SAPPHIRE Announces NITRO+ AMD Radeon RX 9070 Series Graphics Cards

[MxPhenom 216]

I think they're just as liable to(wards) industry norms, as any other company:

That would seem adequate for a 300w GPU

 

[Sir Beregond]

You really think the engineers of these cards would allow it to do that? There's no world to you where the board makers for these cards with understanding to what is happening on Nvidia cards to just not account for any of that?

That isnt even happening on 5090s without the line balancing unless you force it to do that by cutting wires like Jayztwocentz did in his video.

Well supposedly the RX 9000's have been done for months and been sitting in retailer's warehouses since Dec or Jan. It's absolutely possible they did the same garbage on 12V input handling as the 50-series because it was all complete before the 50-series even launched.

Will have to wait and see.

 

[Dave65]

Love Sapphire GPUs but hate their warranty denials.

 

[MxPhenom 216]

Well supposedly the RX 9000's have been done for months and been sitting in retailer's warehouses since Dec or Jan. It's absolutely possible they did the same garbage on 12V input handling as the 50-series because it was all complete before the 50-series even launched.

Will have to wait and see.

The same issue was happening on 4090 though, and these Sapphire implementation has fuzes and shunt resistors. Most of Nvidia cards implementations have basically nothing (Asus Astral being a known exception)

 

[LabRat 891]

That would seem adequate for a 300w GPU

You should know better than I.

But, based on

-Der8auer's FLIR imaging,
-Buildzoid's technical breakdown(s),
-2+ generations of low medium- and high-power nvidia cards w/ burnt connectors,
-3+ revisions to said connector,
-generally-accepted current ratings for 16 and 18 AWG wire
and
-Sapphire's own marketing materials

I have ZERO trust in that presumption.

 

[Sir Beregond]

The same issue was happening on 4090 though, and these Sapphire implementation has fuzes and shunt resistors. Most of Nvidia cards implementations have basically nothing (Asus Astral being a known exception)

It's not just about shunts. The Astral still has the same input problem despite the addition of shunts. The shunts will merely serve as a way to warn of a problem and is all they do. It will be down to how the different power phases are split up and if different 12V inputs feed different phases or if it's all one big blob of 12V.

 

[LabRat 891]

Love Sapphire GPUs but hate their warranty denials.

Sadly, increasingly common across the AIBs. -Asus recently went through some bad press for doing that and worse.

W/in the last month, I did have to RMA my RX 7900 GRE NITRO+ w/ Althon Micro/Sapphire.
It was a more obtuse process than my RMAs w/ G.Skill, AMD, etc. but, overall was a 'fine' process. The Replacement card is lightyears better than the one it replaced (and, is actually stable! )

 

[MxPhenom 216]

You should know better than I.

But, based on

-Der8auer's FLIR imaging,
-Buildzoid's technical breakdown(s),
-2+ generations of low medium- and high-power nvidia cards w/ burnt connectors,
-3+ revisions to said connector,
-generally-accepted current ratings for 16 and 18 AWG wire
and
-Sapphire's own marketing materials

I have ZERO trust in that presumption.

What med end Nvidia cards have reported burnt connectors, and how many relative to amount sold? All media, and reports I have seen from users have all been xx90 cards. -> and a lot of them were using some shitty 90 degree adapter from companies like cablemod, or badly made cables from Corsair.

I have also built probably more systems than total number of reported cases with xx80s and xx90s with not a single burnt connector. One common denominator on all those systems is ZERO used of any kind of adapter, and all cables used were supplied with PSU.

This doesnt mean im writing off the issues as user error. This connector/cable is 100% an issue on a card like a 5090 when lack of fail safes are used

 

[Dave65]

Sadly, increasingly common across the AIBs. -Asus recently went through some bad press for doing that and worse.

W/in the last month, I did have to RMA my RX 7900 GRE NITRO+ w/ Althon Micro/Sapphire.
It was a more obtuse process than my RMAs w/ G.Skill, AMD, etc. but, overall was a 'fine' process. The Replacement card is lightyears better than the one it replaced (and, is actually stable! )

Yeah, but Sapphire needs that exposure right now.. they are terrible.

 

[MxPhenom 216]

It's not just about shunts. The Astral still has the same input problem despite the addition of shunts. The shunts will merely serve as a way to warn of a problem and is all they do. It will be down to how the different power phases are split up and if different 12V inputs feed different phases or if it's all one big blob of 12V.

Did you miss the fuses part?

 

[Sir Beregond]

Did you miss the fuses part?

Guess so. Still more replying to the Astral part of your post. The simple addition of shunts did nothing to change how the power input is handled.

 

[MxPhenom 216]

Guess so. Still more replying to the Astral part of your post. The simple addition of shunts did nothing to change how the power input is handled.

With regard to the Astral, the Asus solution seems to be implemented with the idea that it would be used with a PSU with load balancing done on the PSU side. Their ATX 3.1 PSUs do that so it seems like its a way for them to sell their PSUs with their cards. Not exactly uncommon.

 

[Random_User]

Possible $899 for PULSE Triple RX 9070 XT

If not more. $899 will probably is the Sapphire "MSRP" for their Pulse series. The VAT/tariffs/gouging is to be inevitably added on top.

Show me the 8D report and all the specs. Than we can talk. Especially step 7.
years ago quality claims had to be finished for the automotive branch in less than 7 calendar days. We had nice customers who filed quality claims on friday 11 am. So no early friday end of work shift for myself in the quality department

How many days do we have these issues wiht that nvidia gpu connector? Or should I already write, how many years do we have problems with that nvidia connector?

Also show me the FMEA for that connector, which dates back to the first issue reported.
Note: I talk about this: https://asq.org/quality-resources/fmea

I do not want to use an adapter or buy a cable just for a graphic card. Other graphic cards do not need a cable adapter or a newer cable or a newer power supply unit which ships with that connector.

go on - give me a non - absurd statement.

Whataboutism: It's the same why I did not bought any gigabyte mainboard. Because of those nice gigabyte power supplies which exploded like a fire cracker. Or any SAMSUNG NVME - because of those Firmware issues over all those years.

edit: other thread implies that this issue is well known since nvidia 3090 ....

This, sadly depends on the AIB and direct partner's QA, laziness and "cutting corners". Especially, since how the AIBs like Asus are keen to undermine the Radeon branded crads quality and public image. So, considering them having the issues with the top end solutions of "top end"nVidia GPU (GB202-300-A still might be a cut down silicon, since... GB202, and potentially a Ti, or "Super" refresh)
Also the AIBs probably are just very lazy/greedy, and are just doing the old trick, by trying to utilize the PCB design, already made for the nVidia counterparts. Though, having Sapphire joined AsRock with 12V 2x6 "endeavour". Perhaps, more of the manufacturers are about to join soon, sadly. Hopefully there would be a choice of power connector variant.
And considering, for how long the 9070 series cards been laying in the warehouses, the amount of potential issue, might be even bigger, than in case of nVidia's miserable supply.

You really think the engineers of these cards would allow it to do that? There's no world to you where the board makers for these cards with understanding to what is happening on Nvidia cards to just not account for any of that?

That isnt even happening on 5090s without the line balancing unless you force it to do that by cutting wires like Jayztwocentz did in his video.

Engineers might not. But everyone knows the final decisions and marketing are not made by engineers.

Now add here the fact of RDNA4 being a stopgap, and the AIBs habit to "save" as much as possible, even on crucial components/design, and the things start looking very grim...

The connector, is raw. By connector, it also means the corresponding specifications, and standards. It should have been ironed out, before entring the consumer market. The emphasis is on the consumer, because it has the infinite variety of conditions and configs. Thus making the troubleshooting and the result fluctuation way higher, than the enterprise market, with strict submission to standards, and thus barely any difference between the HW.
Something like the balancing of the power delivery across the cables, should have been the established and obligatory spec of the connector from the get go. There shouldn't have been any further 12VHPWR iterations, to begin with. Let alone the iteration/the version, which is still as faulty as the original. So there is ATX 3.1 and PCIE 5.1, exist only due to nVidia was pushing the bugged proprietary version of the existing standard/connector, and was unable/unwilling to sort their cr*p out.

5090 is basically the top of the crop. And even it is being treated like the bottom of the barrel garbage. Not only by the comple absense of any shunts on the nVidia's own so called "premium" and "superrior" FE solutions. But it also has the silicon quality "wobble" regarding the specs, which is unacceptable for the same SKU, even if it is the bottom end chip. And Asus, MSI and others has the hubris, to ask/demand the hefty "tax" for "better" power stages/shunts (for having them at all, vs FE), something what has to be a obvious, and obligatory...

Well supposedly the RX 9000's have been done for months and been sitting in retailer's warehouses since Dec or Jan. It's absolutely possible they did the same garbage on 12V input handling as the 50-series because it was all complete before the 50-series even launched.

Will have to wait and see.

This all are speculations... However the recent events, have shown, that there's little hope and trust on even the top videocard variants. But yeah, it will all be crystal clear, soon.

It's not just about shunts. The Astral still has the same input problem despite the addition of shunts. The shunts will merely serve as a way to warn of a problem and is all they do. It will be down to how the different power phases are split up and if different 12V inputs feed different phases or if it's all one big blob of 12V.

Something, that is non issue on older "non smart" PCIE/EPS 8 pin. Amazing, how they have managed to royally screw up with "new" "better" standard.

Yeah, but Sapphire needs that exposure right now.. they are terrible.

Indeed. They've been cheaper and more reliable at same time, before RDNA3. Then, they've scr*wed up with the coolers, namely the blatant spacers/mounting pressure issue. And also, having arrogance to ask more, over much better, reliable and also cheaper alernatives from TUL and XFX. Their RMA seems to be horrible outside Taiwan/Asia.
Sadly. Sapphire was more or less one of the most reliable supplier, for almost two decades. And how fast they've managed to flop their entire accomplishements/heritage.

 

[MxPhenom 216]

If not more. $899 will probably is the Sapphire "MSRP" for their Pulse series. The VAT/tariffs/gouging is to be inevitably added on top.

This, sadly depends on the AIB and direct partner's QA, laziness and "cutting corners". Especially, since how the AIBs like Asus are keen to undermine the Radeon branded crads quality and public image. So, considering them having the issues with the top end solutions of "top end"nVidia GPU (GB202-300-A still might be a cut down silicon, since... GB202, and potentially a Ti, or "Super" refresh)
Also the AIBs probably are just very lazy/greedy, and are just doing the old trick, by trying to utilize the PCB design, already made for the nVidia counterparts. Though, having Sapphire joined AsRock with 12V 2x6 "endeavour". Perhaps, more of the manufacturers are about to join soon, sadly. Hopefully there would be a choice of power connector variant.
And considering, for how long the 9070 series cards been laying in the warehouses, the amount of potential issue, might be even bigger, than in case of nVidia's miserable supply.

Engineers might not. But everyone knows the final decisions and marketing are not made by engineers.

Now add here the fact of RDNA4 being a stopgap, and the AIBs habit to "save" as much as possible, even on crucial components/design, and the things start looking very grim...

The connector, is raw. By connector, it also means the corresponding specifications, and standards. It should have been ironed out, before entring the consumer market. The emphasis is on the consumer, because it has the infinite variety of conditions and configs. Thus making the troubleshooting and the result fluctuation way higher, than the enterprise market, with strict submission to standards, and thus barely any difference between the HW.
Something like the balancing of the power delivery across the cables, should have been the established and obligatory spec of the connector from the get go. There shouldn't have been any further 12VHPWR iterations, to begin with. Let alone the iteration/the version, which is still as faulty as the original. So there is ATX 3.1 and PCIE 5.1, exist only due to nVidia was pushing the bugged proprietary version of the existing standard/connector, and was unable/unwilling to sort their cr*p out.

5090 is basically the top of the crop. And even it is being treated like the bottom of the barrel garbage. Not only by the comple absense of any shunts on the nVidia's own so called "premium" and "superrior" FE solutions. But it also has the silicon quality "wobble" regarding the specs, which is unacceptable for the same SKU, even if it is the bottom end chip. And Asus, MSI and others has the hubris, to ask/demand the hefty "tax" for "better" power stages/shunts (for having them at all, vs FE), something what has to be a obvious, and obligatory...

This all are speculations... However the recent events, have shown, that there's little hope and trust on even the top videocard variants. But yeah, it will all be crystal clear, soon.

Something, that is non issue on older "non smart" PCIE/EPS 8 pin. Amazing, how they have managed to royally screw up with "new" "better" standard.

Indeed. They've been cheaper and more reliable at same time, before RDNA3. Then, they've scr*wed up with the coolers, namely the blatant spacers/tension issue. And also, having arrogance to ask more, over much better, reliable and also cheaper alernatives from TUL and XFX. Their RMA seems to be horrible outside Taiwan/Asia.
Sadly. Sapphire was more or less one of the most reliable supplier, for almost two decades. And how fast they've managed to flop their entire accomplishements/heritage.

"Look grim", except the RDNA4 cards that have this connector on them has components (fuses and shunt resistors) to mitigate potential issues with the connector. Same can't be said for 99% of Nvidia card implementations.

Sure doesnt seem like these AIBs tried to save as much as possible. Ill have you know, the 3090Ti has similar on board circuitry and it is known to be nearly absent of issues with the connector.

 

[sethmatrix7]

You really think the engineers of these cards would allow it to do that? There's no world to you where the board makers for these cards with understanding to what is happening on Nvidia cards to just not account for any of that?

That isnt even happening on 5090s without the line balancing unless you force it to do that by cutting wires like Jayztwocentz did in his video.

A GPU series just released with melting connectors, underbuilt power delivery, missing ROPs, and BSOD/black screen issues.

The AiBs and Nvidia are either fucking stupid or more likely- don’t care. After all, chuds are still lining up to buy graphics cards 25% over MSRP so they know the consooomers will eat up the slop.

 

[MxPhenom 216]

A GPU series just released with melting connectors, underbuilt power delivery, missing ROPs, and BSOD/black screen issues.

The AiBs and Nvidia are either fucking stupid or more likely- don’t care. After all, chuds are still lining up to buy graphics cards 25% over MSRP so they know the consooomers will eat up the slop.

Its a "dont care" kind of thing. People still buy their cards, but its not their money maker right now