PSU: Measuring Current From SATA Connection?

[DW1961]

How would I go about measuring current from the SATA power connection from the PSU? I need to make sure I'm not overloading the 4.5Amp connectors.

Is there a pass through SATA connector available I could use to connect a DMM so I can see the current?

[mariush]

you could just get a molex -> sata adapter cable or a sata -> 3-4 sata extension cable , cut the 5v and/or 12v wires and solder in current sense resistors (ex 0.1 ohm or something like that)

or just cut the wire of your atx power supply if you don't care about it.

[Berni]

If you have a current clamp you can just put the wire you want to measure trough it.

But for splicing into things i like to cut up various random adapters that i might have laying around. Those molex to sata power adapters are pretty common, there are some sata power splitter adapters, sometimes fan power adapters have a passtrough design having both a female and male molex/sata on it, there are some molex to 6 pin PCIe power adapter cables (for graphics cards)...etc. Old ATX power supplies can also be cannibalized for the connectors.I keep a box of random computer cables for cases like these for when i need to hack something together.

But keep in mind that adding a multimeter in series introduces extra voltage drop. Probably fine for the 12V rail but might cause too much drop on 5V or 3.3V rails if the currents are high.

[David Hess]

Is there a pass through SATA connector available I could use to connect a DMM so I can see the current?

There are SATA power splitters which could be used as a pass through connector.  The have a female SATA power connector on one side and a male SATA power connector on the other.

[mariush]

But keep in mind that adding a multimeter in series introduces extra voltage drop. Probably fine for the 12V rail but might cause too much drop on 5V or 3.3V rails if the currents are high.

IF you put the multimeter in series yes, you'll have to deal with the multimeter's burden voltage and the resistance of the probes.

BUT if you solder a low value resistor with a good temperature coefficient so the value won't change significantly as it heats up from all the power going through it, then you can use the multimeter in voltage mode and measure the drop across the resistor... and you can also measure the actual voltage after the resistor with a second multimeter and then you can multiply actual voltage with current and get your power.

For example, a 1 ohm resistor will cause a 1v drop across it at 1A  of current.
But you can go down to 0.1 ohm and then you'll only have 0.1v drop across the resistor at 1A of current... and you'd just have to multiply by 10 the voltage measured across the resistor to get your current value.

It's the technique Dave's uCurrent uses ... only that device uses much lower value resistors and  a precision opamp to amplify the voltage drop across the resistor before it's measured by a multimeter

IF your device needs 5v, you can use a linear power supply, set the voltage to something like 5.2v and measure your device - I doubt your device must have 5v and really can't deal with +/- 0.25v

[David Hess]

IF you put the multimeter in series yes, you'll have to deal with the multimeter's burden voltage and the resistance of the probes.

A multimeter would always have to be used on its 10 amp range in this application but that should provide plenty of resolution.

When I have done this in the past, I added banana plugs to the adapter so it could be plugged directly into the multimeter without using test leads.

[DW1961]

All great ideas.

I have 5V ARGB LEDS in my computer that  I want to measure. Currently, they are connected to an SATA ARGB hub and the hub is powered by the SATA PSU connection. The data cable comes from the motherboard. That way I can use the software to control the LEDs.

Incidentally, there is a power cable that also comes from the MB. I was getting some weird feed back through the MB or PSU SATA connection that would slowly turn either the PSUs fan or two of the 12V fans, very slowly, when the computer was off. I found that the phenomenon is the way Gigabyte does the ARGB circuitry. It's kind of a feed back from the hub that is powered by the SATA connection.

I traced the power feedback in off mode to the ARGB power lead connected to the MB.

The solution was simple. I just cut the ARGB wire's power connection from the motherboard. That's how I would prefer it anyway, since the MB ARGB riser has it's own set of ARGB LEDs and they are and should be separate from the SATA powered hub.

Anyway, after reading your comments, I think all I need to do is plug the SATA connection into my bench power supply, then read the bench power supply's amp reading. The power can't get to the MB from the bench supply because the ARGB connection's power wire is cut. I think that would be pretty safe.

In short, I'll be using the MB's ARGB controller for the data cable, in order to turn the lights on,  and the bench PS to power them through the SATA connection. Then I can just read the current from the bench PS's read out.

The best solution is to have an ARGB controller that I can power directly from the bench supply. I do have one, but it is current limiting, so I can't use it to measure maximum current.

It has been impossible for me to find a simple ARGB 5V controller that does not limit current. Otherwise, I'd use that connected to the bench supply to measure current, leaving  the MB out of the loop (which is my preference).

Anyone see any problems doing it like that?

[Berni]

This is more of a problem that you are shoving power back into what is meant to be an output.

A most of semiconductor switching technology is designed to only pass current in one direction, so you get weird behavior once you try to shove power the other way. Also chips usually have ESD protection diodes between the pins and power rails, for this reason its a bad idea to apply a voltage to the pin of a chip that is powered down.

It's usually pretty easy to fell if a connector is overloaded because it will be noticeably warmer compared to a non loaded connector (easier to see on a thermal camera, but you can feel it by hand too)

[DW1961]

This is more of a problem that you are shoving power back into what is meant to be an output.

A most of semiconductor switching technology is designed to only pass current in one direction, so you get weird behavior once you try to shove power the other way. Also chips usually have ESD protection diodes between the pins and power rails, for this reason its a bad idea to apply a voltage to the pin of a chip that is powered down.

It's usually pretty easy to fell if a connector is overloaded because it will be noticeably warmer compared to a non loaded connector (easier to see on a thermal camera, but you can feel it by hand too)

You would think Gigabyte would use a diode or something to prevent back feed. ASUS doesn't have that problem. I could have just soldered in a 12V rectifier inline and problem solved. Also, the hub itself could have used a rectifier. Interestingly enough, I contacted Phantek's tech department and explained the above. They said that's why their ARGb hub won't work with Gigabyte MBs. They had the same back feed problem. ---This only happens when the computer is off.--- After I figured out simply cutting the power wire solved the problem, I told Phanteks. You would think they would solve that problem with a rectifier diode so they can sell more units. Anyway, who knows why their engineers didn't catch that and solve it. It's quite easy.

Anyway, I think I can just plug in an SATA connection with the 5V side energized +/- and run the hub and all of the ARGB lights like that. I'll just cannibalize an SATA cable end and alligator clip it to the Bench PS at 5V and see what I get.

It would be much easier and safer to have an ARGB 5V controller that DOES NOT limit current. Then I could take the MB out of the loop too (would not need its controller). If anyone knows where I can get one, lease let me know.

[Berni]

It sounds like you are trying to do something that is not a actual supported feature of it.

I don't see why you would need to have the LED controller hub connected and powered from both the motherboard and SATA. And even then SATA is supposed to run from the same power rails that the motherboard runs from, so they should always both have power or both not have power simultaneously.

But i have no experience with this PC RGB lighting, I was never a fan of it, so none of my computers ever had it. Maybe they have some odd wiring conventions that i am not familiar of.

[DW1961]

It sounds like you are trying to do something that is not a actual supported feature of it.

I don't see why you would need to have the LED controller hub connected and powered from both the motherboard and SATA. And even then SATA is supposed to run from the same power rails that the motherboard runs from, so they should always both have power or both not have power simultaneously.

But i have no experience with this PC RGB lighting, I was never a fan of it, so none of my computers ever had it. Maybe they have some odd wiring conventions that i am not familiar of.

I just cut the MB's power wire to the ARGB hub. The ARGB hub is already powered by the SATA PSU connector. So, the only difference is that instead of the hub getting both the SATA power and the MB's ARGB riser power, the hub now only gets SATA power, which I am more ok with than the the way around. That, because cutting the MB power wire simply isolates the MB from the SATA ARGB hub, which is connected to the PSU SATA connection. In short, there is now no way for power to travel back trough the hub into the MB. I'm pretty sure soldering in a rectifier to the MB\s ARGB power wire going from the motherboard to the ARGB hub would do the same thing, while allowing the MB power to integrate into the hub along with SATA power and bringing up total current to 6.5 Amps (MB ARGB riser is rated at 2 Amps). In other words, the devices plugged into the ARGB hub would share both the MB and PSU's power.

The reason I like my solution better is that the MB will keep giving it's ARGB riser power until it melts the FETS. MB manufacturers haven't plugged that hole in ARGB/RGB yet. Back in the say, the same thing was the case with fan risers, where you could over load your fan risers and smoke them. From what I've researched, today MB's protect their fan risers by limiting current.

So to answer your question about powering from the MB and the PSU, I agree! I think that is a bad design on the ARGB hub maker's part. I solved it by cutting the power wire from the MB, while leaving the ground and digital signal wire intact.

I'd still like to have an external ARGB controller that doesn't limit current.