I want to estimate the power dissipation of this IGBT SGL160N60UFD but I can't find any Rdson, I leave the datasheet:
can this TO-264 package handle currents of 100A?
\$\begingroup\$
\$\endgroup\$
2
-
2\$\begingroup\$ IGBTs have a saturation voltage like BJTs, not an Rdson. \$\endgroup\$– HearthMar 31, 2021 at 13:54
-
1\$\begingroup\$ There is no RDSon on IGBT, they have a saturation VCE like bipolar transistors. In fact that's their strong point \$\endgroup\$– Lorenzo MarcantonioMar 31, 2021 at 14:14
Add a comment
|
4 Answers
\$\begingroup\$
\$\endgroup\$
There is no Rds(on) on an IGBT because it does not behave like a MOSFET- it has a Vce(sat).
According to the SOA curve, 100A continuous is not safe even at Tc = 25°C. The limitation is thermal. Note that the Vce(sat) is not guaranteed above 80A so there is uncertainty in the maximum power dissipation at 100A.
300A is the upper limit for a single short pulse starting off at 25°C.
answered Mar 31, 2021 at 14:40
\$\begingroup\$
\$\endgroup\$
I want to estimate the power dissipation of this IGBT SGL160N60UFD but I can't find any Rdson
Well, it's an IGBT and it doesn't have a drain (d) or source (s) pin. They are actually called collector and emitter like an ordinary BJT. It does have a gate pin (like a MOSFET) of course.
can this TO-264 package handle currents of 100A?
From the mouser data sheet there is this graph: -
So, clearly it can handle 100 amps (if kept cool) but, the power dissipation will be about 250 watts with a gate-emitter voltage of 10 volts. However, it's a simple jump to convert 2.5 volts and 100 amps to an equivalent on-resistance of 2.5/100 = 25 milli ohms. But don’t run it at this level of current without substantial heat sinking and, not continuously.
answered Mar 31, 2021 at 14:27
\$\begingroup\$
\$\endgroup\$
This part is marked obsolete on D-K site but not Mouser(?), so no complete datasheet in your link.
But consider for all saturated switches;
$$\dfrac{V_{ce(sat)}}{I_{c(sat)}}=R_{ce}$$
1.3V/80A typ = 16.25 mOhm. 1.6V max. =20 mOhm Max.
1.6V/160A typ = 10 mOhm typ. Not max.
Lookup plot for SOA with PW50 duration @ 100A and consider margin for this “typical” plot.
https://www.mouser.com/datasheet/2/149/fairchild%20semiconductor_sgl160n60ufd-320888.pdf
Always search major distributors instead of others... for datasheets.
Then add 15nH 5mm from device with C values.
\$\begingroup\$
\$\endgroup\$
6
The way you model a minority carrier device (BJT, Diode, SCR, IGBT) is via the output characteristic transfer function.
\$V_{ce} = V_{ce0}+ R_0\cdot I_c\$
\$V_{ce0}\$ = on-state voltage threshold \$R_0\$ = On-state resistance approximation
These figures are not provided in the datasheet but can be derived from the Ic vs Vce curve, around an operating point. Once you have derived these figures, the loss can be calculated fairly accurately
In this example I am working around 80A and thus the tangential straight line clips 80A and intersects the x-axis at the approximate \$V_{ce0}\$ ~ 1.4V The ΔI and ΔVce around this point is then used to approximate \$R_0\$ ΔI ~ 80A, ΔV ~0.7V => \$R_0 \approx\$ 8.7mR
Thus the approximate transfer function, around 80A is \$V_{ce} = 1.4V + 8.7m\Omega\cdot I_c\$
Now whether this device should be conducting 80-100A? that is 160-170W, which is only viable IF the die is kept at 25C, which it won't be
-
\$\begingroup\$ OP's concern is power dissipation, and r = ΔV/ΔI aka dynamic resistance relates to small-signal analysis (also, the green slope line crosses the Vce-axis at about 1.4 V, but it's a moot point). \$\endgroup\$– arielCoApr 1, 2021 at 0:20
-
1\$\begingroup\$ @arielCo in this instance it doesn't. This is the accepted way to determine an associated Ron to determine power loss. This is covered in whitepapers from: IR, OnSemi, ST, Infinion, Dynex as part of their power electronics section \$\endgroup\$– user16222Apr 1, 2021 at 6:23
-
\$\begingroup\$ How would you get the 160W that the device is actually dissipating back from that 8.7mohm and 80A? \$\endgroup\$– arielCoApr 1, 2021 at 10:12
-
\$\begingroup\$ who is talking about 160W? it is dissipating closer to 230-250W IF conducting 80A (no switching) \$\endgroup\$– user16222Apr 1, 2021 at 10:27
-
\$\begingroup\$ You picked Ic=80A, right? The drop at that point is Vce≈2V i.imgur.com/JuY8spW.png \$\endgroup\$– arielCoApr 1, 2021 at 11:15
