Surge/Spike Spec vs TVSS Spec. 1

[daKlone]

Can anyone explain how I can correlate the surge/spike specifications I've been given to a TVSS spec?

I have a spike spec like:spikes shall not exceed +90V and/or -60V from a steady-state voltage of 26.4V.

But TVSS spec's say things like 8x20uS:160 Joules.

How do I convert my spike spec into an equivilent TVSS spec?

Thanks in advance,
daKlone

 

[jbartos]

Suggestions:
1. Visit

http://www.itvss.com/techfaq.asp

http://www.itvss.com/subpage.asp?pageid=5

http://www.itvss.com/subpage.asp?pageid=26

http://www.gensemi.com/product/AppNotes/appnotes.htm

http://www.gensemi.com/appnotespdf/tvs_sche.pdf

http://www.gensemi.com/appnotespdf/quik102.pdf

http://www.gensemi.com/appnotespdf/auto_tvs.pdf

http://www.gensemi.com/appnotespdf/quik104.pdf

http://www.littelfuse.com/ASP/news/Detail.asp?id=137

etc. for more information
2. There is more info needed. The TVSS is rated in Joules=Wattseconds. 160Joules=160Wattseconds. If your voltage spike produces Joules = Integral from 0 to t in seconds of (current x volts) x dt of energy which is less than 160 Joules then the specified TVSS is adequate. Some design margin is prudent to incorporate depending on a repetition of spikes in time. E.g. Joules = 100 spikes x integral from 0 to 0.00002 seconds of (90 volts x 10 amps) dt = 1.8 Joules, which leaves good margin to 160Joules. However, if there happen to be 10000 spikes in succession, then resulting 180Joules will exceed the rating of your specified TVSS. This may end up in a thermal runaway. Therefore, one must have a good idea about number of spikes in the succession and TVSS ratings. Continuous spikes must then have a TVSS sufficiently sized so that the heat dissipation takes place. The above calculation is approximate, only for illustration. When a number of spikes is unpredictable, then the selected or estimated TVSS may burn, which happens.

 

[daKlone]

Thanks.

Just so I'm clear on this, if I have a spike of +90V into an Rin of 0.28R, then I have a spike current of about 320A.
If I assume the spike is an 'applied voltage'(K=1.0) with a peak current of (I=)320A lasting (t=)50mS. Then from the info from general semi, if I use a TVSS with a clamp voltage of (Vc=)32V the spike energy (E) is;

E = K x Vc x I x t
= 1.0 x 32 x 320 x 50e-3
= 512 Joules

Correct?

If I have 5 pulses in sucession I therefore need at least 2.6kJ?

How do I know if a device that only states 8/20uS or 10/1000uS ratings is suitable?

 

[jbartos]

Suggestion to the previous posting marked ///\\\:
Correct? ///Yes. At least approximately.\\If I have 5 pulses in sucession I therefore need at least 2.6kJ?
///Yes. There should also be some time interval after those 5 spikes so that the TVSS cools off. Some design margin is prudent to incorporate. The bigger it is, the better it will be.\\How do I know if a device that only states 8/20uS or 10/1000uS ratings is suitable?
///This is not adequate information. Normally, information on the device will give the some additional info, e.g. Joules, or Amps and Voltage. Check the offerings on web sites.\\\