PSU Problems

[P4r4gr4ph]

Hey,

 

So I recently got a new graphics card for my crappy rig and put my old graphics card on the bottom slot for video encoding and my 2nd monitor. However, when I overclocked my new card, my system would often suddenly crash with my monitor becoming a mosaic of sorts. I decided to turn down the power since I applied an extra 20% to the power limit, and my card didn't crash anymore, although it seems to be having trouble hitting my overclocked speeds. I decided to take a look at my PCIE cables, and both of them don't have a pin at the top middle hole, and there isn't a cable for that hole at the back. Either I'm dumb and PCIE cables never had a pin at the top middle or the manufacturer decided not to put one there for some reason. After all, the entire base of my PC is a crappy prebuild. I'm guessing that missing pin is what provides the extra 20% power and isn't required. After all, I'm able to use the card, so it's obviously working. Plus, I think my PSU is struggling to run everything in my rig. I'm not sure if I'm able to measure how much watts is being used from my PSU, but I'd like to know if I can measure that. I definitely need to get a new PSU, but until then, I want to figure out how to fix the issue if at all possible.

 

Thanks,

Paragraph

[KarathKasun]

3 minutes ago, P4r4gr4ph said:

Hey,

 

So I recently got a new graphics card for my crappy rig and put my old graphics card on the bottom slot for video encoding and my 2nd monitor. However, when I overclocked my new card, my system would often suddenly crash with my monitor becoming a mosaic of sorts. I decided to turn down the power since I applied an extra 20% to the power limit, and my card didn't crash anymore, although it seems to be having trouble hitting my overclocked speeds. I decided to take a look at my PCIE cables, and both of them don't have a pin at the top middle hole, and there isn't a cable for that hole at the back. Either I'm dumb and PCIE cables never had a pin at the top middle or the manufacturer decided not to put one there for some reason. After all, the entire base of my PC is a crappy prebuild. I'm guessing that missing pin is what provides the extra 20% power and isn't required. After all, I'm able to use the card, so it's obviously working. Plus, I think my PSU is struggling to run everything in my rig. I'm not sure if I'm able to measure how much watts is being used from my PSU, but I'd like to know if I can measure that. I definitely need to get a new PSU, but until then, I want to figure out how to fix the issue if at all possible.

 

Thanks,

Paragraph

Thats a sense pin, and its not required.

[seon123]

4 minutes ago, P4r4gr4ph said:

, when I overclocked my new card

That's the issue. Not the PSU. 

[KarathKasun]

Just now, seon123 said:

That's the issue. Not the PSU. 

Cant really say that, dont know what PSU and what two GPUs are being used.

[seon123]

1 minute ago, KarathKasun said:

Cant really say that, dont know what PSU and what two GPUs are being used.

Yes, we can. An issue that's caused by either an unstable GPU overclock or a faulty GPU, and that happens when a new overclock is applied. 

[P4r4gr4ph]

4 minutes ago, seon123 said:

That's the issue. Not the PSU.  

I understand how you think my OC is the issue, but I've never had a GPU straight up cause a crash without even showing a BSOD before. Every other card I've ever overclocked too aggressively before, Windows would recover the driver and revert to stock settings. Besides, I'm not sure if its me or if 1050MHz Core and 1500MHz Mem is pretty reasonable for an R9 270, and I'm not overclocking my 2nd card. If anything, I'm probably going to underclock it.

[KarathKasun]

Just now, seon123 said:

Yes, we can. An issue that's caused by either an unstable GPU overclock or a faulty GPU, and that happens when a new overclock is applied. 

Or an insufficient PSU trying to power two GPU's with a pretty big power limit boost.

[seon123]

Just now, KarathKasun said:

Or an insufficient PSU trying to power two GPU's with a pretty big power limit boost.

No. 

Tell me, how would an overloaded PSU cause artifacting? 

1 minute ago, P4r4gr4ph said:

I understand how you think my OC is the issue, but I've never had a GPU straight up cause a crash without even showing a BSOD before. Every other card I've ever overclocked too aggressively before, Windows would recover the driver and revert to stock settings. Besides, I'm not sure if its me or if 1050MHz Core and 1500MHz Mem is pretty reasonable for an R9 270, and I'm not overclocking my 2nd card. If anything, I'm probably going to underclock it.

Artifacting can happen with an unstable overclock. How far a card overclocks depends on the card you have. 

A PSU that's overloaded, and that isn't complete crap, will shut down. A crappy PSU that's overloaded will also eventually shut down, though possibly permanently. 

[r2724r16]

4 minutes ago, KarathKasun said:

Or an insufficient PSU trying to power two GPU's with a pretty big power limit boost.

The PSU would explode or burn out, it wouldn't cause artifacting.

[KarathKasun]

Just now, seon123 said:

No. 

Tell me, how would an overloaded PSU cause artifacting? 

Artifacting can happen with an unstable overclock. How far a card overclocks depends on the card you have. 

A PSU that's overloaded, and that isn't complete crap, will shut down. A crappy PSU that's overloaded will also eventually shut down, though possibly permanently. 

Overloaded PSU = voltage droop = out of spec core/vram/display controller voltage regulation.  Its pretty simple, and is known to happen.

[KarathKasun]

Just now, r2724r16 said:

The PSU would explode or burn out, it wouldn't cause artifacting.

Not true.  Voltage regulation can just go to crap and cause instability. Increased ripple/droop.

[Princess Luna]

Your artifacts is probably due to the VRAM frequency overclock. Also it'd be nice if in middle of the huge text wall of OP to at least mention the hardware.

[seon123]

Just now, KarathKasun said:

Overloaded PSU = voltage droop = out of spec core/vram/display controller voltage regulation.  Its pretty simple, and is known to happen.

With a group regulated PSU, you don't need to overload it for the voltages to go out of spec. Adding a few more watts won't change much at all. 

The system would just shut down if the voltages get too low. It would not cause artifacting. 

[P4r4gr4ph]

4 minutes ago, Princess Cadence said:

Your artifacts is probably due to the VRAM frequency overclock. Also it'd be nice if in middle of the huge text wall of OP to at least mention the hardware.

As I said, when my PC kept crashing, I turned down the power limit and left the rest of the settings the same. It stopped crashing, and although it had trouble hitting the OC settings, it's hitting my VRAM overclock. I don't see how it could be my GPU.

[KarathKasun]

2 minutes ago, seon123 said:

With a group regulated PSU, you don't need to overload it for the voltages to go out of spec. Adding a few more watts won't change much at all. 

The system would just shut down if the voltages get too low. It would not cause artifacting. 

If you are riding the edge of what the PSU is capable of, yes... a few watts can make a difference.

 

Its almost like nobody here understands that PSU's have more operational states than dead or alive.

 

Voltage regulation is much more than what you read on a multimeter.

[Princess Luna]

Just now, P4r4gr4ph said:

.

When you turn down your power limit you're also probably interfering with the clocks since that's simply how GPU Boost 3.0 works, it has to find the best balance between power, temperature and clock, when you force an overclock setting you're just telling it what you want but the Boost 3.0 is still what's in charge.

 

That been said you should use a nice benchmarking application like Heaven and find the true balance of power and temperature along side the extra mhz you want overclocking, for instance my GTX 1080 Ti can do 2002mhz on core clock on FE cooler which is known be bad, that however was only possible without maxing power limit as that caused it to use more power getting more hot and actually downclocking.

 

On your case the better cooling likely is allowing the frequency to kick in however it reached a point it turned software side unstable, so you have to find the correct balance always.

 

All that said, you still haven't said what you actually have, so we can make a better assume on whether or not your PSU is insufficient for your system.

[Guest]

 And yet we still don't know the OP's system's specs....