Ok, so new to this
Been busy with reengineering of a LED-LCD monitor - largest Ive tried (28 in) for a gas-tight requirement. Still more art than engineering on thermal given taking OPD ( other peoples design ) and reengineering packaging n power with regard to thermal but added temperature measurement to uLAN handling button functions. Repackaging seems good, crazy external snafu caused reseal ( currently underway ) but all working. As in this case, Usually end up throwing away plastic packaging ( bezels n covers w_ vents etc and add new mounts etc. Now with LCD & esp LED-LCD and cexternal WORTS able to go passive cooling ( sweet power n quiet ) using aluminum.and no fins ( of course the large package surface and vertical orientation to lower power dissipation really helps!
In past, have successfully reengineered a 12 in & 17 in CRT ( had to use WATER COOLING on 17in but then actine internal fan and passive fins on a re- re engineering ) Oh yeah water next to high voltages !! Used hall effect n 'magic pencil' for controls to maintain sealing and very expensive military circular connectors.
Then DVI and better graphics boards allowed using a sealed a CCFL 17 in LCD, It was hot - but after testing I went passive internal n external with no fins and still working many hundreds hours later ( still use it as my lab 'beater' CRT. On this version, I used touch buttins and reengineered existing connectors, removing some ( now very easy to use touch controls of course )
Then LED LCDs made much easier thermal design. Also used uLAN controls instead of buttons and FO for front indicator and existing connectors like third gen re-engineered technique for seal ( but of course not drip proof like CPC or milspec ) Did four by two different mfrs and very hard to tell by even back panel/ electronis buldge and unless OSD no tell on front.
SO what Im doing - as soon as i wont get frostbite Im cutting metal forsealed RPi and USB Hub for USB Stick NAS.- Designed n going try new button/ indicators for USB / NAS control.
Later