Bridge rectifier question.

Is it feasible to parallel bridge rectifiers for the purpose of increasing the current capability?

Yes, but just buy the larger rectifier, please. Paralleling diodes (or switch gates in general) is not without some real hazards. Depending on the gate technology, two parallel devices might share load nicely, or one device might take the majority of the load onto itself, blowing up nicely, leaving the surviving device to take all the load and blow up even faster. For example, MOSFETs can usually be paralleled quite nicely, but even then you need reasonably matched devices and really good thermal coupling across all the devices for load sharing to work properly. BJTs, on the other hand, need additional circuitry to be paralleled successfully. MOSFETs have a positive coefficient of temperature to on-resistance, while BJTs have an analogous negative coefficient. This makes paralleling MOSFETs fundamentally stable, and BJTs fundamentally unstable.

The considerations for paralleling diodes are covered in this application note. If that's too technical, just buy the larger rectifier, please.

Whatever you do, please do not parallel a mix-and-match set of different bridge rectifiers. That's just asking for trouble.

I'm building a router plane and I need to build a DC power supply for the DC 24v carriage motor. The motor label says 35 amp .83HP. I bought a 40 VA transformer and I was thinking that the unit would be more robust using two matching 40 AMP bridge rectifiers. I just like over doing things.

My old fashioned math says 24 volts x 35 amps is 840 va.

Red James said:

My old fashioned math says 24 volts x 35 amps is 840 va.

Click to expand...

Yep - 40VA is spit in the wind compared to your load.

Kevin Beitz said:

Is it feasible to parallel bridge rectifiers for the purpose of increasing the current capability?

Click to expand...

you better off with a DC motor controller.

DC PWM Motor Controller Reversible 12V 24V 36V 4A High Power Motor Speed Controller Switch CCM6Ds-B: Amazon.com: Industrial & Scientific

dee
;-D

It can be done, but you probably can't do it safely.

As diodes heat up their forward voltage drops significantly, which causes the hottest diode to handle the most current, get hotter, take more current, then fail.

So if you do it wrong you just end up blowing all of your diodes in a sequence, or reduce their reliability substantially.


Generally the only time you need to parallel diodes is when the voltage is too high for you to select a high current enough diode, in which case I typically use avalanche rated diodes in series.


The easiest way to parallel diodes is to thermally couple them together as best you can, and then to figure out what the negative resistance is at equilibrium temperature at multiple currents. From there you can extrapolate what the mV drop per amp is, (you'll get a negative number with the unit ohms or milliohms) and add that much positive resistance in series with it. Done properly you will get perfect current balance under all load conditions, but you may need to add a little resistance to one side or the other to match diode differences. More resistance will help with current balance, but too much resistance wastes power.

As it sits, unless your rectifiers are significantly overrated for the application, your circuit will probably not work well.

If you are still set on this approach I can walk you through the math to size resistors based on information from the datasheet.

You can do it quite nicely, but the extra components needed are much more expensive than the bigger rectifier (balancing transformer, or a couple identical high current inductors, etc).

Those $2-3 bridges are worth what you pay. Power diode pack modules come with configurations like cathode-to-cathode, anode-to-anode, independent.

https://media.digikey.com/pdf/Data Sheets/Vishay Semiconductors/IRK 56, IRK71.pdf

Buried on page 10 of sfriedburg's application note link is the brute force solution:

If the layout is not symmetrical, the connection resistors will increase the current imbalance.
In contrast, a symmetrical layout will balance the current in each diode. To have a real
impact, the values of these resistors have to be of the same order as the resistance of the
diodes. Adding the small resistance in series with each diode can be a good way to balance
the current. On the other hand this solution will generate more power losses and will
decrease the efficiency of the converter.

A bridge rectifier is four identical diodes wired in a particular manner. For your situation it might be cheaper to buy four suitable diodes and connect them in a bridge configuration.

I put my parts together 40va transformer and one rectifier. The 40va was not big enough. So I bought a 1000va transformer and I'm looking at a 400amp rectifier. Will this be big enough for this motor?

And thanks for everyone's help.

Could you specify the 1000VA transformer part number? A 400amp rectifier is way over-building. Doubling the current number from 35 to 70 is what I would choose, or slightly lower than 70 depending on price and availability.

If you meant 40amp rectifier then specify what the part number is.

400amp is an expensive bridge rectifier. Look for 1000v 50amp rectifiers. That's plenty of derating. If you look for lower voltage rectifiers you run into much smaller amperage, generally 600v at 25amps, without out paying lots of money.
Quick look at eBay for 1000v/50a is $5 US seller, $1 China.

35 amps at 24 volts is nothing.

Bridge rectifiers or single diodes can be used.

Single diodes may have better wire attachment but so many options.

Heat sink should be used as well.

One may consider either a motor controller or pre made power supply as there are many sources of pre made units and they usually are switch mode which is far more efficient than brute force transformer and diodes plus whatever caps or other devices to regulate.

Many are sensing supplies where a second set of wires connect to the load point to regulate voltage at the load.

Sent from my SAMSUNG-SGH-I337Z using Tapatalk

Just get a 24V 40A switch mode power supply.

I can buy a 400 amp rectifier for around $20.00 off E-bay. I got the 1000va transformer tral cheap...

Kevin Beitz said:

I can buy a 400 amp rectifier for around $20.00 off E-bay. I got the 1000va transformer tral cheap...

Click to expand...

Link the rectifier or give us a part number.

If it is the MDQ400 part that seems to pop up everywhere when I search for a 400A recitifier, be aware that due to its 1600V rating its diodes have a forward voltage over 1.5V, which adds up to 3V across the rectifier.

This means over 12% of your power is lost through the rectifier, and you will have to dissipate over 100 watts from that rectifier within your enclosure. The datasheet doesn't even mention a thermal resistance from junction to atmosphere, so that's a dead giveaway this can't dissipate much heat on it's own without a big heatsink. It probably won't burn out, but will run so hot (if the case reaches 90C under full load conditions then the junction is 100C which is its maximum operating temperature. This isn't a rule of thumb, I got that number after some fairly complex math) that it isn't any better than an appropriately sized rectifier in the first place.



My recommendation is to get 4 discrete diodes, such as MBR60100 which has a forward voltage of 0.84V, which results in a much lower loss, less heat to dissipate, and four areas to remove that heat from instead of lumping all of the hot components together and trying to cool them all together. Rated for 60A continuous and 100V it's about as overbuilt as you can get without compromising on performance.

If you can get them in DO5 packages (you'll want two of MBR60100 and two of MBR60100R) they are pretty fun to wire and look really nice if you use two beefy chunks of aluminum to connect the DC rails. The TO247 package that is more commonly available is fine too but you'll need (dirt cheap) isolation pads where you attach them to the heatsink. I can walk you through all of this if you need.

A quick note on diodes in rectifier applications: Diodes see double the output voltage across them, and your transformer will output way more than 24V.
If you have a 24V output transformer, chances are it's going to be around 15% higher when under low load, so 27.6V . However, that's RMS, not the peak voltage you will see, which will be 1.4 times higher than RMS, so 38.7V . Doubling that (this is due to the configuration of full bridge rectifiers, and is not a safety margin. You will actually see this voltage across the diodes at some point in the cycle) gives 77.4V . You could go with 80V diodes, but that is a little close for comfort even though they would most likely outlast the motor.

Pictures of the machine, motor,rectifier and transformer that I would like to use.

Kevin Beitz said:

Pictures of the machine, motor,rectifier and transformer that I would like to use.

Click to expand...

That stuff will work, so long as you heatsink that rectifier, but expect 12.5% lower motor speed due to rectifier losses. Typically you would use a 26V or so transformer to counter this.

I would really not suggest using that rectifier. It's like putting a giant transmission on a tiny car. It crosses the line from overbuilt to underengineered.

In case you were wondering that motor is 77% efficent at full load.