Because so many people that were sleeving their own cables were complaining about the double wires on the 24-pin, I decided to do an experiment.

I told everyone that the sense wires were not required for the PSU to work properly, but I was shot down as "not knowing what I was talking about". ¯\_(ツ)_/¯

Fact of the matter is, a good PSU should deliver nominal voltage on all rails. Sense wires will tell the PSU if the voltage at the end of these wire are below ideal. And while I agree sense wires do help, the lack of sense wires do not hurt to any extent.

So here's what I did:

Not pretty, but a quick and dirty "mod". I essentially cut every sense pin and taped it off.

Yes. That''s painter's tape. And I think that might be a cat hair stuck to one of the pieces of tape too.

That is a fairly standard 650mm long cable. Obviously, the longer the cable the more voltage drop, the shorter the cable the less. Ohm's law, you know. The PSU I'm using is a Corsair RM650. The loads are done by a Chroma 8000.

Ok... So here's the before and after.....

80 PLUS 100% load test:

Ok? So that's the load for each rail.

By the way, 12V1 is the 24-pin, 12V2 is the EPS12V and 12V3, 4 and 5 are PCIe.

Now the results.... With sense wires:

Without sense wires:

So definitely a drop in voltages. But very, very minor and still within spec by a long shot (ok.. maybe not the +3.3V. But who really uses that and definitely who puts a 12.78A load on it!?).

The +12V, on average was 12.037 volts. Without the sense wires, it sees a drop to 11.95 volts. 0.087 volts.

Now I'm just going to do the +12V full load test with and without sense wires. That load table looks like this:

First, here's the results with the +12V sense and ground sense wires in place:

And now here are the results with the cut sense wires:

Again, we see a drop, but not a significant one. The +12V on average was at 12.05 and dropped to 11.946. A difference of 0.104.

To make this all easier to read, I've made a table with all of the loads in it and the results:

These final tests are more realistic. The above tests are static loads, like what we see most reviewers do. This next test takes the PSU from 0 to full load and a snap shot is taken of the voltages before the PSU can fully recover. This is actually more realistic as the loads in your PC are not static.

First, with the 80 PLUS 100% load:

Vout-1 is the no load voltage and Vout-2 is with the load. We see almost the same as our static load with a little more of a drop on the PCIe cables. Here's what that drop looks in numbers:

Now we use the cable with the cut sense wires:

Quite a bit of a difference there:

Now let's do the +12V full load again, back to the cable with the sense wires:

Here's the math:

And now without the sense wires:

Again, a drop in voltage, as is to be expected:

So definitely a drop in voltage, but nothing dangerous or anything that would prevent a PC from working properly.

UPDATE: Since I made this page, I posted a link to it on Reddit to help other people realize that you can make your own custom cables without sense wires. Someone pointed out that the SF750 DOES require sense wires and that my RM650 example doesn't apply. When I told them the SF750 works the same way, my post was deleted as being "not safe". LOL!

To prove my point, I used the same cables on an SF750 and did the tests again. Funnily, the cable is 650mm, which is WAY LONGER than the cables that come with the SF. So this is EXTREME WORST CASE SCENARIO... and yet...

This is the 100% load test as performed by 80 PLUS:

Here are the results with the cable with the sense wires:

And here are the results of the PSU with the cable with the cut sense wires:

Here is the +12V rail full load parameters:

And here are the results with the regular 650mm cable:

And here are the results with the sense wires removed:

That's right. We see a voltage INCREASE. But why?

Different PSUs with different designs. The RM650, like most PSUs, have multiple voltage sense points within the PSU. So there's ALWAYS a reading being fedback to the PSU. The smaller SF Series relies completely on the reading from the sense wires and sees no reading, so assumes there's a drop and automatically compensates, even though it doesn't have to. Since there's a fail safe that limits how high the voltage can be raised, it never becomes an unsafe voltage.

To help with the legibility of the results, I've made a table just like the one I made for SF750:

Now, we do see the sense wires make a difference when we spike a heavy load and take a snapshot immediately after the load is applied. Just as we did with the RM650 above, we start at zero load, spike it with a full load and immediately grab a voltage measurement before the PSU has time to "recover".

As before, with the 80 PLUS 100% load:

And the drop in voltages:

And here's the voltage results with the same load with the cable without the sense wires:

First thing you'll note is that the zero load voltages are higher without the sense wires than they were with the sense wires. And while the voltages are higher than before once the load is applied, the difference between load and no load is much greater:

Back to the cable with the sense wires in-tact; here's the +12V full load results:

And the drop in voltage we see from above:

Obviously more of a drop on the +12V rails since that's where all of the load is....

Now let's cut those sense wires (I know... they're already cut. I use the same cable over and over) and do it all over again!

Resulting in a difference of:

Pretty significant, but still within spec and we should also take into consideration the length of the cable.

So, in conclusion: Ideally, yes. Leave the sense wires in there. But if you're sleeving your own cables and you're finding double wires are making your cables more difficult or that they're looking uglier than you'd like, just remember this is an option. You don't want your cable to end up looking like this: