Salta al contenuto principale

Haswell business laptop introduced by Dell in 2013. This system is notable for being the first Haswell business laptop released, but has multiple known problems - including AMD GPU failure. A new battery will operate the laptop for ~4-7 hours, depending which one is installed.

10 Domande Visualizza tutte

What causes the GPU blackout issue on these laptops?

I recently semi retired this due to the age of the system when the hard drive failed. However, it has a longstading GPU problem I was okay with living with as long as the laptop worked in regards to the AMD GPU. When I was using the laptop with a second display (GPU always kicks in on these with multiple displays), the screen has a tendency to black out without warning and my only recovery option is a hard shutdown. I gave up on the laptop because of this issue, especially since it is a 7 year old platform (2013 model). In addition to GPU problems, the other reason is the Haswell CPUs are hit pretty hard by the CPU exploits in this system due to the subpar cooling combined with the performance penalty - in which I only patched Spectre V1 and left the others be because I already took a noticeable penalty with Spectre V1 - yes, even with the Win10 high speed patch. I stopped trying to keep up with it after the A24 BIOS and called it good because I’m well aware of the risks.

Is there any known way to reliably fix the GPU issue or is my advice on the device page (board swap or scrap) apply with how mine has failed?

I never messed with the board, but I repasted it years ago to try and help the issue and it didn’t do much outside of deferring the worst parts of the problem. The catch with mine is it has the 4810MQ CPU (47W TDP), so I had to deal with more heat then the people with M series processors. The AMD 8790M is ~35W (AMD isn’t open about the actual TDP), so the problem appears to have been made worse by the cooler since it is 80W of heat through one fan and vent. If I had the i7 4610M it would have been 35W for the CPU, but the situation would still have been pretty bad at 70W.

Risposto! Visualizza la risposta Anch'io ho questo problema

Questa è una buona domanda?

Punteggio 2
Aggiungi un commento

1 Risposta

Soluzione Prescelta

@nick sound st ome like you are running into typical flip chip ball grid array issues with overheated solder bumps. As you know there is no cure for it other than replacing the processor with a new one.

Questa risposta è stata utile?

Punteggio 2

2 Commenti:

And I've always deemed it to be not worth it on these due to the poor cooling. The issue now is the lack of new chips makes it so you're chancing it as to how long the repair will last, especially if it came out of one of these. I was afraid of it being a flip chip failure, which is why I decided to sideline it and replace it (and try a repair if viable). The actual CPU is modular, so if I found a cheap Intel HD version of the machine (although I'd want to be sure it's the eDP version and not LVDS if possible) I may be able to carry my 4810MQ over but the base system is dead end, but more then likely I'd want to find a reliable laptop to do that with and these have a few issues. I'll probably keep it around for the few E6540 specific parts known to be intact and things like the processor and use it as a parts machine.

The other issue is even if you replace the board, they will die the same way since it doesn't fix the fundamental cooling issues (and I'd bet a lot of the 8790M boards are reflowed). Changing the paste probably helped a lot with mine in getting some extra time out of it, but clearly didn't spare the GPU from failing like everyone else had to deal with early on (first 1-2 years).

At least I got 5 years of use out of it before it happened to mine so severely it was toast since I got it in 2015. It started with bluescreens in Win7 when the GPU kicked in with no grace period to adjust and got to the point it started having hard video crashes that got worse with time - roughly when I installed Win10 ~6 months prior to 1/14/20. I think it was just time for it to get worse more so then 10 straining the card and making it worse.

da

@oldturkey03 sadly I know what probably finished it off. I took it apart to see what parts I have left that I can reuse and the CPU cooler burst open, leaking coolant on the system board and burning the GPU out.

Yep; it's beyond repair. Good thing I didn't store it with the (functional, but dead) battery installed. It's just a series known for running hot, so I figured it was just running hot as usual. Hopefully the CPU isn't trashed as well - that was the best part :(. Doesn't look burned up so it may be fine but the board is a lost cause.

da

Aggiungi un commento

Aggiungi la tua risposta

Nick sarà eternamente grato.
Visualizza Statistiche:

Ultime 24 Ore: 0

Ultimi 7 Giorni: 0

Ultimi 30 Giorni: 0

Tutti i Tempi: 792