Are both of my AX1600i bad?!

[4EvrYng]

I own two AX1600i that are in two systems. First one has been purchased approximately year ago and second one approximately three months ago.

 

Today I’ve looked at them using latest version of iCue and found two things that concern me:

 

1. Voltages don’t seem to be what they should be with these units.

 

On first one I’m getting 12.06V, 4.97V, and 3.34V which is a difference of 0.5%, 0.6%, and 1.2% from perfect values respectively.

 

However, on second one I’m getting 12.19V, 4.94V, and 3.28V which is 1.58%, 1.2% and 0.61% difference.

 

In other words values for first unit seem to be where reviews indicate they should be (even though not as good as review ones) -BUT- values for second unit, especially 12V one, seem to be much worse than values for first unit and values in reviews.

 

2. Look at rest of values showed that unit #2 (one with higher voltages) displays current on 24-pin and SATA rails as it should be (3A on each).

 

-However- current for those rails on first unit is displayed as -zero- which is impossible because a) if current was really zero on 24-pin this system would not be running at all, and b) I do have devices on SATA rail up and running just fine.

 

Does anyone know what is going on, please? I’m especially concerned about 12V being quite higher than what I expected.

[LeDoyen]

those voltages are fine.

It all comes down to cable resistance, connexion resistance, and stuff like that.

 

the ATX specs are +-5% so you're well within spec.

Also remember, the PSU uses remote sensing wires to measure the voltage at the motherboard end, and adjusts the supply accordingly.

 

So to give it 12.00V at the connector, it may have to push 12.1 or 12.2V at the PSU end to account for voltage drop.

 

If you use different cables, sleeved extensions etc.. it will make readings differ even more.

Regarding the amp reading tho, no idea :)

[4EvrYng]

the ATX specs are +-5% so you're well within spec.

 

First thank you for your reply!

 

I know I'm within ATX specs but that is not my concern. My concern is that I don't seem to be within level of performance I should be getting from this particular supply. If I wanted to be just within ATX specs almost any supply on the market would give me that.

 

Also remember, the PSU uses remote sensing wires to measure the voltage at the motherboard end, and adjusts the supply accordingly.

 

 

So to give it 12.00V at the connector, it may have to push 12.1 or 12.2V at the PSU end to account for voltage drop.

 

Yes, and I've already checked voltages on mb side and according to both its BIOS and HWInfo64 voltages on mb side are practically same what iCue was reading on PSU side.

 

If you use different cables, sleeved extensions etc.. it will make readings differ even more.

 

I'm using Corsair's cables that came with them, no extensions.

[LeDoyen]

Yes, and I've already checked voltages on mb side and according to both its BIOS and HWInfo64 voltages on mb side are practically same what iCue was reading on PSU side.

 

Luck of the draw infortunately ^^'

On board voltage readouts are indicative only really. My motherboard reads higher voltages than what's supplied by the PSU. the GPU reads higher on some pins, lower on others.. but all 12V come from the same regulator, so they should all be identical.

Only a multimeter will tell you what you have.

But you may get an approximation looking at all the 12V readouts you can put your hands on (Mobo, GPU has several, etc.. ).

[4EvrYng]

Luck of the draw unfortunately ^^'

 

Thank you again for your reply and help!

 

I agree that “luck of the draw” variation can cause any equipment to have difference in the performance and that that is acceptable as long as “worse one” is still within tolerances for that particular model. It is just that all these reviews together left me with impression that voltage regulation of AX1600i is much MUCH better than what I’m seeing on my second unit.

 

If I am interpreting those reviews correctly Tom’s Hardware said 0.07% on 12V rail, Anandtech 0.2%, TechPowerUp 0.09%, JonnyGuru 0.25%, KitGuru 0.75%, … yet I’m told I have 1.58% on my second unit which is twice worse than worst of the reviews I came across and five times higher than their average.

 

On board voltage readouts are indicative only really. My motherboard reads higher voltages than what's supplied by the PSU. the GPU reads higher on some pins, lower on others.. but all 12V come from the same regulator, so they should all be identical. Only a multimeter will tell you what you have.

 

Makes sense. Unfortunately multimer is well outside my skillset, otherwise I would’ve done it.

 

But you may get an approximation looking at all the 12V readouts you can put your hands on (Mobo, GPU has several, etc.. ).

 

Excellent suggestion! I just looked and I see what you are pointing out. So, if I am putting 2+2 correctly together:

 

1. It doesn’t matter what iCue says voltages are on PSU side, what matters is accuracy of voltages on mb side.

 

2. In absence of exact multimeter I should log all possible 12V/5V/3.3V readouts over a period of time, average them, and that should give me approximate idea what exactly my mb is getting on average.

 

I have one more question, please, about something I noticed while looking at all these voltages:

 

HWInfo64 reports that voltages sometimes drop significantly on some of sensors. For example, 5V Standby dropping to 1.872V, 12V dropping to 7.5V … but only on some of 12V sensors, not on all of them.

 

What I might be looking at? Something that is (not) normal or “just” incorrect sensor readouts?

[LeDoyen]

incorrect readouts.

HWinfo and iCUE should no be left running together. they both poll data from Commander pro or the PSU at the same time and create bogus readouts.

They can even mess up fan speed because the CoPro may have wrong temperature readings.

It's always better to close one before using the other.

 

The motherboard, or GPU sensors are not calibrated, so you have no way of knowing if they are on the mark or not.

The only really useful way to use them is to see how they average under no load.. then it can give you an idea of how much voltage drops under load. You can compare deltas, but not absolute values.

For that reason, those PSU reviews, if they are at least somewhat serious, have been made using meters, and not motherboard sensors, not even PSU sensors.

So, basically, i fear you may be comparing apples and oranges, comparing values between HWinfo/iCUE and review material.

 

And my "luck of the draw" line was about motherboard voltage readings ^^ you can take two motherboards of the same model, put the same hardware, same PSU, and get different readings precisely because those sensors are dodgy :)

 

They are still useful tho :)

When you benchmark your GPU and see that its 12V reading drops to 11.5, well it may indicate the extension sleeves you were using are crap.. It helped me get rid of mine and get good sleeved cables. The delta are useable, not the absolute values.

[4EvrYng]

HWinfo and iCUE should no be left running together.

 

I’m aware of their possible interference so I haven’t been running them both at the same time. I actually have two separate Win10 installations to test this, one with just iCue, other with just HWInfo64, in order to avoid possible interference.

 

The motherboard, or GPU sensors are not calibrated, so you have no way of knowing if they are on the mark or not.

…

… my "luck of the draw" line was about motherboard voltage readings ^^ you can take two motherboards of the same model, put the same hardware, same PSU, and get different readings precisely because those sensors are dodgy :)

…

For that reason, those PSU reviews, if they are at least somewhat serious, have been made using meters, and not motherboard sensors, not even PSU sensors.

…

So, basically, i fear you may be comparing apples and oranges, comparing values between HWinfo/iCUE and review material.

 

Now that I’m reading what you said it is making me think, makes sense and makes things clear.

 

Thought did cross my mind to try to switch cables and supply around to see is it something about cables or possible difference in load but ended up not doing it cause it would be too much work and time.

 

However, if I did it and found situation is still same now that I think about it that still wouldn’t clearly implicate supply. Like you are pointing out sensors are not calibrated and, now that I think about it they also have +/- accuracy that doesn’t have to be always consistent. Discarding data for sensor that shows weird huge drops (weird because I don’t think it is possible for 12V rail to really drop to 7.5V without system crashing and without that drop showing on all other 12V sensors) and averaging all others might give me better data but even that is not to be trusted as it is entirely possible majority of sensors on one mb ended up on different side of accuracy range than on the other mb.

 

So yeah, I too have formed a feeling this has, in the lack of the better method, ended up an apples-to-oranges comparison.

 

They are still useful tho :)

…

The only really useful way to use them is to see how they average under no load.. then it can give you an idea of how much voltage drops under load. You can compare deltas, but not absolute values.

…

The delta are useable, not the absolute values.

…

When you benchmark your GPU and see that its 12V reading drops to 11.5, well it may indicate the extension sleeves you were using are crap.

 

That is a great troubleshooting tip I will have to keep in mind if I ever need to check on voltages! Thank you SO MUCH once again, you’ve been of incredible help and I’ve learned so much from your replies!