MOSFET derating calculation with temperature

[Rajnaveen]

Hi friends;
I am using MOSFET IR840 in a solar charge controller . I have placed a heatsink for cooling purpose . when charger run continuously for 4-5 hour heatsink temp raise to 70 deg at 30 deg ambient . I question is that how could i calculate the derating of mosfet ? so that i can assure that the mosfet i am using will work for 1 year or 2.
please suggest ur views.

 

[kappa_am]

there is a curve in the datasheet current versus temp. I looked you can use it to handle 6A.

 

[andre_luis]

Also must be considered at what conduction operation device is subjected, continuous or pulsed. This reduces the maximum current allowed. That information can be observed at the "Safe Operating Area" graph on datasheet.

 

[Rajnaveen]

it is operated in pulsed mode . actually i want to know that if MOSFET is operating at 70 deg heatsink temp will it derate or leak the mosfet after some time so next time the mosfet could burn in SOA region?

 

[D.A.(Tony)Stewart]

Heatsink rise of 40C is based on what power loss?

Absolute max. is 125 deg C for the junction and I expect heatsink to junction resistance to be 1.5 ~ 2 deg C/W above heatsink.

You could limit Tj to 100deg C max or a Tj rise of 70 deg C.
, based on your power dissipation.

70C rise{Tj}max=40C rise {Tsink} + Watts*2deg/W {Rj-c-hs}

Solving above, if your power exceeds (correction) 15 Watts , Fix you heat sink or better yet , get a better MOSFET with 10x lower RdsOn like 85 mOhm instead of 0.85 Ohm .


Pulsing a PV array to charge a battery is not ideal efficiency for a PV, but more efficient for the switch to dump 6A across low voltage than high in linear mode with less current. Using a steel core choke such as a Microwave oven transformer in series with switch (protected with diode) will give you continuous voltage loading on PV where 75% of no load voltage is close to ideal power transfer to battery, then modulate Duty cycle of pulse or PWM to adjust charge to battery limited by optimal CC limit, CV limit and float voltage limit. Now you have a "real" solar panel battery charger using SMPS switching of MOSFET, optimal PV power loading, optimal battery charging and optimal power transfer to load at minimum temperature rise to MOSFET!!!

A junction temp of 100'C is marginal for short life, not ideal. Tj<85DegC at worst case ambient is a commercial std. Stick to best practises please. Every 10degC rise above 25'C for Tjcn results in 50 % less life (Arhennius law)

 

[kappa_am]

In my opinion use the continuous ID. Since pulse mode means that the switch has enough rest time after duty cycle. I fell is more safe to choose continuous ID however it looks somewhat indulging.

 

[yassin.kraouch]

in page 5 of the datasheet you can find the current VS case temperature, you can use this as a refernce
https://www.vishay.com/docs/91070/91070.pdf

 

[D.A.(Tony)Stewart]

Power in MOSFET is I²R where R=0.85Ω The RdsOn spec for your part. The curve on page 5, using this indicates 2.2°C/W derating of case temp with A vs Tc curve. This is close to my estimate of 1.5~2°C/W. Get a better MOSFET.

- - - Updated - - -

In my opinion use the continuous ID. Since pulse mode means that the switch has enough rest time after duty cycle. I fell is more safe to choose continuous ID however it looks somewhat indulging.

Kappa_am< on linear vs switched current, compare LDO vs Buck regulator, which will drop more power? using duty cycle of pulse for average power vs continuous I²Rds or (Vpv-Vbat)²/Rds? The PV panel might appear to have an effective source impedance of W/V² but actually not. It is the change in max power with the change in max voltage, V² that gives the ESR of the PV. With max power transfer theorem, matching battery lower impedance to PV with a Buck regulator at same ESR gives the maximum power loading of PV panel for a max. battery charge rate. Then using a Buck switch & high current low loss choke (e.g. MOT or microwave oven transformer ) can then give more current without saturating. This is not the only way!

 

[Rajnaveen]

Sunnyskyguy ; Is is necessary to match impedance ? Because while designing buck converter we did not do any calculation for impedance matching . If it is necessary how to do this?