient 2 Ohm source
CAT III 1000V 8000V peak impulse transient 2 Ohm source
CAT IV 600V 8000V peak impulse transient 2 Ohm source
CAT IV 1000V 12000V peak impulse transient 2 Ohm source

[Neilm]

In addition to all that a well designed meter will consider the aspects of using these components that you may not think of at first.

 For instance - imagine that you have a multimeter. It is well designed and has a fused 10A connection. Now imagine that you have just connected your test leads to a high power voltage source with the leads plugged into the 10A connections. First thought is the fuse would blow and everything would be fine. Right?

WRONG.

The 10A fuse might have 50A+ going through it when it blow. It is also flowing through the test leads you connected. Those leads will have some inductance, depending on the leads this could be many nanoHenries. When the fuse opens the voltage at the fuse will start to rise (basic physics). Depending on many factors, this could be many kV. This voltage rise is why you don't find cheap glass fuses in a decent meter.

More over, if the designer does not consider this voltage rise they might find that the voltage goes so high it might precipitate an arc. If your input terminals are too close together the arc will form between them (before the fuse) and the high power voltage source then continues the arc and you get an explosion within your meter. This is why you will see plastic between terminals on expensive meters - I know that Dave has remarked on them when he was tearing down meters.

You might be interested to know that I did once test a multimeter that suffered exactly this problem and exploded when the voltage was connected.

Yours

Neil

[saturation]

Gas discharge tubes for DMM can run 50¢ to $1 or so each, and unlike MOVs, do not slowly wear out with use. Not sure its rare, but the maker has many options for transient protection.  In this old eevblog tear down of the Agilent 1272a, the diode like elements near the green pea like PTC? MOV are likely miniature discharge tubes, one for each line.
 
http://www.bourns.com/ProductLine.aspx?name=gas_discharge_tubes




IEC rules for at least CE rating, and if they have separate certification like UL, TUV, or ETL, will test the meters for conformance to CAT and Pollution degree rating regardless of the electronics [ MOV or TVS diodes, etc.,] and technique used, in the end that's what counts.

IIRC, glass fuses are not allowed in good DMMs, so whatever meters those are likely not IEC qualified; maybe for CAT 1 or 2, but as DMM need to be CAT 3 and up, a simple rule is it should be capable of the transient anticipated in its highest rated range, say 600V or 1000VAC.

Gas discharge tubes are quire rare in meters, I noticed. Maybe they are on the expensive side.
I find it strange that national regulations ignore these specific protection devices, ony requiring meters to survive high transients on all ranges, in accordance with their claimed categories. Proper trace separation to avoid arc-overs seems to be the main concern.
I would have thought that HRC fuses would be required for all CAT III and IV devices, but that is not the case. Hence, some manufacturers are still using glass fuses in their equipment. Perhaps they are worried about burden voltage. 

[ptricks]

I have a meter that implements a crowbar on the voltage lines so that when excess power is detected it blows the fuse.

[Wytnucls]

I have a meter that implements a crowbar on the voltage lines so that when excess power is detected it blows the fuse.

Fuses in a DMM are only there to protect the A/mA/uA range circuitry. They have nothing to do with the voltage ranges (V/mV). Most meters are limited to 10A measurement. Mine actually has a time limit too for currents above 5A. Fast glass fuses are fine for that job.
All ranges (A/mA/V/mV/Ohm/Cap/Temp) usually have a built-in 1000V protection (crowbar circuit? PTCs?).
If the manufacturer wants to protect the equipment and user from high and fast transients and misuse of the meter, like trying to measure very high currents or high voltages while still on the mA/A range, then some serious protection is required to minimize the damage: HRC fuses, MOVs, Gas discharge tubes, etc..

[Wytnucls]

IIRC, glass fuses are not allowed in good DMMs, so whatever meters those are likely not IEC qualified; maybe for CAT 1 or 2, but as DMM need to be CAT 3 and up, a simple rule is it should be capable of the transient anticipated in its highest rated range, say 600V or 1000VAC.

Can you double check on that? I have never seen any regulation mentioning types of fuses to be used in DMMs.
Ceramic, sand-filled fuses, are cheap and would be, in my humble opinion, an improvement at a minimal cost.
HRC fuses on the other hand, for some unknown reason (to me, at least!), are quite expensive.

[Wytnucls]

Gas discharge tubes for DMM can run 50¢ to $1 or so each, and unlike MOVs, do not slowly wear out with use. Not sure its rare, but the maker has many options for transient protection.  In this old eevblog tear down of the Agilent 1272a, the diode like elements near the green pea like PTC are likely miniature discharge tubes, one for each line.

I know Dave identified them as gas discharge tubes, but they look more like diodes of some sort to me. I could be wrong of course. It looks like they are marked as RV on the board, whatever that means (Resistor Variable?).

[ptricks]

I have a meter that implements a crowbar on the voltage lines so that when excess power is detected it blows the fuse.

Fuses in a DMM are only there to protect the A/mA/uA range circuitry. They have nothing to do with the voltage ranges (V/mV).

Not always true. The one I referred to was designed for very high precision and the circuit is there for voltage protection. Using other methods to protect the input would have altered the readings. The meter is used for measuring voltage references to 0.00000 places

[saturation]

IEC 60127.  USA rules require fuses be replaced as specified by the maker; IEC suggests you can substitute an equivalent fuse, but if HRC fuses are specified, it would not be recommended to substitute it regardless of cost. Sand filled fuses are typically HRC but if not tested or specified as HRC, it should not be considered so.

Separately, the DMM would not even meet IEC safety regulations for > 500V if it didn't use HRC type fuses of any sort; there are essentially no glass fuses rated above 500V.

Jump to the fuse section.

http://www.ce-mag.com/archive/02/09/lamothe.html

IIRC, glass fuses are not allowed in good DMMs, so whatever meters those are likely not IEC qualified; maybe for CAT 1 or 2, but as DMM need to be CAT 3 and up, a simple rule is it should be capable of the transient anticipated in its highest rated range, say 600V or 1000VAC.

Can you double check on that? I have never seen any regulation mentioning types of fuses to be used in DMMs.
Ceramic, sand-filled fuses, are cheap and would be, in my humble opinion, an improvement at a minimal cost.
HRC fuses on the other hand, for some unknown reason (to me, at least!), are quite expensive.

Enclosed is the spec sheet of 2 types of high voltage fuses, one is an equivalent and the other the OEM Bussmann fuse for the Fluke 87V; on Amazon the Fluke fuses are priced about $8, the high prices you see elsewhere is likely price gouging.  Even premier sellers like Grainger.cim sell the fuses at around $10. 

[Wytnucls]

Those fuses, you linked up to, are rated at 1000V. If I'm not mistaken, the Fluke 1A fuse has a rating of 600V.
The recommended fuses (M205) for my meter are rated at 240V.
What would be the implications of using a fuse with a higher voltage rating than the original low voltage one?

There are some cheap ceramic fast blow fuses in the M205 dimensions (20x5mm) and 3AG (30x6mm) rated at 10A/1A/240V.
SIBA also makes some DMM fuses in an odd size (32x6.35mm), which are a bit more expensive.
Fluke uses 5AG fuses in their meters (38x10mm)
Ceramic Bussman fuses, like those found in British plugs (24x6.35mm), are also rated at 250V, but I'm not sure if they are fast blow or not.

[SeanB]

Plug fuses are normally a slow blow type, in most cases, as they are there to protect the cable.

Higher voltage fuses are capable of breaking and withstanding a high voltage when open and during the time they are breaking. Using a low voltage fuse at high voltage runs the risk that on breaking an arc can be sustained for a period after the fuse has opened, which allows a destructive current to flow or which bursts the fuse and can cause the surrounding material to burn.

[Wytnucls]

Datasheet for Bussmann BS1362 says they are rated as Medium/Fast, so probably not a great choice for DMMs.
http://www.farnell.com/datasheets/10867.pdf
Some retailers list them as 'Fast Blow'.

What happens if a high voltage fuse, say 1000V, is used in a DMM, instead of the regular 250V DMM manufacturer fuse with the same rating (10A)? Is the DMM circuitry likely to be damaged in case of a high voltage transient?

Some interesting paper from Bussman:
http://www.cooperindustries.com/content/dam/public/bussmann/Electrical/Resources/Data%20Sheets/Bus_Ele_DS_8005_Fuseology.pdf

[Amarbir[Lynx-India]]

Gas discharge tubes for DMM can run 50¢ to $1 or so each, and unlike MOVs, do not slowly wear out with use. Not sure its rare, but the maker has many options for transient protection.  In this old eevblog tear down of the Agilent 1272a, the diode like elements near the green pea like PTC are likely miniature discharge tubes, one for each line.

I know Dave identified them as gas discharge tubes, but they look more like diodes of some sort to me. I could be wrong of course. It looks like they are marked as RV on the board, whatever that means (Resistor Variable?).

Sir ,
 looks like mov

[Wytnucls]

Get with the program Amarbir; we're not talking about the big green peas on the board, which are PTCs actually, but the 2 diode-like devices in front of them. 

[T4P]

There are some cheap ceramic fast blow fuses in the M205 dimensions (20x5mm) and 3AG (30x6mm) rated at 10A/1A/240V.
SIBA also makes some DMM fuses in an odd size (32x6.35mm), which are a bit more expensive.
Fluke uses 5AG fuses in their meters (38x10mm)
Ceramic Bussman fuses, like those found in British plugs (24x6.35mm), are also rated at 250V, but I'm not sure if they are fast blow or not.

32x6.3mm fuses are not odd! They are very popular in low-end fluke meters actually. Like the Fluke 18B
But yeah the next time i see a meter with a BS1362 i would be sure to swap the fuse out ( If it was a factory intension on a 32x6.3mm socket ) or the socket out (like on the UT61E)

[Tobias89]

Hello,

I apologize for bringing up such an old topic, but I have a question that seemed appropriate for it. Do multimeters have any kind of overvoltage protection on ohms measure function? For instance, multimeter is rated to about 1100 V, and user is measuring 1000 V legally. Suddenly, he changes function to ohms measurement. Do ordinary serious multimeters (eg Fluke 27) have chance of surviving that?

I ask because I am currently developing an USB multimeter and I am looking for a solution that could survive that. I have found old HP service manual and some Fluke 27 schematics, but I don't see any circuitry there that I think would help a multimeter survive this scenario.

I would appreciate any help!

[ModemHead]

Yes, most decently-designed multimeters should survive this scenario.  The protection mechanism usually involves a voltage clamp and a PTC thermistor in series with the input.  The voltage clamp is often some arrangement of NPN transistors configured as zener diodes (base-collector shorted), as this provides a very low reverse leakage.  When the clamp conducts, the thermistor will heat up and its resistance will rise, limiting the overload current.

That's why there's sometimes a bit of "recovery time" after an overload, where the multimeter will read a little off as the thermistor cools.  You may also find other fairly high-wattage resistors in series with the input as well, that also helps to dissipate excess energy during overloads.  Most Flukes have either a large 3.5K wirewound or a 1K fusible resistor in series with the input.  These components come into play during over-voltage events when the varistors (or sometimes spark gaps) conduct.

The attached image is the input section of the venerable Fluke 77-II, which illustrates the thermistor/voltage-clamp arrangement.

[retiredcaps]

In addition to modemhead's explanation, Dave has 3 videos that are relevant to your question.

http://www.eevblog.com/2012/03/04/eevblog-252-multimeter-ohms-overload/

http://www.eevblog.com/2012/10/19/eevblog-373-multimeter-input-protection-tutorial/

http://www.eevblog.com/2012/10/26/eevblog-376-multimeter-fuse-diode-followup/

[Tobias89]

Thanks for the help!

ModemHead, I'm familiar with that kind of protection, but the problem with my particular chip (HY3131) is that multiple pins are exposed during ohms function (one of them isn't protected). To make the long story short, this is the schematics of HY3131 eval board.



My concern is pin PB0. Here is a schematics of my proposed solution (not too sophisticated ).



The same kind of protection is duplicated from pin RLD to PB0. The problem is that pin RLD is current source, and PB0 is voltage input, so I couldn't just connect left side of R4 to the left side of R6 (then I would measure voltage drop on PTC also).

Do you have any better idea how could I protect PB0 pin?

[ModemHead]

I have little to no design experience, so you can take what I say with a grain of salt.  But I note that Fluke adds only a spark gap to protect their ohms voltage sense line, presumably relying instead on the high impedance of the coupling resistor (1Meg) to limit current during an overload.  Their ASIC input is probably clamped internally?  For 1000 V input, it would only have to deal with at most 1mA current source or sink.  In any case, why clamp the left side of R4 when you could clamp the right side, taking advantage of R4 as a ballast.  No thermistor required as long as R4 is sized appropriately (1W or more.)

[Kleinstein]

The difficult part can be protecting the current source for low resistor ranges (e.g. 200 Ohm or less). Usually this a PTC (or a few in series). To get a decent current of e.g. 1 mA out the PTC can not be very high value, especially if only a realtively low battery voltage is available.  Some benctop DMMs even use a current source that can withstand up to about 1000 V in one direction and a diode for the reverse direction. But this is an expensive solution.

The voltage reading input during ohms measurement is often the same as for normal voltage readings. So no extra protection is required here.

In cheap DMMs spark gaps are sometimes part of the PCB and not extra parts.

[Tobias89]

The voltage reading input during ohms measurement is often the same as for normal voltage readings. So no extra protection is required here.

In cheap DMMs spark gaps are sometimes part of the PCB and not extra parts.

But the normal voltage reading input has a voltage divider (there is no high voltage on the pins of the chip), while here on the ohms voltage input, there would be directly high voltage on pin. That's what concerns me...


ModemHead, on your attached image, I assume the green circuitry is part of current source, and R2 and E1 are protection for voltage input? Correct me if I'm wrong.

[Kleinstein]

The circuit diagram from ModenHead (fluke 77?) has some protection even in the Ohms range:

R2 and E1 provide only very basic protection. E1 will help in voltage ranges only - but this is more against static electricity. In Ohms range, it is only R2 and the chip internal diodes that save the chip. The spark gap will not help in Ohms range. The more critical part is R1, RT1 and Q1/Q2. Here ther PTC RT1 is essential. It will have most of the voltage if the external voltage is there for a longer time. So it will hopefully llimit the current to a level Q1+Q2 can handle.

Deccent new DMMs will survive external voltage in the ohms range. With older meters, especially those before PTCs became polular / available, might get damaged or blow a fuse, that might be used instead of the PTC.

[Tobias89]

Thanks for the answer!

So you are saying that in ohms range, if we connect 1000V to DMM, the only thing that protects it on pin 20(10) is R2 and internal diodes?

[ModemHead]

ModemHead, on your attached image, I assume the green circuitry is part of current source, and R2 and E1 are protection for voltage input? Correct me if I'm wrong.

Yes, the protection circuit marked in green is basically part of the current source, which comes out through the divider resistors and goes out to the unknown resistance through R1.  (The divider resistors are re-purposed as reference resistors in ohms configuration.)  As noted by Kleinstein, this has to be a low-impedance path back to the ASIC to develop enough test current with a low supply voltage, so this node has the extra protection.  Note that there is a voltage sense input for this node as well (RRS=reference resistor sense).  The ratio of this voltage to the one across the input terminals (OHS), and the reference resistor value, yields the unknown resistance.

So you are saying that in ohms range, if we connect 1000V to DMM, the only thing that protects it on pin 20(10) is R2 and internal diodes?

The voltage sense for the unknown resistance on the input terminals (OHS) is protected by E1 and R2, although E1 would only conduct if a massive voltage overload is present, not rated voltage.  So yes, it looks like R2 and internal diodes are key.

[001]

Sorry for necroposting But...

Is it good idea to protect DMMs with new TVS diodes?
TVS diode much faster than varistors etc https://en.wikipedia.org/wiki/Transient-voltage-suppression_diode

Maybe some typical circuit avaliable now?