Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronic Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Register Log in

32kW peak power in inrush NTC

treez

Advanced Member level 5
Joined
Sep 22, 2008
Messages
7,677
Helped
568
Reputation
1,139
Reaction score
543
Trophy points
1,393
Location
cambridge
Activity points
76,860
Hi,
We are using the B57234S0509M NTC.
Its for inrush into a 190-265VAC offline flyback SMPS with an input electrolytic capacitor of 150uF.
The schem and inrush power waveform is as attached (this is with mains applied at mains peak and when VAC = 265VAC. (Also, LTspice schem is as attached).
Do you think this NTC can handle this ….how many times?
The datasheet gives no indication of i^2.t or allowable power vs time waveforms.

The 3R inrush NTC energy is 16Joules, and this is over 500us. Datasheet doesn't verify this.

If ambient is 40degc at the time of switch ON, then there is a 33kW peak power pulse (as shown). I strongly doubt that this little NTC can handle this without some damage?

B57234S0509M NTC datasheet
 

Attachments


barry

Advanced Member level 5
Joined
Mar 31, 2005
Messages
5,038
Helped
1,104
Reputation
2,218
Reaction score
1,089
Trophy points
1,393
Location
California, USA
Activity points
27,598
Instead of asking a bunch of strangers on the internet their opinion about whether or not a device will fail, why don’t you, oh, I don’t know, ask the manufacturer?

I presume you are building some product that you are going to try and sell to the unwitting public. I, for one, would not feel comfortable buying a product that was verified by some anonymous party with unknown credentials.
 
  • Like
Reactions: treez

    treez

    points: 2
    Helpful Answer Positive Rating

dick_freebird

Advanced Member level 5
Joined
Mar 4, 2008
Messages
7,140
Helped
2,091
Reputation
4,186
Reaction score
1,938
Trophy points
1,393
Location
USA
Activity points
57,287
From datasheet RATINGS AND CHARACTERISTICS

Heat capacity Cth [mJ/K] approx. 1530

So 16J of dissipated inrush energy ought to
produce a temp rise of about 11 degK.

There is a table column for "maximum
switchable energy" (in units of capacitance)
with an explanation and test figure later on
in the document. It indicates that your 150uF
is OK for 230VAC line.
 

treez

Advanced Member level 5
Joined
Sep 22, 2008
Messages
7,677
Helped
568
Reputation
1,139
Reaction score
543
Trophy points
1,393
Location
cambridge
Activity points
76,860
Thanks, its interesting the "maximum switchable capacity" test is only done 1000 times. Which makes you wonder if it does some kind of damage to the NTC, partially wearing it out? 11K is not a big temperature rise, but being smashed up by 11K in 500us sounds like a potential problem for sudden expansion type damage, which may be very slight , but cumulative.
Its disappointing that they put the 1R in there in the test jig, because it means any "sudden high voltage effects" are not shown up as much.
Theres no need for the 1R, since they wait for the NTC to cool down before reapplying the voltage.
Also, we would like the test to be done at 265VAC peak, ie, 373v.

In diagnosing failed SMPS, its often noticeable how the NTC has blown, even though it was "correctly rated for the job".
 
Last edited:

dick_freebird

Advanced Member level 5
Joined
Mar 4, 2008
Messages
7,140
Helped
2,091
Reputation
4,186
Reaction score
1,938
Trophy points
1,393
Location
USA
Activity points
57,287
1K cycles I expect is just some reliability engineer's
guess at what the customers would consider
adequate. Which maybe it is not, for a frequently
cycled power supply. You'd perhaps like to drill into
the vendor reliability data and see if they had any
failure population at end-of-test. But still it won't
give you fittable life vs stress curves unless you
go full test-to-fail (or test-to-N%-fail at least, N
being substantial enough for statistical happiness).
 

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Top