Body
Hello Ole, Warren, the big heatsinks with the stronger diodes work wonderfully, the coolers are lukewarm. The 350 amp module does not feel warm at all. now I still need suitable reactor inductors. Someone an idea? Greetings Sven
Von Samsung Mobile gesendet
-------- Ursprüngliche Nachricht --------
Von: "[email protected] [EVGRAY]" <[email protected]>
Datum:21.01.2018 16:39 (GMT+01:00)
An: [email protected]
Betreff: Re: AW: [EVGRAY] Re: buck plug ...
Hi Warren,
I think this plan could work well. One thing against it is the higher dielectric constant of oil. But this will only matter at fast transients or high frequencies and weak signals.
There are commercial water cooling blocks totally closed systems for CPUs.Here is someone using them for coooling LEDs 'Water Cooled 72,000 Lumen LED Flashlight!!' (4:44 minutes):
https://www.youtube.com/watch?v=-vgNh3fLxJc
Regards
Ole
---In [email protected], <schoonersolsticemoon@...> wrote :
Ole
Is it a crazy idea to take the entire circuit board and put it in a large container of clear mineral oil?
What components would be harmed doing this?
A pump could circulate the fluid around the board and carry it to cooling fins on the outside of the tank keeping all temperatures close to ambiant room temperature , or more actively, pass the oil through a heat exchanger, where it could be refrigerated, if need be, using a small heat pump using a juice can sized compressor.
The big issue is to use components that don't dissolve in oil, which few seem to, unless wax based.
You are highly unlikely to boil that tub of oil with your components, and the heat transfer should be pretty good.
Cheers Warren
Sent from Yahoo Mail on Android
On Wed, Jan 17, 2018 at 7:12 PM, onielsen@... [EVGRAY]
<[email protected]> wrote:
Hi Sven,
The 150 degrees Celsius is the max temperature inside the device (the temperature of the silicon crystal). When cooling the case to 20 degrees Celsius then the max current is what heats the crystal to the 150 degrees. If cooling with a Peltier element or some other kind of refrigeration system the max current can be increased without meltdown. Some people put their computer motherboard in a refrigerant to perhaps double the clock speed without frying the CPU. The limit for removing the heat is the thermal resistivity of the device. The thermal resistivity is given in the datasheets. There may be given values of thermal resistivity (Rth) from junction to case and from case to heat sink or from case to ambient (when using no heat sink). Those values are used to calculate the max power dissipation at different ambient temperatures. The higher the thermal resistivity is the more increase in temperature during a given power dissipation. A good heat conducting heat sink having low thermal resistivity between the case and ambient is what is needed to remove the dissipated power. The heat conduction is given as K/W (Kelvin per Watt. Say if the heat sink is rated 2K/W then the device case will increase its temperature by 2K (or 2 degree Celsius) for each Watt being dissipated in the device. The max temperature given in the data sheet mustn't be exceeded if the device is to survive.
Heatsink Calculations: https://www.re-innovation.co.uk/docs/heatsink-calculations/
How to Select a Suitable Heat Sink: https://www.designworldonline.com/how-to-select-a-suitable-heat-sink/
Regards
Ole
---In [email protected], <s.friedrich@...> wrote :
Hello Ole, in the case of the large modules, the data are also given at 150 degrees Celsius. The fact is that the more powerful modules generate less heat. I have removed and used aluminum heatsinks from various defective converters. I do not want to install expensive water cooling as long as I can get hi power module cheaper. When I think about how small the igbts are on the huge radiators, there is a big thermal problem here. It is especially this circuit. I have seen in my other experiments yet no such burden. Otherwise, the thyristors, igbt modules have not even felt warm. The diodes I have now lowered with 200a modules on hand warming. Finally, I would like to further increase the coupling capacitors capacity so that I can transfer more energy while more heat is generated again. I think it will be short but very strong currents flow otherwise I can not explain that. Greetings Sven
Von Samsung Mobile gesendet
-------- Ursprüngliche Nachricht --------
Von: "onielsen@... [EVGRAY]"
Datum:17.01.2018 20:49 (GMT+01:00)
An: [email protected]
Betreff: [EVGRAY] Re: buck plug ...
Hi Sven,
The max current given in the datasheets is when the surface of the device is held a 20 degrees Celcius. This requires water cooling. Else the current has to be derated according to the datasheet. Don't dissipate more power in the device than allowed according to the datasheet and ambient temperature. Use appropriate heat sink to dissipate the power.
Regards
Ole
---In [email protected], <s.friedrich@...> wrote :
Hello Hector,
why did you use these powerful diodes 180 amps, I'm assuming that thyristors must also be so stable.
I was lucky and cheap get a power block.
Is it only possible to perform diode plug experiments without the components dying the heat dead. It is of course possible to cool small components with huge cooling effort, the only question is what is cheaper.
Who still transverter experiments performs or experiment please report.
regards
Sven