Gas Discharge tube based mains transient protector......upstream of mains fuse?

[betwixt]

I would anticipate a holding current of just a few mA. Once the gas has ionized it will stay so until the voltage is low enough. I think 10V at 1 Amp is a measure of it's clamping voltage at that current, not an indication that it will start or sustain conduction at that voltage.

As already mentioned, a GDT is not a clamping device in the sense of a MOV or TVS, it has a 'foldback' characteristic so once ignited, it will only add about 10V to the overhead the other device sees. They are also relatively slow and I suspect you are reading the dv/dt the wrong way around, they need higher voltage to ignite them if the pulse is shorter.

The data sheet is headed with 'applications' which clearly show it is intended for signal rather than power line protection. For example you will find them in many telephone 'master sockets' to protect against static discharges on over head wires or on line modem inputs to ground for the same reason. These are all applications where they are the first line of defense but further insulation is present afterwards, not power line situations.

GDTs also suffer a degradation due to sputtering of the electrodes around the interior of the enclosure. The high temperatures of ionized gas eventually causes migration of the electrode metal to the internal surface of the chamber and reduces their ignition voltage and in extreme case can make them go completely short circuit. Look at a neon lamp that has had many hours of use (power indicator for example) and you will notice the glass has a metallic sheen and the light output has diminished. At the much higher current of a GDT discharge the effect is greatly speeded up.

Brian.

 

[Kajunbee]

Littlefuse has a ac120 and ac240 series gdt for mains power. There breakdown voltages are nearly twice the Bourns gdt though.

 

[treez]

Thanks, is the bad case scenario described in the first paragraph of post #15 a potential occurrence?
ie the problem of a mains transient occurring at mains peak and then even when the mains transient has stopped after approx 60us, the GDT still stays conducting and unfortunately tries to "discharge" the normal mains peak voltage....

 

[Kajunbee]

There is a chance of this happening because of follow through current. But the mov if it is working properly should eleviate the chances of this happening. As said before the mov helps extinguish the arc quickly. If the mov stops conducting when voltage returns to normal there's little chance of the problem described in post #15.

What I'm not sure about is what causes the follow through current in the first place. I know that it is a problem in low impedance higher current loads. How does higher current to the load create follow through current in the GDT.
Maybe your lights draw so little current that this may not be and issue.

 

[treez]

Thanks, and i am sure you agree that we need a higher voltage rated MOV than the 270V one we currently have.....We in fact need a 400V MOV. Do you agree?

 

[Kajunbee]

I don't know think going to a higher rated mov is the answer either.

 

[treez]

Thanks, it we dont have a higher mov, then the gdt/mov will start clamping the mains peak if the transinet happens just before mains peak.

 

[treez]

I think we will need the following 400V TVS in series with the GDT instead... do you agree?
SMCJ400A
https://m.littelfuse.com/~/media/el...s/littelfuse_tvs_diode_smcj_datasheet.pdf.pdf

 

[FvM]

Unless you show a substantiated reference for GDT + MOV series circuit application, I prefer to consider it as a useless design idea.

 

[Kajunbee]

Search " EC-640 combining gdt and mov -littelfuse". Hopefully this will help in your decision making.

 

[Kajunbee]

For other related information try " gdt in series with varistor- stack exchange".

 

[FvM]

Search " EC-640 combining gdt and mov -littelfuse". Hopefully this will help in your decision making.

Unfortunately not related to my comment. Did you read EC640 at all? It's discussing industry standard staggered protection ("hybrid surge protection") rather than series circuit.

https://www.littelfuse.com/~/media/...plication_notes/gas_discharge_tubes/ec640.pdf

Also the other reference isn't related to the circuit discussed in this thread. It's about a special MOV + GDT combination where the GDT is connected to earth. You find the same in Epcos/TDK surge arrestor application notes.

 

[Kajunbee]

Yes I understand but I wonder if it would work in line to neutral. Say mov gdt and then cap. to neutral. Do you think when the mov starts conducting the capacitor would absorb some of the energy until the gdt fires off. I guess what I'm getting at is if the capacitor could compensate for the slower reaction time of the gdt.

 

[treez]

Unless you show a substantiated reference for GDT + MOV series circuit application, I prefer to consider it as a useless design idea.

Thanks, one reason for it is as kindly given by Kajunbee in post #12.
That is, as the MOV wears out and starts conducting and heating up, the GDT stops this conduction, and so the lifetime of the MOV is increased.
Also, we in actual truth use this series GDT/MOV combination upstream of the mains input fuse…….so it can’t actually crowbar the fuse when it flashes over. This is good because when the fuse blows the product obviously no longer works.
Downstream of the fuse, we have the mains rectifier, and then a 400V TVS which sits behind a current clamp…and the load (the led driver) is across the downstream TVS.
Our product is very small, and the GDT/MOV combination is the best we can fit into the tiny PCB space…..and also they are both Surface mount components, which fits with our all_surface_mount PCB.
Since the series GDT/MOV combination is upstream of the mains input fuse, we want it to fail open circuit. By putting the two transient protectors in series, we are almost guaranteed that at least one of them will fail open circuit. It would be extremely unlikely for both of them to fail short circuit at the same time. This is good because we don’t want a short across the mains as the supply to the entire installation would go down.
Also, consider using a single MOV or TVS…..in surface mount, you cannot get any higher than 440V flashover capability. That’s the SMCJ440A. You cant get a surface mount TVS with higher voltage rating than that. Surface mount MOVs are virtually non-existent….the Panasonic one stated here is virtually the only one in the world……so, basically, we need some front-end, surface mount transient protection with voltage flashover rating at some 700V……and the combination of GDT and MOV shown are the only things available for this. Two lower voltage TVS’s cannot be placed in series, neither can two MOVs be placed in series to get a higher overall voltage rating….but a TVS or a MOV can be placed in series with a GDT. -That's how we achieve the higher voltage rating for our front-end transient protection.
Do you agree?

By the way, is the ERZVF2M271 a TVS or a MOV?...its datasheet just calls it a "surge absorber"..

ERZVF2M271 DATASHEET:
https://industrial.panasonic.com/cdbs/www-data/pdf/AWA0000/AWA0000C4.pdf

 

[betwixt]

I'm not sure I agree with your design but the device uses ZnO material so it is an MOV. TVS works more like a pair of Zener diodes back-to-back in series so it would more likely be using Silicon as main material.

I feel the missing link in your design is something to absorb the transient rather than trying to clamp the power station output.
Brian.

 

[treez]

I feel the missing link in your design is something to absorb the transient rather than trying to clamp the power station output.

Thanks, do you mean that we need something like this to add in?...ie, the first post of this...

https://www.edaboard.com/showthread.php?t=372892&p=1597190#post1597190

 

[betwixt]

No, I have already replied to that thread.

What I mean is this kind of protection needs two steps, for sake of argument imagine the transient has unlimited power, placing anything across it that clips that power will destroy it. Obviously, in real life the power in transients is very small but the principle still applies. What you are currently proposing is a clipping circuit directly across the incoming power lines so think about where the transient energy will go. All the energy has to go into the clipping components.

If you place an impedance before the clipper, the energy is then dissipated in that instead. It has the difference between the transient voltage and the clipped voltage across it. Given the nature of your lighting designs, it would make sense to use an inductor to do that so it can convert the spike to a magnetic field and also act as an EMC filter. If ringing in the inductor itself worries you, shunt it with a resistor to reduce it's Q.

Even a resistor before the clipper will work. It is quite common practice to use a fusible resistor as well as, or instead of a traditional fuse. The resistor then converts the transient to heat but given the brief period you can probably still use a physically small one. It has the benefit of acting as an inrush limiter as well. Given the low current your product draws, you can afford to use a relatively high resistance, certainly many Ohms.

Brian.

 

[mtwieg]

What makes GDTs very useful is their foldback/avalanche behavior. This enables them to clamp much higher current impulses than MOVs or TVSs of the same power rating. Putting a MOV (or really anything with a substantial resistance or voltage drop) in series completely nullifies this advantage.

This advantage relies on the assumption that the surge source has some impedance, which will then dissipate most of the energy. This impedance may come from other components in your circuit, or from the wiring itself.

 

[treez]

The series GDT/MOV makes no sense.

According to the following , putting a GDT in series with a MOV makes sense because the MOV has no voltage across it at first, because the higher impedance GDT takes all the normal line voltage......then when the GDT breaks over the capacitor across the MOV gets advantageously charged up from zero by the transient , and thus absorbs the transient.

https://electronics.stackexchange.com/questions/220937/gdt-in-series-with-varistor

This advantage relies on the assumption that the surge source has some impedance,

The surge source is the energy in the stray inductance of the mains wiring......there is much energy stored in such wiring after say a high power piece of equipment has had a short circuit, and then blown a fuse......the fault current puts 100's of amps in the stray inductance of the mains, and this energy is what the mov will absorb..since one cannot break an inductive current...as you know, it is pointless to make the input filter resistive enough to be able to dissipate the surge current, -if it was that resistive it would be too lossy....the mov/gdt/tvs has to quench virtually all of it.

 

[FvM]

One more unrelated link, not referring to mains power line protection, mostly speculative, without any specification or application information...