Jump to content

Can someone explain OCP to me?

Matsozetex

I've been trying to understand how OCP works, I've read through JG's articles and Toms PSUs 101 but still have trouble grasping the logic of it. For example, the Corsair CX650M has 54A on its single 12v rail but wouldn't the OCP trip at 20A (240VA at 12v) or does OCP trigger when someone is trying to pull more power than a cable is specified to handle (i.e An 8 pin from the PSU to a splitter in a 1080ti). Do single rail PSUs even need OCP?

 

Help. I am confused, complicated and detailed answers are welcome.

Link to comment
Share on other sites

Link to post
Share on other sites

by how does OCP work, you mean the logic behind the way of operation, or the electrical components used to get it to work?

CPU: i7-2600K 4751MHz 1.44V (software) --> 1.47V at the back of the socket Motherboard: Asrock Z77 Extreme4 (BCLK: 103.3MHz) CPU Cooler: Noctua NH-D15 RAM: Adata XPG 2x8GB DDR3 (XMP: 2133MHz 10-11-11-30 CR2, custom: 2203MHz 10-11-10-26 CR1 tRFC:230 tREFI:14000) GPU: Asus GTX 1070 Dual (Super Jetstream vbios, +70(2025-2088MHz)/+400(8.8Gbps)) SSD: Samsung 840 Pro 256GB (main boot drive), Transcend SSD370 128GB PSU: Seasonic X-660 80+ Gold Case: Antec P110 Silent, 5 intakes 1 exhaust Monitor: AOC G2460PF 1080p 144Hz (150Hz max w/ DP, 121Hz max w/ HDMI) TN panel Keyboard: Logitech G610 Orion (Cherry MX Blue) with SteelSeries Apex M260 keycaps Mouse: BenQ Zowie FK1

 

Model: HP Omen 17 17-an110ca CPU: i7-8750H (0.125V core & cache, 50mV SA undervolt) GPU: GTX 1060 6GB Mobile (+80/+450, 1650MHz~1750MHz 0.78V~0.85V) RAM: 8+8GB DDR4-2400 18-17-17-39 2T Storage: HP EX920 1TB PCIe x4 M.2 SSD + Crucial MX500 1TB 2.5" SATA SSD, 128GB Toshiba PCIe x2 M.2 SSD (KBG30ZMV128G) gone cooking externally, 1TB Seagate 7200RPM 2.5" HDD (ST1000LM049-2GH172) left outside Monitor: 1080p 126Hz IPS G-sync

 

Desktop benching:

Cinebench R15 Single thread:168 Multi-thread: 833 

SuperPi (v1.5 from Techpowerup, PI value output) 16K: 0.100s 1M: 8.255s 32M: 7m 45.93s

Link to comment
Share on other sites

Link to post
Share on other sites

1 minute ago, Jurrunio said:

by how does OCP work, you mean the logic behind the way of operation, or the electrical components used to get it to work?

The logic behind how it operates, triggers and some example scenarios please.

Link to comment
Share on other sites

Link to post
Share on other sites

1 minute ago, Matsozetex said:

The logic behind how it operates, triggers and some example scenarios please.

OCP is rail-sensitive and rail-independent, so you get multiple shunts for OCP. Their job is simple: the rail they monitor exceeds current draw limit, they cut the power (of the whole PSU, not just the rail exceeding the limit). As modern independently regulated PSUs can do nearly 100% (if not exactly 100%) output on the 12v rail, some single 12v rail PSUs omit OCP on the 12V as OPP trips when 12v OCP is supposed to trip anyway (and of course in more situations than that). The number of cables plugged in don't bother OCP, it can't detect that in the first place.

 

In contrast, OPP is incapable of monitoring independent rail power draw as it usually sits on the post-PFC stage, in other words it has no idea you're pulling 50A from the 5V or the 12V in case of the CX650M. That's why you still need OCP. Some PSUs have larger sum of 12v rating and 5v + 3.3v rating than the power rating of the entire unit, that's when you need OPP.

CPU: i7-2600K 4751MHz 1.44V (software) --> 1.47V at the back of the socket Motherboard: Asrock Z77 Extreme4 (BCLK: 103.3MHz) CPU Cooler: Noctua NH-D15 RAM: Adata XPG 2x8GB DDR3 (XMP: 2133MHz 10-11-11-30 CR2, custom: 2203MHz 10-11-10-26 CR1 tRFC:230 tREFI:14000) GPU: Asus GTX 1070 Dual (Super Jetstream vbios, +70(2025-2088MHz)/+400(8.8Gbps)) SSD: Samsung 840 Pro 256GB (main boot drive), Transcend SSD370 128GB PSU: Seasonic X-660 80+ Gold Case: Antec P110 Silent, 5 intakes 1 exhaust Monitor: AOC G2460PF 1080p 144Hz (150Hz max w/ DP, 121Hz max w/ HDMI) TN panel Keyboard: Logitech G610 Orion (Cherry MX Blue) with SteelSeries Apex M260 keycaps Mouse: BenQ Zowie FK1

 

Model: HP Omen 17 17-an110ca CPU: i7-8750H (0.125V core & cache, 50mV SA undervolt) GPU: GTX 1060 6GB Mobile (+80/+450, 1650MHz~1750MHz 0.78V~0.85V) RAM: 8+8GB DDR4-2400 18-17-17-39 2T Storage: HP EX920 1TB PCIe x4 M.2 SSD + Crucial MX500 1TB 2.5" SATA SSD, 128GB Toshiba PCIe x2 M.2 SSD (KBG30ZMV128G) gone cooking externally, 1TB Seagate 7200RPM 2.5" HDD (ST1000LM049-2GH172) left outside Monitor: 1080p 126Hz IPS G-sync

 

Desktop benching:

Cinebench R15 Single thread:168 Multi-thread: 833 

SuperPi (v1.5 from Techpowerup, PI value output) 16K: 0.100s 1M: 8.255s 32M: 7m 45.93s

Link to comment
Share on other sites

Link to post
Share on other sites

3 minutes ago, Matsozetex said:

Do single rail PSUs even need OCP?

Single rail PSUs typically use OPP (Over Power Protection) rather than OCP on the 12V rail.

 

26 minutes ago, Matsozetex said:

but wouldn't the OCP trip at 20A (240VA at 12v)

Uhhh? Not sure what you mean here? I think you might be confusing AC input for DC output? OCP is on the output of the unit (what it supplies to the system), not the input from the wall.
 

4 minutes ago, Matsozetex said:

For example, the Corsair CX650M has 54A on its single 12v rail but wouldn't the OCP trip at 20A (240VA at 12v) or does OCP trigger when someone is trying to pull more power than a cable is specified to handle (i.e An 8 pin from the PSU to a splitter in a 1080ti).

It will trip when whatever is connected to the rail is trying to draw more current than whatever the trigger point is set for OCP on that particular unit.

CX650M is a poor example since it is a single rail PSU and uses OPP on 12V rather than OCP... But here is the tripping points according to the Tom's Hardware article by Aris you linked to. It will trip OPP and shut down when it exceeds the ~760W limit.

Quote
OCP 12V: -
5V: 32.5 A (130%)
3.3V: 32 A (128%)
5VSB: 5.8 A (193.3%), 78.48 mV Ripple
OPP 762.51 W (117.3%)

 



In comparison looking at a multi-rail PSU with OCP on 12V, it will trip and shut down when it reaches the current set to trigger it on that specific rail.
https://www.tomshardware.com/reviews/be-quiet-straight-power-11-1000w-psu,5465-6.html
Quote
Protection Features
OCP 12V1,2: 30.6A (139.1%), 11.85V
12V3,4: >40.75A (135.8%), 11.75V
5V: 40.1A (160.4%), 4.989V
3.3V: 29.85A (154.8%), 3.3V
5VSB: 4A (133.3%), 5.006V
OPP 1228.13W (122.8%)


So in that case on 12V1 & 12V2 rails it will trigger OCP at 30.6A. 12V3 & 12V4 rails will trip at 40.75A.
Compared to its specs on the box which is 22A for 12V1, 12V2  and 30A for 12V3, 12V4.
image.png.4751cf6f65b7efc960a2e357f6d01fe3.png


In the event that you have say, a faulty graphics card that for whatever reason starts drawing 70A through the PCIe connector, the multi-rail PSU should trigger OCP and shut down. The single rail PSU will only shut down when OPP is tripped, which can be a much higher value and result in the full 70A being delivered to the card without shutting down, which can lead to further damage.

At least, that's as much as what I think I know about OCP :D I'm sure Stefan, JG, Seon, etc will tell me I'm and idiot and completely wrong, and that's okay cause I'll probably learn something from it.

CPU: Intel i7 6700k  | Motherboard: Gigabyte Z170x Gaming 5 | RAM: 2x16GB 3000MHz Corsair Vengeance LPX | GPU: Gigabyte Aorus GTX 1080ti | PSU: Corsair RM750x (2018) | Case: BeQuiet SilentBase 800 | Cooler: Arctic Freezer 34 eSports | SSD: Samsung 970 Evo 500GB + Samsung 840 500GB + Crucial MX500 2TB | Monitor: Acer Predator XB271HU + Samsung BX2450

Link to comment
Share on other sites

Link to post
Share on other sites

2 minutes ago, Jurrunio said:

The number of cables plugged in don't bother OCP, it can't detect that in the first place.

 

So say if some 'smart person' tried to power a Vega 64 with a single 8 pin using splitters the wires would melt/be damaged and the power supply wouldn't to jack?

Link to comment
Share on other sites

Link to post
Share on other sites

1 minute ago, Matsozetex said:

So say if some 'smart person' tried to power a Vega 64 with a single 8 pin using splitters the wires would melt/be damaged and the power supply wouldn't to jack?

the power supply won't, but in this case the graphics card could detect the below-standard voltage and refuse to turn on.

CPU: i7-2600K 4751MHz 1.44V (software) --> 1.47V at the back of the socket Motherboard: Asrock Z77 Extreme4 (BCLK: 103.3MHz) CPU Cooler: Noctua NH-D15 RAM: Adata XPG 2x8GB DDR3 (XMP: 2133MHz 10-11-11-30 CR2, custom: 2203MHz 10-11-10-26 CR1 tRFC:230 tREFI:14000) GPU: Asus GTX 1070 Dual (Super Jetstream vbios, +70(2025-2088MHz)/+400(8.8Gbps)) SSD: Samsung 840 Pro 256GB (main boot drive), Transcend SSD370 128GB PSU: Seasonic X-660 80+ Gold Case: Antec P110 Silent, 5 intakes 1 exhaust Monitor: AOC G2460PF 1080p 144Hz (150Hz max w/ DP, 121Hz max w/ HDMI) TN panel Keyboard: Logitech G610 Orion (Cherry MX Blue) with SteelSeries Apex M260 keycaps Mouse: BenQ Zowie FK1

 

Model: HP Omen 17 17-an110ca CPU: i7-8750H (0.125V core & cache, 50mV SA undervolt) GPU: GTX 1060 6GB Mobile (+80/+450, 1650MHz~1750MHz 0.78V~0.85V) RAM: 8+8GB DDR4-2400 18-17-17-39 2T Storage: HP EX920 1TB PCIe x4 M.2 SSD + Crucial MX500 1TB 2.5" SATA SSD, 128GB Toshiba PCIe x2 M.2 SSD (KBG30ZMV128G) gone cooking externally, 1TB Seagate 7200RPM 2.5" HDD (ST1000LM049-2GH172) left outside Monitor: 1080p 126Hz IPS G-sync

 

Desktop benching:

Cinebench R15 Single thread:168 Multi-thread: 833 

SuperPi (v1.5 from Techpowerup, PI value output) 16K: 0.100s 1M: 8.255s 32M: 7m 45.93s

Link to comment
Share on other sites

Link to post
Share on other sites

3 minutes ago, Spotty said:

In the event that you have say, a faulty graphics card that for whatever reason starts drawing 70A through the PCIe connector, the multi-rail PSU should trigger OCP and shut down. The single rail PSU will only shut down when OPP is tripped, which can be a much higher value and result in the full 70A being delivered to the card without shutting down, which can lead to further damage.

So this is why the S12/M12 is shunned. 

Link to comment
Share on other sites

Link to post
Share on other sites

3 minutes ago, Matsozetex said:

So say if some 'smart person' tried to power a Vega 64 with a single 8 pin using splitters the wires would melt/be damaged and the power supply wouldn't to jack?

Potentially, yes.

... Are you planning on powering a Vega 64 with the CX650M you mentioned earlier?

 

Just now, Matsozetex said:

So this is why the S12/M12 is shunned. 

Amongst other reasons... It's fair to say that any ~10 year old PSU should be avoided.
@Stefan Payne can probably list off a thousand reasons not to buy a S12/M12.

CPU: Intel i7 6700k  | Motherboard: Gigabyte Z170x Gaming 5 | RAM: 2x16GB 3000MHz Corsair Vengeance LPX | GPU: Gigabyte Aorus GTX 1080ti | PSU: Corsair RM750x (2018) | Case: BeQuiet SilentBase 800 | Cooler: Arctic Freezer 34 eSports | SSD: Samsung 970 Evo 500GB + Samsung 840 500GB + Crucial MX500 2TB | Monitor: Acer Predator XB271HU + Samsung BX2450

Link to comment
Share on other sites

Link to post
Share on other sites

Just now, Spotty said:

Potentially, yes.

... Are you planning on powering a Vega 64 with the CX650M you mentioned earlier?

No its just a hypothetical as I have seen a lot of people say if the wattage of the power supply can support the graphics card, then using splitters is alright. Realizing that now and a bit of googling the 8pin has a 150W max while the 6pin has a 75W max.

 

2 minutes ago, Spotty said:

Amongst other reasons... It's fair to say that any ~10 year old PSU should be avoided.
@Stefan Payne can probably list off a thousand reasons not to buy a S12/M12.

I've seen his comments, fair to say that he has convinced the entire internet to not buy it. I knew about the OTP and cross load but since I wasn't sure about OCP, I wasn't sure about its OCP issue.

Link to comment
Share on other sites

Link to post
Share on other sites

10 minutes ago, Matsozetex said:

So this is why the S12/M12 is shunned. 

the S12ii and M12ii are bad mainly because they are group regulated. You can not trip OPP (they dont have OCP) while the voltage regulation gets out of spec (they have UVP and OVP, but they kick in too late). There's no OTP either, so nothing to stop a rail from burning out and try start a fire

CPU: i7-2600K 4751MHz 1.44V (software) --> 1.47V at the back of the socket Motherboard: Asrock Z77 Extreme4 (BCLK: 103.3MHz) CPU Cooler: Noctua NH-D15 RAM: Adata XPG 2x8GB DDR3 (XMP: 2133MHz 10-11-11-30 CR2, custom: 2203MHz 10-11-10-26 CR1 tRFC:230 tREFI:14000) GPU: Asus GTX 1070 Dual (Super Jetstream vbios, +70(2025-2088MHz)/+400(8.8Gbps)) SSD: Samsung 840 Pro 256GB (main boot drive), Transcend SSD370 128GB PSU: Seasonic X-660 80+ Gold Case: Antec P110 Silent, 5 intakes 1 exhaust Monitor: AOC G2460PF 1080p 144Hz (150Hz max w/ DP, 121Hz max w/ HDMI) TN panel Keyboard: Logitech G610 Orion (Cherry MX Blue) with SteelSeries Apex M260 keycaps Mouse: BenQ Zowie FK1

 

Model: HP Omen 17 17-an110ca CPU: i7-8750H (0.125V core & cache, 50mV SA undervolt) GPU: GTX 1060 6GB Mobile (+80/+450, 1650MHz~1750MHz 0.78V~0.85V) RAM: 8+8GB DDR4-2400 18-17-17-39 2T Storage: HP EX920 1TB PCIe x4 M.2 SSD + Crucial MX500 1TB 2.5" SATA SSD, 128GB Toshiba PCIe x2 M.2 SSD (KBG30ZMV128G) gone cooking externally, 1TB Seagate 7200RPM 2.5" HDD (ST1000LM049-2GH172) left outside Monitor: 1080p 126Hz IPS G-sync

 

Desktop benching:

Cinebench R15 Single thread:168 Multi-thread: 833 

SuperPi (v1.5 from Techpowerup, PI value output) 16K: 0.100s 1M: 8.255s 32M: 7m 45.93s

Link to comment
Share on other sites

Link to post
Share on other sites

Group regulation is not that bad. Just use it to run a regular PC and you'll never crossload it to where voltages go out of spec.

Link to comment
Share on other sites

Link to post
Share on other sites

1 hour ago, Matsozetex said:

I've been trying to understand how OCP works, I've read through JG's articles and Toms PSUs 101 but still have trouble grasping the logic of it.

Essentially you guess the Current on a Rail and when that limit is reached the Supervisior tells the Control Chip to turn off.

Its similar to the Breaker in your breaker bOX.


With the example you've given or Single Rail in General it is uncertain wheter the OCP trips or something else, in this case its what we call "OPP" aka Over Power Protection. Wich is more or less the same, just implemented in a completely different area (here: usually PFC Section).

The Problem with OPP is that it is slower to react than OCP as more things have to happen, so it is advantageous to have both...

 

Quote

For example, the Corsair CX650M has 54A on its single 12v rail but wouldn't the OCP trip at 20A (240VA at 12v) or does OCP trigger when someone is trying to pull more power than a cable is specified to handle (i.e An 8 pin from the PSU to a splitter in a 1080ti). Do single rail PSUs even need OCP?

IIRC it doesn't even have OCP on 12V, only on minor rails...

19 minutes ago, Matsozetex said:

So this is why the S12/M12 is shunned. 

One of the few things.

 

Other things are:

They lied on the Label on earlier models and claimed two 12V Rails, when there is none.
Price -> it is still sold in the higher price range, where you can get independantly regulated units from be quiet, Corsair, Cooler Master and others.

Voltage regulation -> Regulates 5V and 12V is regulated dependant on each other. That means if you have high load on 5V and low load on 12V, 12V goes high and 5V low. And vice versa with the other way around. In short: Its crap with modern stuff as both rails influence each other heavily.

Protection  -> If you short something "the right way", it might catch fire but the PSU will not switch off but it might die.

Fan Speed -> at medium load, the fan rotates at 2000rpm or so. And we're talking like 250W or so. Reasonable low loads.

 

 

In short: If you'd look at other manufacturers stuff, you'd get a unit that's way better than the S12II for the same Price...

 

And another thing that is really annoying is the "Wapanese Caps" thing, that force people to think to have a good unit because the Caps are from Japanese brands.

"Hell is full of good meanings, but Heaven is full of good works"

Link to comment
Share on other sites

Link to post
Share on other sites

2 minutes ago, Matsozetex said:

No its just a hypothetical as I have seen a lot of people say if the wattage of the power supply can support the graphics card, then using splitters is alright. Realizing that now and a bit of googling the 8pin has a 150W max while the 6pin has a 75W max.

By splitters, do you mean a PSU cable that comes from one cable attached at the PSU side and is designed to split in to two 6+2 connectors like THIS?
Or do you mean using some cheap Y splitter off of eBay like THIS to split a single connector in to two?

CPU: Intel i7 6700k  | Motherboard: Gigabyte Z170x Gaming 5 | RAM: 2x16GB 3000MHz Corsair Vengeance LPX | GPU: Gigabyte Aorus GTX 1080ti | PSU: Corsair RM750x (2018) | Case: BeQuiet SilentBase 800 | Cooler: Arctic Freezer 34 eSports | SSD: Samsung 970 Evo 500GB + Samsung 840 500GB + Crucial MX500 2TB | Monitor: Acer Predator XB271HU + Samsung BX2450

Link to comment
Share on other sites

Link to post
Share on other sites

2 minutes ago, Spotty said:

By splitters, do you mean a PSU cable that comes from one cable attached at the PSU side and is designed to split in to two 6+2 connectors like THIS?
Or do you mean using some cheap Y splitter off of eBay like THIS to split a single connector in to two?

Not 6+2 but 8 to 2x8 pin converters.

Link to comment
Share on other sites

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×