Regarding High Current Tripping

[sabu31]

Dear All,

I need to use a tripping mechanism for AC current typically in range of 150 to 200A. I tried one SSR which had rating of 200A. However, even when around 120A flows through it, the SSR heat sink (along with fan) temperature rises to around 70Degree Celsisus.
Are there any SSR for use in continuous operation in this current range. Or can we parallel the SSR for the same. Also I would like to know if there any triacs which have similar rating. I checked mouse, but could find triace up to 88A.

What are the other low cost options for achieving this. Thanking you.

 

[KlausST]

Hi,

However, even when around 120A flows through it, the SSR heat sink (along with fan) temperature rises to around 70Degree Celsisus.

I guess the datasheet gives informations what power dissipation to expect. And then it´s on you to calculate and use a suitable heatsink.
Another guess: the SSR works with Triac / SCR thus the power dissipation calcualtion is rather complicated when the current waveform is non sinusoidal.
And ... paralleling of Triacs / SCRs does not reduce power disspation that much (compared to transistors)
And ... a Triac / SCR can not be switched OFF immediately ... it stays ON until zero cross of current.

And there we are... specifiactions and requirements:
* What is your operating voltage?
* How fast do you need to switch OFF?
* What is the load?
* any other information we need to know?
* what does "low cost" mean for you?

As first idea .. I´d use a relay.

Mosfets or IGBTs could be a benefit ... depening on the requirements and specifications...

Klaus

 

[cupoftea]

Yes, breaking such high currents is problematic for a MOSFET, (hence MOSFETs are still not "de rigeur" in BLDC drives) even paralleled MOSFETs.....it is more IGBT territory, and Triacs etc. But yes, a relay with a snubber is also good.

In for example , high power BLDC drives, MOSFETS get little use since they cant break the short circuit current. (they just pop and die).
MOSFETs just cant handle the linear region of overlapping voltage and current that they get when breaking a very high current and reasonably high voltage aswell. They just pop.
Hence the reason that MOSFETs arent used in very high power trailing edge heating controllers.

Even IGBTs struggle , but if a decent RC snubber is used aswell, then you can be OK with an IGBT to break the high current. -Switch the IGBT off as fast as poss and have a snubber RC to take the inductive spike that you'd otherwise get.

 

[Easy peasy]

A parallel combo of SiC devices and IGBT's is used to break higher currents - and to carry them.

we have designed electronic fuses along similar lines.

 

[cupoftea]

...yes and i know you are saying SiC due to the high voltage/low rdson requirement...but ayk, a normal Si FET is OK when the voltages are lower..
Ayk, SiC is slightly more delicate than standard Si FET....but with the IGBTs with it and your other bells and whistles, it will of course be fine.
Even then, for breaking super high current at high volts you still need a RC snubber, and basically some kind of RCD clamp type thing.

 

[Easy peasy]

but ayk, a normal Si FET is OK when the voltages are lower.

have you testing this much ? our experience is that the IGBT // SiC combo is 10 x more resilient than a Si mosfet.

Ayk, SiC is slightly more delicate than standard Si FET

Hardly - the 1200V 50A SiC actually have 3uS short ckt ratings - no such luck for a Si mosfet => poof

 

[cupoftea]

Thanks, yes i noticed this has a high single pulse power dissipation capability

https://www.onsemi.com/download/data-sheet/pdf/nthl040n120m3s-d.pdf

Though with a combo of IGBT/SiC, ayk, the SiC would switch off, and then the short circuit current be instantly diverted to the IGBT, which would not be switched off so quick.......then the IGBT takes all the grief of the linear region when it switches off.
So whether Si or SiC, it would be nicely cocooned by the IGBT.

SiC FETs are AKA "SiC Schottky" FETs. And Schottkys are well known for being much more delicate than either UF or Gen Purp diodes. So maybe its the internal diode of the SiC thats more delicate than the Si FET i believe.

 

[Easy peasy]

SiC FETs are AKA "SiC Schottky" FETs. And Schottkys are well known for being much more delicate than either UF or Gen Purp diodes. So maybe its the internal diode of the SiC thats more delicate than the Si FET i believe.

Where do you get this mumbo jumbo from ?

 

[FvM]

Confusing SiC diodes (Schottky) with SiC MOSFET...

 

[cupoftea]

I will double check, but im fairly sure that the internal diode in a SiC FET is a Schottky, or rather, it doesnt suffer reverse recovery as badly as the internal diode in an Si FET.
yes i thought so, as follows...

Body Diode Characteristics | What are SiC-MOSFETs? – SiC-MOSFET Features | TechWeb

The last time, we explained differences with IGBTs. This time, we will discuss

techweb.rohm.com

This is why the SiC FET is the friend of the LLC designer, and in many ways, the Phase Shift Full Bridge aswell.
As is known there are scores of docs on the web about reverse recovery failures in both PSFB and LLC.

 

[FvM]

Looking closely at the above posted SiC MOSFET structure clarifies that the body diode is a regular PN diode. No Schottky junction involved.

 

[mtwieg]

A parallel combo of SiC devices and IGBT's is used to break higher currents - and to carry them.

we have designed electronic fuses along similar lines.

Just curious, I'm guessing you control the devices such that the SiC always closes first and opens last? What sort of delay is appropriate?

 

[cupoftea]

Just in Case EP doesnt answer, i guess when breaking a short cct current , you have to open the SiC first, then just a few us later, open the IGBT.
Though one would rather wait longer, and let the IGBT warm up a bit before switching it off and subjecting it to the super Megawatts of the overlap region.
Sudden expansion being very damaging.

 

[Easy peasy]

obviously - they can close at the same time - IGBT's actually turn on very fast, at turn off, only 500nS is needed after SiC turn off for IGBT turn off.

 

[sabu31]

Hi all,

Sorry for the late reply.
Based on my experimentations, the data sheet mentioned for 200A rating was not suitable for even 120A with heatsink provided by manufacturer. I had to use bigger size heatsink and high speed heatsink fan to operate at around 140A. Still here the heatsink temperature rose to 60 Degree Celisus after continuous operation.

I found from data sheet of SSR manufacturer ATO, that even for resistive current, the load current can be up to 60% of rated SCR for reliable operation. So typically for 140-150A operation, we need SSR of 400-500A if load is inductive (around 0.8pF).

The other alternate is to use two thyristors and connect them in a triac formation, however, the costing goes up to 100 Dollar per module (two scrs). Required cost is around 50-60 Dollars per SSR (rated around 400-500A)

Is there any single-phase contact with such high ratings as 150A and lesser cost than SSR/(two SCR combination)?


Thanks

 

[KlausST]

the data sheet mentioned for 200A

What device and what datasheet are you talking about?

Usually one can rely on datasheet infromations .. so without more details (device, datasheet, waveform, gate drive voltage, heatsink specifications...)
we are not able to give detailed assistence.

I can only guess ... you did something wrong....

even for resistive current

You mean resistive load?
--> It does not matter whether the load current is resistive, capacitive, inducitve or mixed ... all that counts is the RMS value (for IGBT and MOSFET).
(For SCR and triac you need to consider RMS current as well as average current to calculate power disspation.)

if load is inductive (around 0.8pF).

0.8pF is nothing when we talk about magnitude.
And pF means "capacitive", so not inductive at all.

--> I´m really confused by your informations. They make no sense to me.

Klaus

 

[Easy peasy]

Surely you had done the rough thermal calculations before hand ?

say 1.5V drop in the Triac @ 120A rms = 180 watts - so obviously a very large heatsink and fan needed....

good SCR's cost money - you cannot make a silk purse SSR out of a sows ear.

--- Updated Nov 3, 2024 ---

Let us look at this device from RS components in India: ( 5205 rupee = 62 USD ) 160A capable 1200V

Vishay VS-VSKT162/12PBF, Dual Thyristor Module 1200V, 160A 270mA | Vishay | RS Components India

VS-VSKT162/12PBF Vishay | Vishay VS-VSKT162/12PBF, Dual Thyristor Module 1200V, 160A 270mA | 395-8384 | RS Components

in.rsdelivers.com

you can link this package for a back to back ( inverse parallel ) AC switch

from the data sheet we see: 1.54 volt @ 160A device and just over 200W dissipation for 180 deg conduction from the graphs ( top right of page 8 ) - so over 400W total package - and you HAVE to keep the case below 85 deg C !

So a real heatsink and fan needed for this as an AC switch - no matter how you do it with SCR's !

 

[KlausST]

say 1.5V drop in the Triac @ 120A rms = 180 watts

accurate power dissipation for SCRs work differently.

beccause SCRs have a
* constant voltage part that needs to be multiplied with average_current
* resistive voltage part that needs to be multiplied with RMS_current

For an over the thumb esitmation it still gives a good clue.

Klaus

 

[Easy peasy]

Here you go, single phase SSR, 350A ( 400 on request ) USD $ 49.00 - voila !

https://www.amazon.com/Industrial-Control-Petrochemical-Equipment-MGR-H3350Z/dp/B08NT6S2NH/ref=sr_1_3?crid=26DN13NWBFI5E&dib=eyJ2IjoiMSJ9.Ljbu_CwaUo05UXhJP-u-Q2ksHUNcgZKUN-i8d4wmzpUJ7tupccZpN7L_eKIk_Km2yd705goFDVEcMmR6JNj6yzs9VXiP2YcUCwzlOHp21HkhTkyslyW4-zuTKCs5YcXWnbZPy_1B_OM1WrHmfp4ECzsAxta8t_6yIR5FdUlelt_WU32jnkWPiqmJpfnkOMbpE6uysFjkYTh4wket1O3xtlWSRi0kMcoVVjUxPowQHMw.1a9HgOWjgjI9Gtpm-J3PvzgagE2VltP0quo_kCq5dyo&dib_tag=se&keywords=Solid+State+Relay+400A&qid=1730595354&sprefix=solid+state+relay+400a%2Caps%2C298&sr=8-3

just need to add freight, heatsink, fan, fan psu ... and a wee ckt to advise you the h/sink is over temp - i.e. fan fail !

[ just remember - you always get just that quality that you pay for ]

and that 400A x 1.5V ( approx ) = 600 watts to dissipate ( for sine shaped current ).

 

[sabu31]

What device and what datasheet are you talking about?

Usually one can rely on datasheet infromations .. so without more details (device, datasheet, waveform, gate drive voltage, heatsink specifications...)
we are not able to give detailed assistence.

I can only guess ... you did something wrong....


You mean resistive load?
--> It does not matter whether the load current is resistive, capacitive, inducitve or mixed ... all that counts is the RMS value (for IGBT and MOSFET).
(For SCR and triac you need to consider RMS current as well as average current to calculate power disspation.)


0.8pF is nothing when we talk about magnitude.
And pF means "capacitive", so not inductive at all.

--> I´m really confused by your informations. They make no sense to me.

Klaus

Hi KlausST,

I am sorry, i posted quickly without sufficient information. The datasheet of SSR being used is attached. I tested with 200A model along with heatsink provided. However, as mentioned in my post, with heatsink provided, I could not continuously operate the SSR for 140A for more than 3-4minutes. I changed heatsink (27cm (L) *12.5cm (W)*14cm(H) and along with two fans rated at 18W, the it reached around 65 degrees Celsius after operating for 1hour.


There is a datasheet by ATO that mentions relative possible load current with respect to SSR current (page six of solid-state-relay-catalog).

By 0.8pF, (it was a typo), I meant 0.8 Power Factor lagging which is generally the load in a distribution transformer.