| Electronics > Projects, Designs, and Technical Stuff |
| Best high current linear regulator configuration |
| (1/3) > >> |
| ricko_uk:
Hi, What is the best design/solution for a regulator for a PCB with the following characteristics: - linear (because of stringent noise requirements) - 2A out (regardless of Vin and Vout) - Vin from 12V to 26V supplied by an external wall adaptor. - Vout is variable (through an ADC connected to a uC or other means) from 2V to (Vin-2V). - the PCB with the regulator (not the wall adaptor obviously) is to be placed inside an enclosure without any air vents (and possibly even potted). That means that with respect to heat dissipation the worst case scenario is when Vin is 26V and Vout is 2V. At 2A load that is 48W dissipated just by the regulator. 1) What is the best solution using linear regulators? The classics are two regulators (but then have to add balancing resistors) or bypass transistor. Which is best and why? Any parts/designs suggestions? 2) Any other topologies/solutions? 3) Any other concerns? For example, I'm thinking the 48W dissipated inside an enclosure without ventilation? The elements that heat up (regulator, bypass transistor etc) would have heat sinks but how safe is that? 4) Would any potting (of the entire PCB) improve or worsen the thermal situation and general safety? Many thanks :) |
| Ian.M:
3) With 48W dissipation, it will cook itself to death unless its in a large metal enclosure, with enough surface area to dissipate the heat and with the high dissipation parts on heat spreaders thermally coupled to the enclosure, or alternatively if you fit a large enough exterior heatsink with all high dissipation parts mounted to its inner face. |
| Whales:
Common alternative topologies: A. Multi-tap transformer followed by linear reg. Relays select which tap gets used. Dramatically lowers P loss in linear reg. Requires an array of comparators or a micro. B. Switch-mode followed by linear reg. Special attention may be needed for common-mode noise, depending on your requirements. If this is a one off: computer heatsinks + fans make cheap and easy coolers. Edit: yes for 48W you will need forced airflow; or a crazy amount of thick-whole-case-side-multi-kg heatsinking if you want it passive. FETs may be easier to parallel than BJTs - can someone comment here regarding runaway? I believe (but I may be wrong) that FETs have better thermal coeffs? Make sure the check the SOA of your linear transistors. If the datasheet doesn't provide a DC SOA then don't trust it. I can't recall exactly, but I think the higher-voltage capable FETs tend to have better SOA. Requires lots of looking around, the only obvious search term hits (like the Linear L2 series) cost a lot per FET even thought other manufacturers make similarly specced items for less. |
| Whales:
> 4) Would any potting (of the entire PCB) improve or worsen the thermal situation and general safety? You have to think of the individual thermal and safety problems. Potting should have nil effect on your main pass transistors, because they have to already be secured to a big heatsink anyway. As long as you don't pot the fins of the heatsink, that is :) "Safety" is not a one-word topic. You need to look at what parts of your supply can fail and what bad stuff could happen because of this. I would avoid potting anyway unless this is for production or you have a specifically harsh environment (eg wet), it makes it very hard to repair/modify. |
| Circlotron:
--- Quote from: Whales on April 19, 2020, 12:39:29 am ---FETs may be easier to parallel than BJTs - can someone comment here regarding runaway? I believe (but I may be wrong) that FETs have better thermal coeffs? --- End quote --- MOSFETS parallel great when they are fully switched on. The hotter one increases it’s drain-source resistance and so conduction backs off. BUT... A MOSFET partially switched on as per a linear regulator, the hotter on will have a lower gate threshold voltage causing it to turn on even more than the increased drain-source resistance tries to back it off and so hogging the current. Not what you want. A resistor in each mosfet source lead may help somewhat. |
| Navigation |
| Message Index |
| Next page |