# What efficient techniques exist for rectifying high voltage?

I want to rectify 120V (mains North America) at a relatively high amount of current (5 amps), so naturally my first idea is to just find a bridge rectifier. After finding one capable of the current and voltage, I realize that the power dissipated would be high enough to warrant a heat sink due to the large forward voltage of the diodes.

V(fwd) = 1.1v per diode 2 diodes are always in conduction so 2.2v drop 2.2 * 5 = 11 watts of heat dissipation

I started looking into alternatives but I did not seem to come up with much. So, back to the question, what solid state techniques exist to rectify high voltage AC that remain efficient even under high currents?

• Why would you need to rectify the mains directly? Do you need that voltage also in DC? Feb 17, 2012 at 7:37
• Maybe he wants to do power factor correction, which involves rectifying the mains and then passing it through an DC/DC like converter (actually based on a boost converter if I recall properly) for an as close to ideal 'resistor' load as possible. This is often done in high-power supplies. A high probability that your PC's ATX supply got one too.
– Hans
Feb 17, 2012 at 20:33
• A transformer to step it down first at the required currents and 60 hz would be too large for me, so I will do it with a SMPS instead. Also, I am probably going to do PFC as Hans said. And you are right, it is basically a boost converter. Feb 19, 2012 at 4:11

At 5A power handled = 120 x 5 = = 600 Watt so your 11Watts of dissipation is about 2% loss - this oild be acceptable in most cases. The cost of getting lower losses may exceed the cost of heatsinking 11 Watts and the reliability may be lower.

BUT: Schottky diodes may about halve power losses. High Voltage SiC Schottky are available. Beware of the can-be-horrendous reverse leakage at elevated temperatures.

As Martin says - synchronous rectification will work BUT the cost of switch devices and controller may be dearer than a heatsink. If eco-greenery insists on low losses then the expensive solution may be needed. Your 1.1V/2 diodes at %a = equivalent R of R = V/I = 0.55/5 = 0.11 ohms = 110 milliohms.To be useful a synchronous rectifier MOSFET would need to be << 110 milliohm in a 200 V +++ FET.

Doable, but a heatsink seems attractive.

• Thanks for your response, and I completely agree that a 2% loss is totally fine for rectifying mains, but I want to do away with as much heat sinks as possible as a challenge to myself. It would make me more proud of my work. After all, I am making only one unit, for myself, so a cost difference of say 10 dollars is fine with me. :P Feb 19, 2012 at 4:08
• The SiC shottkies have a real foward drop of say 1.5V when you are using reasonable currents .If you used the SiC you could be worse off than you would with the tried and trusted SI bridge .As for normal schottkies I guess that 200V ones would save significant power when rectifieng 120V mains BUT you would have to be careful about power surges . Feb 25, 2016 at 19:48
• As Vr spec rises the Vf gets worse and it is best to keep Vr as low as is usable in your application. Feb 26, 2016 at 14:46
• @Autistic Vf varies with spec (as I know you know). When comparing similarly specd parts at not too high voltages the Schottky will have lower Vf and you can easily get parts that are well below 1V Vf at 25C.| As an example only these Vx5202 from Vishay are about \$0.75 each in 1's from Digikey and 200V 5A rated. I'd use higher spec parts - but these have worst case Vf of 0.88V at 25C dropping to 0.73V worst case, 0.65V typ at 5A at 125C. As they are from Vishay - if they do not meet that worst case spec they will be replaced by ones that do. Feb 26, 2016 at 14:50
• At 200V rating Schottky can have usefully lower Vf. Much above that and std Si has lower Vfs. Feb 26, 2016 at 14:51

Synchronous rectification using some type of power transistor eg MOSFETs are more efficient but you have to add a control circuit to switch them at the right time. For 5 amps, I would stick with diodes and mount the bridge on a heatsink, that is what the hole in the middle of high power bridges is for.

• Thanks for your response, and I agree that just using a heat sink on a diode bridge would be most sensible and cheapest, but as I commented in the above question, I want to have as little wasted power as possible as a challenge to myself, and the increased cost is not that important for me since I am only making one unit. (for myself) Feb 19, 2012 at 4:12