Talk:Varistor

Varistor vs. TVS

There are two significant differences between the Varistor and a silicon Transient Voltage Suppressor(TVS). Varistors are typically an order of magnitude more leakier, ie they will drain a battery 10X faster. Also they degrade over time. This shows up after repeated transient events. Mike

Incorporated. Mike: you could have done that yourself. -- Tim Starling 03:22, Oct 24, 2003 (UTC)

What about the other figures of merit? Capacitance, max. energy, etc.?-- Tim Starling 03:26, Oct 24, 2003 (UTC)

I think the comparison is incomplete/misleading, at best. Because MOVs are a bulk-effect device, they tend to be able to absorb more surge energy (joules) for a given size package. A diode-derived design puts all the action at a junction, an inherently thin slice of the device, concentrating the energy/heat that needs to be dissapated, so a diode of a given size is likely to be less able to handle as much energy as a MOV. (Also, can someone supply some links/references so that we can check the details on leakage etc.?)

This source says that the leakage current is similar: http://www.lightningtalks.com/LightningProtection.htm detailed Comparison of Surge Suppression devices

Metal-oxide varistor (MOV)
+ Up to 70 000 Amps surge
+ Lifetime @ 100 Amps, 8x20 uS pulse shape: 1000 surges
- Shunt capacitance >500 pF
- Leakage approximately 10 micro amps

Avalanche diode
+ Lifetime @ 50 Amps, 8x20 uS pulse shape: infinite
+ Shunt capacitance: 50 pF
- Low surge capability: 50 Amps @ 8x20 uS pulse shape
- Leakage approximately 10 micro amps

69.87.203.165 14:11, 9 November 2006 (UTC)

Reliability

Hi I'm Jeff Reagan. I'm new to this. I take issue with the following:

Varistors will almost always eventually fail for either of two reasons.

I counter that Varistors are extremely reliable. Predicting each varistor will fail gives an overly bleak assessment by my estimate. High continuous power dissipation "thyrite" disks are mounted in stacks to clip reflected power in line-type modulators and induction heaters, where some mismatch is desired or expected, becoming especially brutal with a shorted or open load. Power rated thyrites won't normally fail. Resembling grinding wheels with metalized faces, they don't degrade with use. There is no manufacturer-specified limit to cumulative coulombs transferred. Utilities use thyrites for power line protection. While they can go bad, good reliability is achieved for even the most demanding application. This is done by combining enough devices with overhead ground-shield wires. Lightning will destroy a household appliance on a direct hit. No varistor can change that. Miles away our concern is with indirect transients resulting from capacitive/inductive coupling. These surges are manageable using varistors. Surge severity diminishes with distance. End of line transient voltage could prove excessive if left unprotected. Proliferation of varistors came with widespread adoption of switching power supplies, bringing ubiquitous transient voltage supression. Large transient energy can be spread across numerous devices. Acting together transient voltages are quenched reliably by the lumped mass of distributed Varistors, which is typically quite large. Miles away from lightning strikes such networks of varistors will protect and serve indefinitely.Jeffreagan (talk) 02:28, 21 November 2016 (UTC)

@Jeff, this article seems primarily aimed at MOVs, the small "capacitor shaped" disks. It lacks a lot of information about the big piles used in power transmission. Perhaps you can expand it?

I did update the article a couple of weeks back to emphasize how information is more about MOVs than super high power stuff. I don't know enough (and could not find too much information) about the area you've been talking about here.HiTechHiTouch (talk) 16:54, 21 November 2016 (UTC)

Jeff, your initial comments seem plausible. If you can find WP:RS to back them up, you can add them to the article. Reify-tech (talk) 21:04, 21 November 2016 (UTC)

A capacitor? Hardly!

The lede of the article currently contains the following statement:

and is not a resistor but in fact a capacitor.

This is horsepucky. While many MOVs usually look like capacitors (in the form of flat disks) and have some parasitic capacitance (which can be measured when you're below their threshold voltage), they are most definitely not capacitors in the sense that an Electrical Engineer would typically mean it. Once you're above their threshold voltage, they are primarily resistive, not capacitive. They are non-linear resistors.

If no one objects, I'll (eventually) re-write this part of the lede.

Atlant 11:40, 25 May 2005 (UTC)

Of course they're not capacitors. I rewrote the intro. --Heron 16:47, 25 May 2005 (UTC)

Thanks! Atlant 23:31, 25 May 2005 (UTC)

Action of Varistors

I dont think the varistor acts as a spark gap, but as a non linear resistor as it says in the first line. Under high applied voltage, it does not breakdown like a gap, but merely goes low resistance, maybe the author was thinking of a miniature gas discharge tube (spark gap). Would anyone like to comment before I alter??Light current 18:57, 9 August 2005 (UTC)

The "spark gap" line is so rubbish that I can't wait for you to delete it - it's gone. --Heron 20:58, 9 August 2005 (UTC)

Must have been some confusion with "Gap caps"? Might be useful to mention them in the comparison table?

The mention of gas discharge tubes being used in RF is true- but that doesn't mean they are effective! They normally break down at a higher voltage than the components they supposedly protect! Talk about false sense of security! Diodes are only slightly better. Better designs use a small inductor which effectively bypasses frequencies below the RF being used. I know this gets off topic, but I sure hate to see the reference to discharge tubes in RF remain somewhat misleading- even if it is "standard practice". What was sufficient for vacuum tube technology is NOT sufficient for semiconductors- despite the tendencies of old habit's tendencies to remain unquestioned. 137.164.225.34 (talk) 21:12, 12 February 2010 (UTC)

Vague intent

I am perplexed by the following quote from the MOV section:

When used in communications lines (such as phone lines used for modems), high capacitance is undesirable since it absorbs high frequency signals, thereby reducing the available bandwidth of the line being protected.

Can someone indicate whether a varistor is part of the problem or part of the solution? Vonkje 08:27, 14 September 2005 (UTC)

The varistor is part of the solution to one problem and part of another problem.

It's part of the solution to the problem of preventing lightning transients (and other electrical disturbances) from getting into the equipment and blowing it up. But any parasitic capacitance of the varistor tends to bypass (short out) high frequencies that the modem depends upon and the voice phone likes to have, so parasitic capacitance is a problem. So our ideal varistor (in such applications) would have no parasitic capacitance (and be free, have infinite impedance below the threshold voltage, and other impossible characteristics :-) ).

Atlant 23:51, 22 February 2006 (UTC)

Electrical Symbol

Can someone add an electrical symbol to the article? --PeterMarkSmith 08:18, 17 January 2006 (UTC)

I would second this motion. --Matejhowell 22:26, 22 February 2006 (UTC)

• Varistor symbol and V-I characteristics 69.87.200.87 02:36, 14 November 2006 (UTC)

Should the 'Traditional varistor schematic symbol' shown top right be different from that for a DIAC ? - Rod57 (talk) 18:15, 25 April 2017 (UTC)

Relationship with polyswitches

I ran across the Polyswitch article (referenced by one article as "resettable fuses"), but I believe that there is some relationship between a varistor and a PTC or NTC resettable fuse/polyswitch. Can further research be done? I don't know enough about both, or maybe it is the internal workings that are different... --Matejhowell 21:27, 1 March 2006 (UTC)

Polyswitches are sort of the inverse device of a varistor: Varistors become low resistance when exposed to a voltage greater than their threshold voltage (so are handy for shorting out/clamping transient voltages) while polyswitches become high resistances when forced to carry a current greater than their threshold current (so they are handy for disconnecting an overcurrent condition). Both devices are self-reseting. You could use them together: The polyswitch in series with the supply and the varistor downstream of the polyswitch in parallel with the load.

Atlant 12:37, 28 March 2006 (UTC)

The article says "Polyswitch (trademark) - a combined varistor and PTC thermistor " but I do not think this is true. A polyswitch is certainly a thermistor, but how exactly is it a varistor? The polyswitch article does not mention anything about "varistors" in any case. I think this reference is inaccurate, but I am not sure. 68.228.0.128 07:13, 7 September 2006 (UTC)

Correct. I made a change; see if you like it.

Atlant 12:29, 7 September 2006 (UTC)

Varistors in shunt or series?

I thought that if a Varistor protects a circuit from excessive voltage spikes, etc, then surely the resistance would have to INCREASE with an increase in voltage, not the other way around, as is stated towards the middle of the article. I'm still learning this stuff, so can someone please help? —This unsigned comment was added by 203.87.68.209 (talk • contribs) .

Varistors are usually hooked up shunting (in parallel with) a load to be protected. If an over-voltage condition occurs, they decrease rapidly in resistance, shorting out most of the energy of the overvoltage (at least until the point where the energy dissipated within the varistor heats the varistor to the point of destruction). The energy of the transient is dissipated in varistor and the source impedance of the voltage rail.

By the way, you can sign your posts on "talk" pages by including four tildes (~~~~) at the end of your post. When you press "Save page", this will be converted into your username or ip address in a handy Wikilinked format. A timestamp will also be included.

Atlant 12:32, 28 March 2006 (UTC)

Sacrifice?

When encountering a surge, I have read that MOVs will sacrifice themselves to protect loaded equipment. If this is true, I think it should be included in the article. Chris 13:06, 30 October 2006 (UTC)

Each time a MOV absorbs a surge, it tends to be somewhat degraded. The extent of the degradation depends on the size of the surge (energy in joules) compared to the rating of the device. If the surge is too large, the device may be greatly degraded, or fail, perhaps spectacularly. MOVs tend to fail short -- they do everything possible to sacrifice themselves to protect the load, their mission. 69.87.203.165 14:11, 9 November 2006 (UTC)

An MOV should only degrade when each spike is shunted. Manufacturers provide charts for this degradation - normal operation. At no time must an MOV operate well outside those chart curves; MOV should not 'sacrifice itself' - abnormal operation. That vaporization occurs when manufacturer's "Absolute Maximum Ratings" are significantly exceeded. A 'sacrificed' protector quits early, and may create a fire hazard as shown in pictures from the Gaston County Fire Marshall report (see References section).

Catastrophic MOV failure is an unsafe event. Efforts have been made to limit this unacceptable failure in UL1449 2nd edition. Still this failure exists as the Fire Marshall repeatedly demonstrates.

Unfortunately, too few MOVs inside a protector located on a desk or on a rug can create a fire hazard. MOVs located in less hazardous locations (i.e. circuit breaker box) would significantly diminish this threat. Those protectors are typically better constructed and include more joules.

When a protector is undersized, its manufacturer may install a smaller or faster temperature sensitive fuse in series to disconnect MOVs faster (see 'Gaston' pictures). Faster disconnecting means diminished appliance protection; a trade off between better protection or increased human safety. If fire does not result, then a failure light may indicate protector failure. That light implies a protector was undersized; that MOVs failed or disconnected due to a surge; that a thermal fuse blew to protect human life. Many confuse this with "surge protector failed to save my computer". Instead, a fuse blew to save human life from a protector with insufficient joules.

Install MOV protectors to protect household appliances. To avoid catastrophic failure (fire risk), a protector joules rating is increased. IOW minimum size for a residential 'whole house' protector is typically more than 1000 joules. For plug-in protectors where all MOVs are not used during a surge, the same minimum number may be 3000 joules; not including MOVs (joules) installed for other ports such as phone or cable.

If properly sized (sufficient joules), then a protector shunts a surge without anyone knowing of that surge. It must only degrade; not vaporize or ‘sacrifice itself’. A protector should have sufficient joules to remain functional after each surge.

An MOV is degraded when its voltage changes by 10%. See a discussion in "MOV Testing" for further details. MOV should expire by degrading; definitely not by exceeding "Absolute Maximum Ratings". Vaporizing MOVs (a safety hazard) typically should not happen if a protector was sufficiently sized. Still, a fire risk exists which is why protectors should never be located in high risk areas such as among flammable materials such as a desktop or carpeting, or in a bedroom.

I know! I know! It's w_tom. Lemme do the next one. —Preceding unsigned comment added by 69.95.190.147 (talk) 15:45, 1 October 2007 (UTC)

MOV Testing

How can you test your MOV/surge protectors to see if they have been degraded etc?

Ideally, you would first visually inspect the MOVs to see if they appear damaged. But if they look OK or you cannot get access to them, you can test them electrically.

Ideally, you would do I-V curve tracing under various conditions. (More information here: http://www.repairfaq.org/sam/semitest.htm ) And you would study the datasheets, such as the pdf here: http://www.newark.com/NewarkWebCommerce/newark/en_US/endecaSearch/partDetail.jsp;?SKU=09F2091&CMP=KNC-G10000645 )

But if you are just trying to get by, the first thing to do would be to check the leakage current at the highest voltage you can easily and safely attain, perhaps just whatever your ohm-meter puts out. Beyond that, you could use the power line as a voltage source (120VAC). You could very carefully check the leakage current, ideally using a current limiting resistor in case the device you are testing is leaky or shorted. With a rectifier and capacitor, you could generate a higher DC voltage, 170VDC. With another rectifier and capacitor, you could double that to 340VDC. Then with a current limiting resistor of about 200K ohms, you would be able to properly measure the nominal VDC of a MOV under a standard leakage current of about 1mA. The big safety concerns are to make sure nothing has to dissipate too much power and gets too hot or burns up. To not overload your DMM or other test instrument with over-voltage etc. And most important, not to get a shock!

69.87.193.213 00:23, 10 November 2006 (UTC)

Questionable products

There is no mention yet of the controversy over whether surge suppressors should be used for general household use. While they are often recommended (sales are profitable),

1. computer psus already have better protection built in,
2. surge suppressors' fire risk adds to destruction of life and property
3. the article so far fails to even mention the difference between inductive surges, which is what such units are able to absorb, and near lightning strikes, which they normally have far too small an ability to deal with.

While they're useful for electronics running on portable generators, which do produce surges a MOV can do something about, on a standard low impedance mains supply they're really of little value.

Yes, controversy and argument... but I do think the article needs to acknowledge the real issues, and start by understanding what type of surge an MOV can dissipate, and what it can't. Lack of awareness seems to be most common on this point, and so far the article seems to perpetutate the popular sales line. Tabby (talk) 15:41, 31 December 2007 (UTC)

1) Discussion here is limited to varistor and its applications. The discussion includes varistor failure modes inside application devices. But how those application devices perform is a topic located elsewhere.

2) Protection from inrush current is also an irrelevant electric power event. Similar devices such as gas discharge tubes and avalanche suppression diodes are mentioned here since all conduct current when voltage becomes excessive. Inrush current limiting is a power disturbance and solution irrelevant to what a varistor does.

3) If varistors are not designed for lightning protection, then why do varistor datasheets routinely specify numbers unique to lightning strikes (ie 8/20 microsecond pulses)? Varistors are for protection from lightning strikes and from other similar and lesser current pulses. But varistors are only one component of the application device. Whether an application device (ie power strip protector) provides lightning protection involves other considerations that must be discussed elsewhere. Varistors are designed for use in lightning protection as made obvious by numbers in datasheets such as "8/20 microseconds".

69.72.7.216 (talk) 03:02, 25 January 2008 (UTC)

Hello again, Jeff Reagan here, still a newcomer and unclear about how to proceed. I see Tabby's entry regarding generator surges above. It's almost ten years old. I know from experience how generators behave. Citing a source confounds me. I only consider specifications I was required to meet and my personal memory of over-voltage trip-point settings I've seen in power plant over voltage protection relays. So what do I do with these entries? No curator has visited this talk page in some time. My question also relates to another "talk entry" I made above:

Generators produce excessive output voltage on a step-change going from full-load to no-load, typically returning to nominal within a half second. Sluggish overshoot recovery results from voltage output tracking field decay within the rotor (intrinsically slow, even if critically damped). Magnitude usually crests at 1.5 times nominal voltage. Available power can be quite large, enough to destroy a poorly selected varistor. Clamping such lengthy surges would be cost prohibitive for most cases. Thus exposed devices should be rated to withstand 1.5X nominal voltage momentarily, varistors included.Jeffreagan (talk) 00:58, 21 November 2016 (UTC)

"Hot" and "8/20 microseconds"

October 2008 changes are removed and described below.

Hot refers to a wire with voltage different from ground. "Hot" is the standard term in the US, Canada, and other nations. "Live" is the term in the UK. "Active" is an Australian term and rarely used in most technical publications. Most common term is "hot" especially where varistors are most commonly used. Changing this word to "active" only confuses most of the world. Details also found in Wikipedia under Electrical wiring (United States) Electrical wiring (UK) AC power plugs and sockets

8/20 microseconds is an industry standard waveform that quantifies varistors and is routinely listed in data sheets. "2/5" is not an industry standard parameter. Varistors respond in nanoseconds as discussed in an existing footnote in "Varistors compared to other transient-suppressors". "8/20 microseconds" is an industry standard benchmark to quantify a varistor's response to a typically fast transient.

Better description of MOV polycrystalline structure & mfr

GE's MOV was new in 1972 and fortuitously I stumbled across an article by 4 General Electric R&D people that sets forth how you get from "metal oxide" to "polycrystalline diodes-in-bulk" better than anything I knew before. The ref is:

Metal-oxide varistor: a new way to suppress transients." J.D. Harnden, Jr., F.D.Martzloff, W.G. Morris, and F.G. Golden (all General Electric), Electronics, 9 Oct 1972, p91.

Paraphrasing the article's text:

"The MOV has an encapsulated polycrystalline ceramic body, with metal contacts and wire leads. The essential ingredients of the polycrystalline ceramic are zinc oxide and bismuth oxide. They are mixed with proprietary powdered additives (typically cobalt and manganese oxides) and pressed into disks for sintering at temperatures over 1,200 deg C. Because the bismuth oxide is molten above 825 deg C, it greatly assists in the formation of a dense polycrystalline ceramic through the process of liquid-phase sintering. During cooling, the liquid phase forms a rigid amorphous coating around each zinc oxide grain, yielding a microstructure of zinc oxide grains that are isolated from each other by a thin, yet continuous, intergranular phase. It is this two-phase microstructure that creates innumerable diode junctions and thus the device's nonlinear voltage-current characteristics."

That is so pleasing! I never understood it before. Would one of you who have worked so hard to build the MOV entry put some of this text into the first main section after the intro, "Metal oxide varistor"?
Jerry-va (talk) 01:30, 8 August 2009 (UTC)

article scope

This article is focussed almost entirely on MOVs, it seems like the authors did not appreciate the larger scope of the subject and the history that goes back to the 1920s. Kbrose (talk) 15:39, 13 September 2014 (UTC)

The article coverage could indeed be expanded. If you have relevant information backed by WP:RS, please feel free to add it to the article. Reify-tech (talk) 21:04, 21 November 2016 (UTC)

Antiparallel? That should be antiseries!

Two diodes in antiparallel (such as suggested by the circuit diagram) will form a varistor, but with a breakdown voltage that is no higher than the diode threshold voltage (0.2 for Ge, 0.7 for Si). I am pretty certain that should be antiseries. Grab your electronics kit and try for yourself. This implies that the electrical diagram is just wrong, and that the "DIAC" symbol has a very good reason for not being used anymore. This will probably influence the explanation of the working of the MOV a bit. - Jurjen — Preceding unsigned comment added by Jurjen B (talk • contribs) 19:48, 10 November 2019 (UTC)

Are you assuming that a varistor should have a higher breakdown level than just a diode? The breakdown level is certainly a parameter of interest, but it can be changed by using multiple diodes in series for each direction. For example, varistors used to be constructed with multiple CuO disks for each direction. This seem quite analogous to the ZnO MOV, as the grains are randomly oriented, providing statistically the same number of "anti-parallel" grain orientations for each conduction direction. And this also provides the same number of "series" orientations, so that the overall breakdown level can be adjusted with the size of the device. Kbrose (talk) 19:07, 12 November 2019 (UTC)

Pronunciation

How do you pronounce "varistor"? It initially looks like 'VAIR - iss - ,tor but when one discovers that it means "variable resistor" one tries va - 'RISS - tor. Please can someone let us know which it is? Thank you. UBJ 43X (talk) 12:04, 8 November 2021 (UTC)