power handling of series pass element (mosfet/transistors/bJt's etc )

[eevblogfan]

hello


I'd like to know what is the limit for each pakeg exsist ,

to-220

to 247

to 3

to 18

etc ..

just quik refrence (IE , to 220 can handle X rating of power for ussual devise )  withuot burn !

I think knowing that key will prevent many of us to burn some pricy elements

thank you in advance !

[T4P]

There's no universal standard, it's different for every manufacture and even every device (YIKES !)
But generally a TO-3P/TO-247 is 50-300W ( That's mostly for BJT's  ) Some mosfets can go up to crazier ratings like 1.8kW
Then we shall put 150W as the ceiling for normal power TO-3P BJT's
TO-3 mostly 90W-300W ( at least as BJT's go ) or 115W by ONSEMI and quite a few other mfrs for the 2N3055
but they depend on the thermal mass of the heatsink and the component
TO-220 mostly 30W MAX for most of the devices i see
TO-18 500mW to amb for 2N2222
TO-3 0.3W to ambient and 0.75W to case (BC108)
Best check a datasheet but these are the standard ones i use

[ejeffrey]

It isn't really needed or possible.  As DaveXRQ has shown, power rating varies quite a bit for the same package type.  Power rating is one of the top-line parameters in any data sheet and always available as a filter field in a parametric search.  There is no reason you need rules of thumb for given packages.

What neither the package type nor the front page number on the data sheet will tell you is how the part needs to be derated for operating conditions.  For instance, BJTs maximum temperature/power are reduced when operating with large voltages.  You need to dig into the data sheet and consider your specific circuit to get these more subtle issues.

[eevblogfan]

thank you for replay

I've read B4 some datasheets (like most of us )

I saw some crazy ratings , but I think that 1.8kW is true only for liquid hydrogen cooling XD ,

really , put it on CPU heat sink and I am quiet shure it will melt after passing the 150+W

what I meant is , with normal usage ( IE , 0.1 C/W to 0.03C/W or so ) heat sink , no more , what is the power you can put into the to-220 etc without lunch it to space

so many times I have manged to blow IRFP064N with 0.06C/W (or close) , just with 50W ( like 49C above 37 C ambient  or 86C junction temp , ) oh can 200W 175+ C blow at 86C and 50W load ?? what have I missed here ?

really trup for young players , I haven't really set and studied that case ,(have no time for some good reading + practising )

can you tell me where am I wrong in here ? (apart from my bad English )

I know that mosfet is design for switching app , but it rated 200W (at 25C of coarse ) and 0.99C/W-175 C junction temp !!! and damm , B4 eBay I bought them for really costy price ! 64$ for only 10 of them !!! (stupid tight ass country !!! all is costy in here ! )

[T4P]

thank you for replay

I've read B4 some datasheets (like most of us )

I saw some crazy ratings , but I think that 1.8kW is true only for liquid hydrogen cooling XD ,

really , put it on CPU heat sink and I am quiet shure it will melt after passing the 150+W

what I meant is , with normal usage ( IE , 0.1 C/W to 0.03C/W or so ) heat sink , no more , what is the power you can put into the to-220 etc without lunch it to space

so many times I have manged to blow IRFP064N with 0.06C/W (or close) , just with 50W ( like 49C above 37 C ambient  or 86C junction temp , ) oh can 200W 175+ C blow at 86C and 50W load ?? what have I missed here ?

really trup for young players , I haven't really set and studied that case ,(have no time for some good reading + practising )

can you tell me where am I wrong in here ? (apart from my bad English )

I know that mosfet is design for switching app , but it rated 200W (at 25C of coarse ) and 0.99C/W-175 C junction temp !!! and damm , B4 eBay I bought them for really costy price ! 64$ for only 10 of them !!! (stupid tight ass country !!! all is costy in here ! )

You have to see the derating curve AND the spikes it's not all the time the derating kills it

Plus a big CPU heatsink is okay with 150W, in fact most can stay below 40C with a fan of course

[eevblogfan]

thanks

BTW , If I'll put 2 CPU heatsinks and creat a tannel "[ ]" shaped and put some 80mm fan , can someone asimate the power handling per C above ambient (IE ,C/W rating ) ?

I plan to build 300W (or more ) power supply , and as for the heatsinking , I plan to use the CPU's HS as b&k did ( ep 281 in dave's channel )  it looks awsome

I plan to use 2 mosfet per heatsink (75W) and 2 heatsink-150W as one "[ ]" tannel , will it do ?

thank you in advance !

[T4P]

thanks

BTW , If I'll put 2 CPU heatsinks and creat a tannel "[ ]" shaped and put some 80mm fan , can someone asimate the power handling per C above ambient (IE ,C/W rating ) ?

I plan to build 300W (or more ) power supply , and as for the heatsinking , I plan to use the CPU's HS as b&k did ( ep 281 in dave's channel )  it looks awsome

I plan to use 2 mosfet per heatsink (75W) and 2 heatsink-150W as one "[ ]" tannel , will it do ?

thank you in advance !

The most important thing is that computer heatsinks are designed for the fan directly on the top
But it looks like mfrs do not say the thermal resistance of the heatsink at first glance so i have to probe around for datasheets

[SeanB]

The channel heatsinks will do around 60W, and if you place 2 to make a channel and use a fan at the one end you could probably do 4 TO220 devices and dissipate up to 200W, or 50W per device.

If you use a newer one with a central boss you will get one device per heatsink, but it will be able to handle up to 120W

The biggest variable is the mounting, to get the best you need to mount the tab to a machined flat surface, absolutely flat ( mirror finish if possible) with only a thin smear of thermal compound, and with the correct size machine screw that is tightened with the correct torque into the right size relieved tapped mounting hole. All these are variables, and affect the heat transfer ability.

[ejeffrey]

so many times I have manged to blow IRFP064N with 0.06C/W (or close) , just with 50W ( like 49C above 37 C ambient  or 86C junction temp , ) oh can 200W 175+ C blow at 86C and 50W load ?? what have I missed here

By "ambient" do you mean the air in the case or the temperature of the heat sink?  Also, are you sure the heat sink is in good thermal contact with the case?  The data sheet I am looking at specified a 0.75 K/W junction-to-case and 0.24 K/W case-to-heatsink.  That is all fine, but you can easily hit 1 K/W or worse on the case-to-heatsink if you aren't careful.  Too much thermal grease, too little pressure, and improperly machined mounting hole, or the wrong insulating pad can all easily cause the thermal resistance to go up.  So can mounting the transistor at the edge instead of the center of the heat sink.  Restricted airflow or inability to dissipate heat to the ultimate heat sink can also be problems.

Another possibility is oscillation, surges, or thermal instability of your circuit causing the power dissipation to be much greater than you expect.

If possible, measuring the temperature of the case directly.  Failing that, probe the heat sink as close as you can to the mounting point of the transistor.   A small surface mount diode or diode-connected transistor makes a good low mass temperature sensor that can respond fast enough to catch some forms of thermal runaway.

[T4P]

For a pretty high power project i made a few months ago involved AMD heatsinks ( you can get them on the cheap or even free, i used AMD because i get more area contact ) with fans all over, it involved 1 heatsink per mosfet each mosfet was dumping about 100W ( or so the CPU that went with the heatsink had as TDP )
and then a few central fans, i managed to keep temps down to 50celsius
Everything was mounted vertically with riser boards for the mosfets and clamp spots for the heatsinks
worked well with some cheap thermal grease, not the expensive silver based ones as that is electrically conductive

It didn't pan out cheap but it was a 600W load but not programmable, might add in the future but 6 heatsinks into a 1U rack was mission accomplished with much space to spare
OT: Mind you i kept the current low and the voltage high, i didn't want high amperage as that is pretty hard to get right and quite .... dangerous to accomplish
(12V 50amps MAX ... )

[eevblogfan]

I've cheked all I could (back then )

the heat sink is 17@12cm , the heat sink has 4 holes ( all centerd ) , the mosfet has well contact , assuming wors case of 1.5C/W (somehow ? )  then -> 1.5*50W=75C above ambient + the ambient (ie 37) = 112CjcT

the coad must be huge ammount of current and it even make sens ( 8mohm Rds-on) +10mohm shunt + say 7mohm wiers = 25 mohm -> at 12V it's max peak of like 480A! , umm yep , that must be the couse of big bang(s)  -_-

in data sheet it's max current peak is 390A at the "best" condition , leasson ?  put in series a good 3-5W , 0.24ohm ( or 2X 0.1 ohm ) and the "max" peak current should be "safe" >= to 50A (at 12V ! ) , or use the 0.24ohm as a shunt ( of coarse , 1shunt per mosfet will give better regulation + multiple shunts give lower series resistance etc )

anyway , when I'll be able to . I'll chek the C/W of the "[ ]" shaped heat sink areggment , I hope for some good results

thank you in advance !