Fan-less power supply needs a Vicor DCDC Module?.....not an "Open Frame" type DCDC module?

[zenerbjt]

Dear Engineers,

We wish to design a modular power supply of the following spec:


Vin = 42-59VDC (14S LiPO)
Vout = 12V
Iout = 11A
Isolated (500V)
Not allowed to use fans to cool it….just heatsink.
Enclosure is a totally sealed carbon fibre box of dimension 50cm x 20cm x 10cm (high).
Max ambient outside of box = 45degC.

Given the non-use of fans, I believe you would agree with me that this more or less confines us to a Vicor module type solution? Due to the greater thermal conductivity of the material of the vicor modules.

For example, A Vicor PRM48AT480T400A00 cascaded with a Vicor VTM48ET120T025A00 would cover our spec. Their datasheets give the thermal resistance of their hottest component to the case as 1 degC/W. You simply can't get that low with "open PCB" type DCDC modules, because they use gap-pad to thermally couple to their little metal base-plate.

“Open PCB” type DCDC modules such as eg the Murata UEE-12/12.5-D48NB-C just can’t compete for such a fan-less application. Do you agree?

The power components of the UEE-12/12.5-D48NB-C are gap-padded to the metal base-plate. Clearly, gap pad isn’t as good a thermal conductor as the Vicor module material. The UEE-12/12.5-D48NB-C module’s datasheet doesn’t even provide any thermal resistance figures.

Clearly , the “open PCB” type DCDC modules must have a fan blowing over them in order for them to operate at anywhere near maximum power. Do you agree?

Vicor PRM48AT480T400A00 datasheet

http://cdn.vicorpower.com/documents/datasheets/PRM48AF480T400A00_ds.pdf

Vicor VTM48ET120T025A00 datasheet

http://www.vicorpower.com/documents/datasheets/VTM48E_120_025A00.pdf

Murata UEE-12/12.5-D48NB-C datasheet
http://www.murata.com/products/productdata/8807039893534/uee150w.pdf?1583754815000

Here is a picture of an “open-PCB” type , “non-Vicor” DCDC module…..
http://www.digikey.co.uk/product-de...2-12-5-D48NM-C/UEE-12-12-5-D48NM-C-ND/4747528
……You can clearly see that its meant to only be used with blown air. There is no high-thermal-conductance potting like you get in the Vicor modules.

Surely you agree, that our Fan-less spec predicates a Vicor DCDC module approach?

 

[FvM]

Enclosure is a totally sealed carbon fibre box of dimension 50cm x 20cm x 10cm (high).
Max ambient outside of box = 45degC.

Where's the heat sink? In- or outside the box?

 

[zenerbjt]

Heat sink for the DCDC module is inside the box, connected to the DCDC module.

 

[FvM]

In this case, major issue is heat dissipation through enclosure wall. Internal fan may helpful, no advantage of sealed modules over open frame converter.

 

[zenerbjt]

Hi, are you sure?

I thought to come back as I don’t believe I emphasised how much worse the “open-frame_plus-baseplate” type DCDC modules are when compared to vicor modules. This is especially the case for us because we have no fan cooling. That cant be emphasised enough. Not only that, but because the PSU is inside an enclosure, then we don’t even have “natural convection” either. So none of the graphs in the “Open-frame” datasheets truly apply to our case. Our performance would be worse than predicted by these graphs. As you know, datasheet “natural convection” means literally that….in free unenclosed air.

The vicor modules have a thermal resistance (JC) of 1 degC/W. That’s extremely impressive…..similar to a TO220 power transistor. None of the open-frame DCDC modules even provide any thermal resistance data. -though it would be far worse than the vicor module case. Looking at the Murata module we have here shows that the power FETs are gap-padded to the thin metal baseplate (heatsink). The case (not the metal tab) of the power transistors contacts with the gap pad…which is a poor thermal situation….thermal resistance through the plastic case of a transistor is at least 20degC/W worse than through the tab. As such, the thermal situation for the open frame type DCDC modules is far worse than the vicor module. The open-frame type dcdc’s are meant for use with a fan, or at least “natural convection”, which is defined as literally in free air, (not in an enclosure like our case).

The murata module datasheets thus cant be exactly used for our case…however, I believe we can definitely say that in an external ambient of 40degC, the Murata DCDC modules wont be able to provide our maximum power over the 14S voltage range.


Another point about the vicor modules is that they can be heatsink-connected on back and front sides…..whereas the open-frame style ones have only connection via one side. Needless to say, this would make a very significant difference to thermal performance, especially since we have no fan.

To really properly test the open frame modules, I think would need a large thermal chamber, which I believe are available for hire useage in places like EMC labs etc. As you know FET rds(on) gets higher with temperature, meaning lower temperature test extrapolation to higher temperatures isn’t too accurate.

so are you sure vicor = open frame thermally?

--- Updated Sep 16, 2020 ---

Hi, are you sure?

I thought to come back as I don’t believe I emphasised how much worse the “open-frame_plus-baseplate” type DCDC modules are when compared to vicor modules. This is especially the case for us because we have no fan cooling. That cant be emphasised enough. Not only that, but because the PSU is inside an enclosure, then we don’t even have “natural convection” either. So none of the graphs in the “Open-frame” datasheets truly apply to our case. Our performance would be worse than predicted by these graphs. As you know, datasheet “natural convection” means literally that….in free unenclosed air.

The vicor modules have a thermal resistance (JC) of 1 degC/W. That’s extremely impressive…..similar to a TO220 power transistor. None of the open-frame DCDC modules even provide any thermal resistance data. -though it would be far worse than the vicor module case. Looking at the Murata module we have here shows that the power FETs are gap-padded to the thin metal baseplate (heatsink). The case (not the metal tab) of the power transistors contacts with the gap pad…which is a poor thermal situation….thermal resistance through the plastic case of a transistor is at least 20degC/W worse than through the tab. As such, the thermal situation for the open frame type DCDC modules is far worse than the vicor module. The open-frame type dcdc’s are meant for use with a fan, or at least “natural convection”, which is defined as literally in free air, (not in an enclosure like our case).

The murata module datasheets thus cant be exactly used for our case…however, I believe we can definitely say that in an external ambient of 40degC, the Murata DCDC modules wont be able to provide our maximum power over the 14S voltage range.


Another point about the vicor modules is that they can be heatsink-connected on back and front sides…..whereas the open-frame style ones have only connection via one side. Needless to say, this would make a very significant difference to thermal performance, especially since we have no fan.

To really properly test the open frame modules, I think would need a large thermal chamber, which I believe are available for hire useage in places like EMC labs etc. As you know FET rds(on) gets higher with temperature, meaning lower temperature test extrapolation to higher temperatures isn’t too accurate.

so are you sure vicor = open frame thermally?

 

[zenerbjt]

In this case, major issue is heat dissipation through enclosure wall. Internal fan may helpful, no advantage of sealed modules over open frame converter.

Thanks but if I explain more then i believe you will agree that the sealed module, in this particular case, is better. We are not allowed any fans. In both cases, the sealed vicor module and the open frame module are on a large heatsink. The thermal resistance from hottest component inside the sealed module, to the heatsink, is far less than that for the “open-PCB-plus-baseplate” module. As such, I believe that you would agree that the sealed module (vicor module) would be much better from a thermal point of view?

 

[FvM]

Don't know, heat dissipation through the enclosure is the major problem. Converter efficiency is respectivly a key parameter, it determines the internal overtemperature for a given power throughput. Internal thermal resistances and component temperatures don't matter as long limit temperatures aren't exceeded.