Dummy Load for Bench Power Supply - Resistors?

[Benta]

Incandescent lamps are cheap.

And unfortunately very non-linear. The cold resistance can be 5...6 times lower than at temperature.

[DW1961]

As a rough rule of thumb, those aluminium cased power resistors are basically halved in power rating when not attached to a heat sink.

I was researching and some say de-rate them by x4 also, not the usual 50%. So a 100 w resistor would be max sustained 25 watts.

[DW1961]

I've been looking at Nichrome wire, also. I could wind that around ceramic posts. Then just connect where I need the wattage to be. However, what you have then is basically exposed heater coils. I'd have to put them in a box with a fan. Then, creating my own space heater. lol

[Stray Electron]

Is there anyway to run two identical resistors together for x2 power resisted?

  If you want to keep the original resistance than you need to use four equal value resisters. Two in series and then that in parallel to two others in series. Each string will have double the resistance but 2R paralleled to 2R would yield 1R. The advantage sn that the heat dissipation is quadrupled.

[Gyro]

I've been looking at Nichrome wire, also. I could wind that around ceramic posts. Then just connect where I need the wattage to be. However, what you have then is basically exposed heater coils. I'd have to put them in a box with a fan. Then, creating my own space heater. lol

Yep, that's what it is!

Remember that Nichrome isn't solderable (aside from the temperature of the joint). You need to crimp the connections - or you can get high temperature ceramic screw  terminal blocks, eg. https://www.ebay.com/sch/i.html?_from=R40&_trksid=p2380057.m570.l1311&_nkw=ceramic+terminal+block&_sacat=0

[Wallace Gasiewicz]

Need lots of nichrome wire and be careful, I almost started my bench on fire as a kid

You would be making a toaster

You can also wind the wire around a ceramic  tube and put it in the Cantenna. Ater measueing the ohms to your satisfaction.

[pqass]

I used to futz around with power resistors but then watched mikeselectricstuff whip-up a crude e-load so I made my own.  It was really really easy and have used it many times to test power supplies and batteries.

For less than $20 you could easily create a 360W load with a pot, LM324, 4xMOSFETs, 4x4x1R 1W resistors, and a handful of passives.  I've used an old 6cm² CPU heatsink but if you're hard-up, just attach the MOSFETs to one end of a long Al plate and submerge the other end in a bucket of water (keeping just the MOSFETs above the water level).

See attached schematic.  Make four of these for the full 360W.  Just tie the MOSFET drains together as your positive post, ground is your negative post. The pot wiper is attached to all +in pins of the opamps.  You could use fewer modules but I wouldn't push more than 100W per MOSFET module.

http://www.falstad.com/circuit/circuitjs.html?ctz=CQAgjCAMB0l3BWcMBMcUHYMGZIA4UA2ATmIxAUgoqoQFMBaMMAKAEMQUUAWEbvKtzh8BIYuBRRk8YTAzdCPbGEKEEebGnlRo8xZRlww2bWEMsATn2H9BxQiKpUzkS9drde3exU9SXrgDu7r68XLwIfkGcPKExXj6uAGZ8PrYUhIKimiDYOiRo3GAI4cTGLMEImakOVR68rgDmGVSZeC0gig1QbnVxfW1SqAgsAMYh6UL1Q7DwEAjQxCaQxHglhEJ4xFw6Rixg2ht8KuB+CGCSEAFOs0gASnQAzgCWjwAubAB2o3QszUfnSQAlDtJwsZYdbCEdp9BS8CBgu6Q6GQtBDay5JwzEZIo7YKGdMJ4UFSIQSdEwEbBMB+OHxPhHVwAB1OXiONIiF3AQxYLK6cX5gwRPWC-MiEWq4pFnVigwG+GlHIoXIBXOZMrZvDFfmF0X5dNhjN5NUcJoQGAcuoqJrl1UG0T6Uu8tSibjxBKORQczgMcGtzs6CoD5u9LAAHp11HwSiAMAgkNx43xOLwAJJ3ABiAAUUJRw51qtwLWJVnwMOJ4e07gATR4ACgA9p8AJQAHQAjjB8AgADStx4ANUajw7aAHfcHw7rbwAFm329gB-miA5uLFsIm+Jhk7EADIAWU03HzahhCHEAnEVUklfAd37Yf73H7z0+-aZbAsbAANt+6N+WAbagjnIIRSAxQhoG6K5MkAkB2hA6xwLMTooO5CAUGgEYgA

[Electro Fan]

That is one pretty load.

+1 

[wizard69]

Well if you are going that way , almost any material with a resistance would work.   Steel banding used for heavy shipping for example could be coiled  up to create a load.  In either case you have two problems.   One is that you are creating an inductor which might be important.   The other thing is that you have a very nice fire starter.

Somebody mentioned above "the resistor in the paint can" approach.   That is a very good idea if you need to test for an extended time.   You can also air cool.  I've seen huge motor drives in industry with massive wire loads mounted above the control panels that are fan cooled.   You can build the same thing if you keep an eye on safety.

I've been looking at Nichrome wire, also. I could wind that around ceramic posts. Then just connect where I need the wattage to be. However, what you have then is basically exposed heater coils. I'd have to put them in a box with a fan. Then, creating my own space heater. lol

[DW1961]

I used to futz around with power resistors but then watched mikeselectricstuff whip-up a crude e-load so I made my own.  It was really really easy and have used it many times to test power supplies and batteries.

For less than $20 you could easily create a 360W load with a pot, LM324, 4xMOSFETs, 4x4x1R 1W resistors, and a handful of passives.  I've used an old 6cm² CPU heatsink but if you're hard-up, just attach the MOSFETs to one end of a long Al plate and submerge the other end in a bucket of water (keeping just the MOSFETs above the water level).

See attached schematic.  Make four of these for the full 360W.  Just tie the MOSFET drains together as your positive post, ground is your negative post. The pot wiper is attached to all +in pins of the opamps.  You could use fewer modules but I wouldn't push more than 100W per MOSFET module.

http://www.falstad.com/circuit/circuitjs.html?ctz=CQAgjCAMB0l3BWcMBMcUHYMGZIA4UA2ATmIxAUgoqoQFMBaMMAKAEMQUUAWEbvKtzh8BIYuBRRk8YTAzdCPbGEKEEebGnlRo8xZRlww2bWEMsATn2H9BxQiKpUzkS9drde3exU9SXrgDu7r68XLwIfkGcPKExXj6uAGZ8PrYUhIKimiDYOiRo3GAI4cTGLMEImakOVR68rgDmGVSZeC0gig1QbnVxfW1SqAgsAMYh6UL1Q7DwEAjQxCaQxHglhEJ4xFw6Rixg2ht8KuB+CGCSEAFOs0gASnQAzgCWjwAubAB2o3QszUfnSQAlDtJwsZYdbCEdp9BS8CBgu6Q6GQtBDay5JwzEZIo7YKGdMJ4UFSIQSdEwEbBMB+OHxPhHVwAB1OXiONIiF3AQxYLK6cX5gwRPWC-MiEWq4pFnVigwG+GlHIoXIBXOZMrZvDFfmF0X5dNhjN5NUcJoQGAcuoqJrl1UG0T6Uu8tSibjxBKORQczgMcGtzs6CoD5u9LAAHp11HwSiAMAgkNx43xOLwAJJ3ABiAAUUJRw51qtwLWJVnwMOJ4e07gATR4ACgA9p8AJQAHQAjjB8AgADStx4ANUajw7aAHfcHw7rbwAFm329gB-miA5uLFsIm+Jhk7EADIAWU03HzahhCHEAnEVUklfAd37Yf73H7z0+-aZbAsbAANt+6N+WAbagjnIIRSAxQhoG6K5MkAkB2hA6xwLMTooO5CAUGgEYgA

Would that only test 12 or 30V? If I could find a kit I would buy it and just assemble it myself.

[pqass]

...
Would that only test 12 or 30V? If I could find a kit I would buy it and just assemble it myself.

Actually, it would depend on the MOSFET. The IRFP250 datasheet says it is rated to 200V or 30A max. so any voltage or current below these limits. However, practical circuits should give a good buffer and stay below 175V. 

But what really limits the circuit is the heat transfer capability of the package.  The TO-247 package is better at transferring heat from the die to the environment then say, a TO-220.  That's why I suggested to keep each MOSFET at 100W max.  So you can do 175V@571mA or 30V@3.3A or 10V@10A.   In the circuit shown, I chose to limit the current to 3A. If you increase this design limit then that changes the wattage of the 0.25R shunt; eg. 25V@4A keeps the MOSFET to the 100W limit but now the shunt must be 4.7W+margin (probably should double that). You can try this yourself in the simulation by lowering the 150K resistor to 100K and the voltage source from +30V to +25V. 

As I suggested, if you parallel 4 of these circuits then you multiply the 3A individual circuit limit by 4 or 12A. And if you stay at 30V or below, you won't overshoot the 100W limit per MOSFET or a combined 400W. If you plan on inputting more than 30V then you'll have to turn-down the pot.  Say, about half-way (1.6A per circuit) for up to 60V, etc.

This circuit can be assembled on a small perfboard; 1 IC+passives with short leads to 4 mounted MOSFETs, shunts, and a pot.  I think it's a good beginner project.

[DW1961]

Is there anyway to run two identical resistors together for x2 power resisted?

  If you want to keep the original resistance than you need to use four equal value resisters. Two in series and then that in parallel to two others in series. Each string will have double the resistance but 2R paralleled to 2R would yield 1R. The advantage sn that the heat dissipation is quadrupled.

I think that's what I'm going to do, but can I run more than 4 and maintain 6ohms? I was thinking 8 100w resistors. Max 300w watts, divided by 8 resistors should keep them pretty cool?

[theHWcave]

Have a look at this video https://youtu.be/Fxp0nxphgYw where I build a power resistor setup for this purpose (and many other uses). These really have become invaluable over time.  The idea is to use 10, 20, 40, and 80 Ohm  (and 1, 2,4,8) because by parallel and serial connections that gives you an incredible wide selection of possible uses and power ratings. I calculated a table of resistances and power ratings of all possible combinations which you find on my Github page https://github.com/TheHWcave/BCD-Power-Resistors

[beanflying]

Have a look at this video https://youtu.be/Fxp0nxphgYw where I build a power resistor setup for this purpose (and many other uses). These really have become invaluable over time.  The idea is to use 10, 20, 40, and 80 Ohm  (and 1, 2,4,8) because by parallel and serial connections that gives you an incredible wide selection of possible uses and power ratings. I calculated a table of resistances and power ratings of all possible combinations which you find on my Github page https://github.com/TheHWcave/BCD-Power-Resistors

Nice idea. The reason I went 1,2,3 & 4 rather than BCD was I could get two matching pairs of 5 No right or wrong just that was my reasoning at the time.

[Wallace Gasiewicz]

Here is a solution a 6.8 Watt 200 Watt resistor, Put it on a big heatsink with a fan or put it on a smaller heatsink in a can of oil.

US Stock 6.8 ohm 6R8 200W Watt Aluminum Housed Metal Case Wirewound Resistors

Just copy the above and past it into e bay

[DW1961]

Got something on the way. I'll post after I wire it up.


Thanks for all of the help and ideas  from everyone.

[Electro Fan]

Here is a solution a 6.8 Watt 200 Watt resistor, Put it on a big heatsink with a fan or put it on a smaller heatsink in a can of oil.

US Stock 6.8 ohm 6R8 200W Watt Aluminum Housed Metal Case Wirewound Resistors

Just copy the above and past it into e bay

Looks like they might also have similar 200 watt 10 ohm resistors for about the same price.

[Kleinstein]

...
Would that only test 12 or 30V? If I could find a kit I would buy it and just assemble it myself.

Actually, it would depend on the MOSFET. The IRFP250 datasheet says it is rated to 200V or 30A max. so any voltage or current below these limits. However, practical circuits should give a good buffer and stay below 175V. 

But what really limits the circuit is the heat transfer capability of the package.  The TO-247 package is better at transferring heat from the die to the environment then say, a TO-220.  That's why I suggested to keep each MOSFET at 100W max.  So you can do 175V@571mA or 30V@3.3A or 10V@10A.   In the circuit shown, I chose to limit the current to 3A. If you increase this design limit then that changes the wattage of the 0.25R shunt; eg. 25V@4A keeps the MOSFET to the 100W limit but now the shunt must be 4.7W+margin (probably should double that). You can try this yourself in the simulation by lowering the 150K resistor to 100K and the voltage source from +30V to +25V. 

As I suggested, if you parallel 4 of these circuits then you multiply the 3A individual circuit limit by 4 or 12A. And if you stay at 30V or below, you won't overshoot the 100W limit per MOSFET or a combined 400W. If you plan on inputting more than 30V then you'll have to turn-down the pot.  Say, about half-way (1.6A per circuit) for up to 60V, etc.

This circuit can be assembled on a small perfboard; 1 IC+passives with short leads to 4 mounted MOSFETs, shunts, and a pot.  I think it's a good beginner project.

Even 100 W from a TO247 part needs a good heat sink.
Beside the total power and voltage there can also be lower SOA limit. So not every combination of voltage and current to give the 100 W may be reliable. Especially high power at a relatively high voltage can be a problem and lead to failure (some samples are OK, some may fail). The SOA specs on MOSFETs are a tricky topic - often there is no curve in the DS and even if there is may not be valid.

The resistor version has definitely some advantage in reliablity and may get a way with a smaller heat sink. However it is less flexible. With the resistors the chances are smaller to get fakes from Ebay or Banggood. So lower cost sources may be acceptable.
As a crude test an old toaster may make a cheap high power resistors, usually better than incandesant lamps.

[DW1961]

Have a look at this video https://youtu.be/Fxp0nxphgYw where I build a power resistor setup for this purpose (and many other uses). These really have become invaluable over time.  The idea is to use 10, 20, 40, and 80 Ohm  (and 1, 2,4,8) because by parallel and serial connections that gives you an incredible wide selection of possible uses and power ratings. I calculated a table of resistances and power ratings of all possible combinations which you find on my Github page https://github.com/TheHWcave/BCD-Power-Resistors

HA, found that a couple nights ago even before you posted. I watched the entire video. Very good. That's sort of what I am doing, except I'm going to use 4 100 watt 6 ohm resistors in various series or parallel to get what I need. I already spent $29.00, so it's enough for now. Heatsinks are expensive. I think I may have gone too small on the heatsink, too, but it is what it is for now. (Any biggest on the sink and the prices doubled or more)

Heat sink:
Resistors: https://www.amazon.com/gp/product/B07KJZYW7J/ref=ppx_yo_dt_b_asin_title_o00_s00?ie=UTF8&psc=1
Heat Sink: https://www.amazon.com/gp/product/B08T8837ZH/ref=ppx_yo_dt_b_asin_title_o00_s01?ie=UTF8&psc=1
The heat sink is only 5.9 x 2.75 x 0.38 but has 28 fins on it.

I'll also attach a decent 12 V CPU cooler to it if I need it.

[DW1961]

...
Would that only test 12 or 30V? If I could find a kit I would buy it and just assemble it myself.

Actually, it would depend on the MOSFET. The IRFP250 datasheet says it is rated to 200V or 30A max. so any voltage or current below these limits. However, practical circuits should give a good buffer and stay below 175V. 

But what really limits the circuit is the heat transfer capability of the package.  The TO-247 package is better at transferring heat from the die to the environment then say, a TO-220.  That's why I suggested to keep each MOSFET at 100W max.  So you can do 175V@571mA or 30V@3.3A or 10V@10A.   In the circuit shown, I chose to limit the current to 3A. If you increase this design limit then that changes the wattage of the 0.25R shunt; eg. 25V@4A keeps the MOSFET to the 100W limit but now the shunt must be 4.7W+margin (probably should double that). You can try this yourself in the simulation by lowering the 150K resistor to 100K and the voltage source from +30V to +25V. 

As I suggested, if you parallel 4 of these circuits then you multiply the 3A individual circuit limit by 4 or 12A. And if you stay at 30V or below, you won't overshoot the 100W limit per MOSFET or a combined 400W. If you plan on inputting more than 30V then you'll have to turn-down the pot.  Say, about half-way (1.6A per circuit) for up to 60V, etc.

This circuit can be assembled on a small perfboard; 1 IC+passives with short leads to 4 mounted MOSFETs, shunts, and a pot.  I think it's a good beginner project.

Even 100 W from a TO247 part needs a good heat sink.
Beside the total power and voltage there can also be lower SOA limit. So not every combination of voltage and current to give the 100 W may be reliable. Especially high power at a relatively high voltage can be a problem and lead to failure (some samples are OK, some may fail). The SOA specs on MOSFETs are a tricky topic - often there is no curve in the DS and even if there is may not be valid.

The resistor version has definitely some advantage in reliablity and may get a way with a smaller heat sink. However it is less flexible. With the resistors the chances are smaller to get fakes from Ebay or Banggood. So lower cost sources may be acceptable.
As a crude test an old toaster may make a cheap high power resistors, usually better than incandesant lamps.

A friend of mine mentioned using an old toaster, or even a cheapo 10 dollar new one. However, wouldn't that have a fixed ohm/power rating? Those even small toasters can use a lot of power.

[Jwillis]

Even 100 W from a TO247 part needs a good heat sink.
Beside the total power and voltage there can also be lower SOA limit. So not every combination of voltage and current to give the 100 W may be reliable. Especially high power at a relatively high voltage can be a problem and lead to failure (some samples are OK, some may fail). The SOA specs on MOSFETs are a tricky topic - often there is no curve in the DS and even if there is may not be valid.

The resistor version has definitely some advantage in reliablity and may get a way with a smaller heat sink. However it is less flexible. With the resistors the chances are smaller to get fakes from Ebay or Banggood. So lower cost sources may be acceptable.
As a crude test an old toaster may make a cheap high power resistors, usually better than incandesant lamps.

For a reliable electronic load I would only suggest Linear MOSFETS .
Theirs really only a couple manufacturers that produce reliable Linear MOSFETs . Infineon and IXYS . And they're not usually cheap . 
Nonlinear MOSFETs are tricky, as you say, because theirs no defined SOA for the DC region and can differ significantly from one chip to the next , even on same makes and models . 

Power resistors from Ebay or Banggood I wouldn't go as far as calling them fakes but they can be of low quality . I've had a couple aluminum shelled power resistors exploded out the ends when pushed close to their maximum rated power. Pretty scary events . Could be just isolated cases but I don't push them far anymore. 

[DW1961]

Power resistors from Ebay or Banggood I wouldn't go as far as calling them fakes but they can be of low quality . I've had a couple aluminum shelled power resistors exploded out the ends when pushed close to their maximum rated power. Pretty scary events . Could be just isolated cases but I don't push them far anymore.

I've read reviews of that happening, but many reviews don't report any problems. I won't be trying my luck with them. Well, maybe I will. lol

[pqass]

Even 100 W from a TO247 part needs a good heat sink.
Beside the total power and voltage there can also be lower SOA limit. So not every combination of voltage and current to give the 100 W may be reliable. Especially high power at a relatively high voltage can be a problem and lead to failure (some samples are OK, some may fail). The SOA specs on MOSFETs are a tricky topic - often there is no curve in the DS and even if there is may not be valid.

The resistor version has definitely some advantage in reliablity and may get a way with a smaller heat sink. However it is less flexible. With the resistors the chances are smaller to get fakes from Ebay or Banggood. So lower cost sources may be acceptable.
As a crude test an old toaster may make a cheap high power resistors, usually better than incandesant lamps.

For a reliable electronic load I would only suggest Linear MOSFETS .
Theirs really only a couple manufacturers that produce reliable Linear MOSFETs . Infineon and IXYS . And they're not usually cheap . 
Nonlinear MOSFETs are tricky, as you say, because theirs no defined SOA for the DC region and can differ significantly from one chip to the next , even on same makes and models . 

Power resistors from Ebay or Banggood I wouldn't go as far as calling them fakes but they can be of low quality . I've had a couple aluminum shelled power resistors exploded out the ends when pushed close to their maximum rated power. Pretty scary events . Could be just isolated cases but I don't push them far anymore.

Okay, if 100W per module is too ambitious then just use 8 parallel modules at 50W per. Surely they can do that much.  The cost per module (MOSFET, 0.25R shunt, opamp, and passives) is less than $4 (Digikey prices). Power resistors aren't cheap either.   The e-load solution is simple, repeatable, and you gain dialing-in any power you may need without thinking; no discrete resistor combinations.  What's not to love? 

Well, there is Kerry Wong's 400W/100A beast but each [linear] MOSFET is $32!   http://www.kerrywong.com/2017/01/15/a-400w-1kw-peak-100a-electronic-load-using-linear-mosfets/

I'd rather go with lot more but less stressed devices.  It's good enough for HPAK.  They use 8 modules of IRFP140's.  I'm sure they were preselected by grey-bearded virgins.  Although, I don't know what's up with the parallel IRF540!  See attachment and/or page 26 here: https://dev.xdevs.com/attachments/download/148/Agilent_N3304_component_shematic_5989-6691EN.pdf
 

[DW1961]

It's not pretty, but it works. Heat sink is waaaaaaaaaaaaaaay to small. I ran it fanless at 50 watts for about 10 minutes. I think it hit around 80C. Vendor said it was good to 130C. I'll put a fan on it tomorrow and test it with quite a bit more power. It's going to work just fine, though. I need to find a bigger sink and a housing with banana female so I can have it in a box and swap resistance easily. I'd like to have it all self contained and use a 120mm CPU fan that moves some air through the box (open on each end.). The hard part is finding parts to do that.

[Gyro]

Hopefully you've used a smear of heatsink compound underneath the resistors. With a milled / extruded surfaces, the actual metal to metal contact area is minimal, so the resistors will end up running significantly hotter than the heatsink.