Hello,
One of the things that will prove to be useful to me is a resistance box. Now I know I could buy one, or a programmable load, but I'd rather make my own. I'll learn more this way.

Now I have two options for creating one.
1: Use resistors.
2: Use a MOSFET (or maybe an IGBT).

Now for a 1/4w box, resistors would be the obvious choice, but if you want something that can handle more power, say 100w, then resistors require heatsinks and become quite expensive. That is not to say that I'm actually planning to throw 100W at a resistance box, rather, I'm fairly certain that is the upper limit of what I'd use it for throughout my lifetime. Even at half that power level, resistors require heatsinks and are themselves expensive.
Not that it's the ideal part, but the ever popular IRFZ44N can dissipate 110W, theoretically. The problem is dissipating the heat.


So, I started reading up on the various heatsinks. I have learned that fin length (often the height), is very important for cooling.
A good example of a heatsink for MOSFETs is the Ohmite brand RA and FA series (datasheet). But they're rather expensive compared to plain stock extruded heatsinks you can find online. For example (HxWxD/L), XXx34x12mm TO-220/TO-247, 36x69xXXXmm, 25x60x150mm, and 6x20xXXXmm.
(XXX is typically 50, 100, 150, 200, or 300, millimeters with the noted exception of the TO-220/TO-247 model listed above which comes in several sizes. I've noted sizes of 38, and 70 millimeters, as they correspond roughly to the size of the Ohmite variety, but the generics are only half as deep and about 3/4 as wide).


Which brings me to my question, has anyone characterized the thermal performance of no-name heatsinks? I did try searching online, but I'm just not getting anything useful.

Thanks!

 

Resistors with Aluminum extrusions as housing have become quite affordable while good Mosfets are trending in the other direction.

You can find the aluminum housed resistors for a few bucks at 50w each. Chassis mount so you can put them in an aluminum box or plate for more heat transfer.

https://www.mouser.com/ProductDetail/ARCOL-Ohmite/HS50-10R-F?qs=bvCpHSZmQjmRqRZ14Wtj0Q==

 

A given heatsink does not have a single number that you can use to determine the suitability for a given application. Heat is dissipated by three different processes. They are:

1. Radiation - black bodies work best
2. Conduction - low thermal resistance works best since you get a smaller temperature rise per watt
3. Convection - heat sinks with fins are effective in removing heat over which there is a "fluid" flow

Most effective strategies use all three.

 

One of the things that will prove to be useful to me is a resistance box. Now I know I could buy one, or a programmable load, but I'd rather make my own. I'll learn more this way.

So you want a resistance box commonly called a decade resistor box or resistor box or do you want a load bank as in a resistive load used for example to test power supplies. You seem to be describing a load bank. If you want a load bank I agree with the suggestion of resistors with aluminum housings. I have used them and just mounted them on sheet aluminum. Whatever you choose as to a load bank a cooling fan or fans will be your new best friend. Loads can be switched In/Out using MOSFETS.

Ron

 

Papabravo I'm planning on using conduction and convection. I'll use radiation if I can. It depends on what's available and at what cost.

 

No-Name electronic parts companies make junk but have lots of profit.
Reputable Name-Brand companies make good well-tested products at fair prices and reasonable profit.

 

So you want a resistance box commonly called a decade resistor box or resistor box or do you want a load bank as in a resistive load used for example to test power supplies.

I'm looking for a resistive load. I'm sorry if this sounds ignorant, but I could not find any different in the descriptions between a load bank and a resistor box. Both can contain resistors. Both load down the circuit under test.

Loads can be switched In/Out using MOSFETS.

I was thinking of using manual switches if I went that route. For a MOSFET load, I'd probably be able to use a potentiometer to adjust how much resistance the MOSFETs are creating.

 

No-Name electronic parts companies make junk but have lots of profit.
Reputable Name-Brand companies make good well-tested products at fair prices and reasonable profit.

Erm... name-brand motherboard heatsinks on the CPU power rails are worse (they have less surface area and fewer fins), than the products I have referred to above -- all of them (for their size, of course).

 

I'm looking for a resistive load. I'm sorry if this sounds ignorant, but I could not find any different in the descriptions between a load bank and a resistor box. Both can contain resistors. Both load down the circuit under test.


I was thinking of using manual switches if I went that route. For a MOSFET load, I'd probably be able to use a potentiometer to adjust how much resistance the MOSFETs are creating.

Well alrighty then. You want a load bank and yes, you can switch loads In and Out using MOSFETS or switches. The only consideration is when switching in a load make sure the switch, relay or MOSFET can handle the current. Years ago, maybe 20 years ago I did a prototype load bank for testing ATX form factor (PC) power supplies. What remains of an old inactive web page can be found here. Thye old schematics can be found here. The design was just a prototype and I used what I had for resistors which were Ohmite Stackohm types. Keep in mind this was over 20 years ago. Note the fan for cooling the resistors. The later design used the mentioned aluminum housing resistors on aluminum 1/4" sheets also with forced air cooling. The final version also had some variable power resistors to trim the current loads. Anyway it gives you some idea about a load bank.

Ron