Nereus

redteamgo
Cross posting this to this thread in case anyone has any ideas.  My 2nd kingpin hybrid died this morning at boot .  The video below shows the error I get on the OLED.  System posts, card's pump powers on and runs.  The card's LED does NOT turn on.  The displayport ports obviously don't work.
 
To troubleshoot, I tried every 8pin PCIE cable I had in every possible combination in the card and in several different combinations from the PSU (EVGA 1000W) and I get the same error on the OLED.
 
Next, I plugged my 1080 TI back in and used two random cables from the test in the same position the kingpin used before it popped .  The card fires right up and the system is fine now.
 
Someone please help if you have gone through this error and fixed it.  I have reseated the card, cables, removed all non-essential devices (fans, controllers, cables, etc) but have not had an success beyond what you can see below in the vid.
 
Separately, I bought a 3090 KP HC kit when my queue position came up and just spent another $800 on water cooling upgrades this morning about an hour before it died .  The hydrocopper plate is a really sweet looking piece of hardware.  However, how in the world can I get comfortable with upgrading my hybrid KP card to HC when the card's life cycle is so unpredictable.  I feel extremely foolish at this point with this level of $ spend.
 
 
https://www.youtube.com/watch?v=Y7MBcJ-Uy4A

Ouch.. you were using 3 x 8-pin cables right? (have to ask). Looks like it's just the 8-pin header that is having an issue, if the same cables are ok with your 1080. For the same issue to happen twice to you is strange. Check every 8-pin outlet on your PSU going to the 1080, in case it's not the card or the cable, but one of the 8-pin outlets on your PSU. Also what sort of PSU and wattage are you using? (ok, I see EVGA 1000W). Have you tried a different PSU (if you have access to one)?
 

My first RMA was a pump failure just to give some credit where due.  This is a new thing
 
Yes, I used 3 independent 8 pin cables.   I even tried the 1+1 cable for ****s and giggles.  
 
I did what you mentioned with the PSU -> 1080.  That was the second troubleshooting step after trying all power cable combinations in the kingpin.
 
While testing the KP hybrid, I removed everything but the CPU, RAM, CPU cooler and the rear case fan.  I tested all 8 pin cables on all kingpin ports in slots on the PSU.  The error always shows the right 8 pin on the OLED screen no matter what.   Whereas, the 1080 TI works fine on any PSU position with any cable from the kingpin setup.
 
I don't have a different PSU handy.  The one in use is about 2 months old.  I got it for the KP because everyone here said 850w wasnt enough lol.  Regarding the PSU PCIE peripheral port, yes, it was powering a fan controller before it died.  The fan controller had the radiator fans (push/pull) and the bottom intake fans - 8 total.

redteamgo
Cross posting this to this thread in case anyone has any ideas.  My 2nd kingpin hybrid died this morning at boot .  The video below shows the error I get on the OLED.  System posts, card's pump powers on and runs.  The card's LED does NOT turn on.  The displayport ports obviously don't work.
 
To troubleshoot, I tried every 8pin PCIE cable I had in every possible combination in the card and in several different combinations from the PSU (EVGA 1000W) and I get the same error on the OLED.
 
Next, I plugged my 1080 TI back in and used two random cables from the test in the same position the kingpin used before it popped .  The card fires right up and the system is fine now.
 
Someone please help if you have gone through this error and fixed it.  I have reseated the card, cables, removed all non-essential devices (fans, controllers, cables, etc) but have not had an success beyond what you can see below in the vid.
 
Separately, I bought a 3090 KP HC kit when my queue position came up and just spent another $800 on water cooling upgrades this morning about an hour before it died .  The hydrocopper plate is a really sweet looking piece of hardware.  However, how in the world can I get comfortable with upgrading my hybrid KP card to HC when the card's life cycle is so unpredictable.  I feel extremely foolish at this point with this level of $ spend.
 
 
https://www.youtube.com/watch?v=Y7MBcJ-Uy4A

This is very unfortunate. Not sure what may be causing these problems. I will say that my 3090 Kingpin was originally a Hybrid, I just installed my KP Hydro Copper kit like 2-3 days ago. And I have only owned this GPU for about 1.5 weeks all together. I have ran it very hard on the pre-installed LN2 520 watt bios daily ever since owning it. I keep the memory at +1700 to +1800 minimum. 


Have you used classified tool? Have you used the dip switches on the back? maybe you accidentally ran too high voltage, with too high of a load line calibration? The default load line is on lvl 1 which is wayyy too much in classified tool. It should be set to maybe 5 or 6, and honestly load level 7 is just fine for ambient water cooling.


I have carefully played around with sending more memory voltage to my card. And I can say from experience here, it’s easy to burn one of these up without being careful, or even realizing it, when using the dip switches and classified tool with PX1. 

You’ve gotta be careful with these cards. But going through (2) of them just doesn’t seem normal at all to me..

Is the OLED display seated and plugged in correctly?  


 
 
 
 

tps34433

redteamgo
Cross posting this to this thread in case anyone has any ideas.  My 2nd kingpin hybrid died this morning at boot .  The video below shows the error I get on the OLED.  System posts, card's pump powers on and runs.  The card's LED does NOT turn on.  The displayport ports obviously don't work.
 
To troubleshoot, I tried every 8pin PCIE cable I had in every possible combination in the card and in several different combinations from the PSU (EVGA 1000W) and I get the same error on the OLED.  Next, I plugged my 1080 TI back in and used two random cables from the test in the same position the kingpin used before it popped .  The card fires right up and the system is fine now.  Someone please help if you have gone through this error and fixed it.  I have reseated the card, cables, removed all non-essential devices (fans, controllers, cables, etc) but have not had an success beyond what you can see below in the vid.
 
Separately, I bought a 3090 KP HC kit when my queue position came up and just spent another $800 on water cooling upgrades this morning about an hour before it died .  The hydrocopper plate is a really sweet looking piece of hardware.  However, how in the world can I get comfortable with upgrading my hybrid KP card to HC when the card's life cycle is so unpredictable.  I feel extremely foolish at this point with this level of $ spend.
 
 
https://www.youtube.com/watch?v=Y7MBcJ-Uy4A

This is very unfortunate. Not sure what may be causing these problems. I will say that my 3090 Kingpin was originally a Hybrid, I just installed my KP Hydro Copper kit like 2-3 days ago. And I have only owned this GPU for about 1.5 weeks all together. I have ran it very hard on the pre-installed LN2 520 watt bios daily ever since owning it. I keep the memory at +1700 to +1800 minimum. 


Have you used classified tool?
>>I have not used the classified tool.  The card itself is unaltered.  The rad has 3 additional fans on it for push/pull.
 
Have you used the dip switches on the back?
>>No.  I flipped the stock bios to LN2 when I benchmarked the card in Timespy to make sure it was in the right performance range and have run it in LN2 stock bios at stable +800m/+84c since.  Temps in the mid 50s under load.
 
maybe you accidentally ran too high voltage, with too high of a load line calibration?
>>I don't know how that could have happened .  The card popped at boot and I have only ever used PX1 with this card.  During troubleshooting, I did try different bios switch positions.  The latter made no difference.
 
 
The default load line is on lvl 1 which is wayyy too much in classified tool. It should be set to maybe 5 or 6, and honestly load level 7 is just fine for ambient water cooling.  I have carefully played around with sending more memory voltage to my card. And I can say from experience here, it’s easy to burn one of these up without being careful, or even realizing it, when using the dip switches and classified tool with PX1. 

You’ve gotta be careful with these cards. But going through (2) of them just doesn’t seem normal at all to me..
>>The first card didn't actually die - the AIO pump died.  Whoever gets that base card as an RMA will be happy.  This is a different issue.  I'm glad its not normal, that means 3rd times the charm perhaps.
 

Is the OLED display seated and plugged in correctly?  
>>I haven't opened the card to install the HC kit yet, I'm still getting all of my parts.  At this point, I don't see the benefit of opening the card and introducing additional risk of damage.
 
 
 
 

 
Thank you for posting your thoughts.  I responded in line to your questions in bold to make it easier to track in case someone else has these issues :(

So folks, help me out here:
 
Card runs fine in AC Odyssey, Valhalla, League of Legends, and CS:GO. BUT it crashes every so often in Witcher 3. Is this a known issue? It crashes even with no OC on VRAM or core, and just the voltage/power slider all the way up just for GPU boost to take care of everything. Running at 58-59 degree C with "stable" 2010-2025 Mhz, 80 degree C on the VRAM Junction Temp. AC Odyssey can take a +25Mhz overclock on the core and a bit on the VRAM no problem (though it stills runs around 2025 Mhz prob due to thermal limitations). Training neural networks for 10+ hours doesn't crash. Windows Event Viewer logs "nvlddmkm stopped responding, restarted" or something, where the driver just crashes. It is a simple crash to windows and no hard reboot. Was wondering if any of yall mad lads want to test out Witcher 3 and tell me if you are running into the same issue (sometimes it runs stable for an hour or two and crash). Using latest driver version (released on 7/19), also tried 466.47. Both crashes. Any help would be appreciated!

tps34433

elbramso

tps34433

elbramso
I just received my card from RMA and have to say the gpu is better on the core but worse on the memory. My old card had issues to hit 2145mhz on the core constantly with the 520w bios but was stable with +1500 on the memory. The new card can easily hit 2190mhz but gets unstable at +750 on the memory.
I will start pushing the card once I receive the HC Kit.

Try the 520 watt LN2 bios, then retry the memory OC again. I believe it sends different voltages or possibly has different timings maybe VS. the OC or standard bios. 
 
Because my normal +1700Mhz on the memory is stable on the 520 LN2 bios in games. But not the normal bios. (It went down after 2 hours of gaming) 
 
Or you could try increasing voltages with dip switches.
 
But, temperature is everything here. So everyone’s clocks will be different. 
 
 
 

I'm on the LN2 BIOS (that came with the card) already. 
Mem is unstable above +750, so I won't touch it for a while.
Using the the full tdp of the 520w bios I was able to score 15288 in PR which is ok for the Hybrid version I guess:
https://www.3dmark.com/3dm/64153340
 

Enable both of the fsvbb dip switches on the back of the card. that’s memory voltage. 


I am testing +1800 right now totally fine. With 1.470V memory voltage.
 
 
I run a KP Hydro Copper block so my memory temps are outstanding! 


 
Enable the dip switches and try again.
 
 

I'll do once I receive my HC Kit. Don't wont to stress the mem temps too much before I have a good cooling.
Btw. is there any advantage using dip switches over the classified tool? Obviously the dip switches will still be set once the PC reboots but other than that?
 

elbramso

tps34433

elbramso

tps34433

elbramso
I just received my card from RMA and have to say the gpu is better on the core but worse on the memory. My old card had issues to hit 2145mhz on the core constantly with the 520w bios but was stable with +1500 on the memory. The new card can easily hit 2190mhz but gets unstable at +750 on the memory.
I will start pushing the card once I receive the HC Kit.

Try the 520 watt LN2 bios, then retry the memory OC again. I believe it sends different voltages or possibly has different timings maybe VS. the OC or standard bios. 
 
Because my normal +1700Mhz on the memory is stable on the 520 LN2 bios in games. But not the normal bios. (It went down after 2 hours of gaming) 
 
Or you could try increasing voltages with dip switches.
 
But, temperature is everything here. So everyone’s clocks will be different. 
 
 
 

I'm on the LN2 BIOS (that came with the card) already. 
Mem is unstable above +750, so I won't touch it for a while.
Using the the full tdp of the 520w bios I was able to score 15288 in PR which is ok for the Hybrid version I guess:
https://www.3dmark.com/3dm/64153340
 

Enable both of the fsvbb dip switches on the back of the card. that’s memory voltage. 


I am testing +1800 right now totally fine. With 1.470V memory voltage.
 
 
I run a KP Hydro Copper block so my memory temps are outstanding! 


 
Enable the dip switches and try again.
 
 

I'll do once I receive my HC Kit. Don't wont to stress the mem temps too much before I have a good cooling.
Btw. is there any advantage using dip switches over the classified tool? Obviously the dip switches will still be set once the PC reboots but other than that?
 

The dip switches are just super easy, theyll
also provide a higher base voltage in classified. No app adjustment, and they'll always apply that voltage no matter the bios you use, or by restoring defaults in PX1, classified etc. 
 
1.375V is the default memory voltage on the OC, and Standard bios position. So they memory won’t go too far on these without assistance.


I would just be careful. I imagine someone could easily damage one of these cards with too much. I’ve read 1.40-1.43 is maximum safe for Hybrid cooler. I have sent as much as 1.51V and going beyond 1800Mhz is stable. I’m gonna keep fine tuning as I have seen guys running 2Ghz 

Anyone here that moved from a HC from their hybrid want to sell their 360 AIO? PM me. 

https://optimuspc.com/products/signature-gpu-block-kingpin-3090

Optimus just released their active block for an eye watering $600

The waterblock kingpin looks too involved compared to the hybrid kingpin. Don't see the real benefit.

Both the hydrocopper and hybrid rely on passive cooling for the rear VRAM. The real benefit with the Optimus is actively cooled VRAM. Useful if someone runs memory heavy operations like mining. Those chips hit the throttling temp pretty fast without active cooling