A ebay fake GTX 1050/2GB video card is really a GK107B4 GTX 650ti/1GB: restoring functionality by flashing video BIOS with CH431A board

I (knowingly) purchased a fake GTX1050 from ebay for about $90. That's how you know it's fake, of course. It performed OK for desktop apps but was way slow for any benchmarks and also crashed with corrupted video garbage whenever more than 1GB of RAM was used. Of course eBay gave me my money back, which I was counting on when I made the purchase. 


The back of the board gave no real hints to the actual GPU, though it did expose a video bios flash ram chip. NVFLASH couldn't reprogram it, probably due to intentional corruption of the firmware, so I had to purchase a CH431a programmer (for all of $3 on ebay, plus a clip to connect it while still soldered onto the board!!). Before flashing it I had to decide what it actually was. Taking off the heat sink revealed a model number on the board: gk107b4, which could be just about anything in the GTX 600 series. Interestingly, the chip had some of the writing on it worn off. Like they were trying to head off possible returns. Only this part remained: 220-a.


 I eventually guessed it was a GTX650ti, and using tech powerup's video bios database flashed the card. The first flash didn't work - no video output. But I hadn't been that careful about matching up the outputs - this card had HDMI out, and the bios I downloaded only had MiniHDMI. So I tried again with a perfect match. Bingo: Working video card. I benchmarked using userbenchmark which allowed me to compare video card performance to all the other 650ti's out there. The card benchmarked right in the middle of the pack so either I guessed right, or just close enough.

I also tried a longer benchmark with unigine heaven 4.0. Worked fine, no glitches or otherwise. Success! But it's a damn slow card compared to a GTX 1050. Who falls for this kind of thing, anyway? On the upside, this card performs way better than a rectangular scrap of cardboard that I scrawled GTX 1050 on, and for about the same cost. If you discount my time that is, which is always a mistake :-(.

Comments

Unknown said…
Hi, can you share bios file that worked for you?
December 23, 2018 at 1:08 PM
Anonymous said…
Kudos, m8. I'm a sucker who fell for the scam, and ended up with the exact same pack as you. Any chance you could share the link to the bios that you used?
January 24, 2019 at 1:05 PM
Post a Comment

Email me

Name

Email *

Message *

Popular posts from this blog

Panasonic TH-42PX75U Plasma TV review: input lag and upscaling tested using the piLagTesterPRO

Image
This 720p Plasma TV from 2008 is very unique and quite nice in many ways. It used to retail for around $1300. If you can find one it's a great choice for retro gaming.  Overview/Image quality This is a plasma TV, which generally means at least the following: nearly infinite viewing angles, and jaw dropping weight. That's true in this case as well. This 42" model weights 75lbs.  At native resolution (720p) this TV crops 20 pixels off the top/bottom and 30 off the sides. 480p/i crops 20 pixels off all sides. Both modes have aliasing, too, though it's not too bad. There's really no excuse for aliasing on a TV this big: you know the panel was designed specifically to be a tv. Why not hit 720p exactly? And as for 480p/i, they could have at least offered a zoomed out mode with black borders and 1:1 pixel mapping. But no such luck. There are zoom/stretch options to crop even more, but none that crop less. While the native resolution supposedly is 1366 x 768 it does accept...
Read more

piLagTester PRO order page

Image
The piLagTesterPRO measures your monitor's input lag and response time. It is a cheap, full featured add-on to the Raspberry Pi; I provide the sensor and required software, you provide the Pi (a $5 pi Zero is more than sufficient). Although it has a DIY aesthetic, it is quite functional and in fact offers quite a lot more features than the commercial alternatives (Leo Bodnar and the Time Sleuth), and has been tested with over 40 TVs, including 4k models .  Here's what it looks like in action: The pi draws a black background, and then roughly once a second displays a set of target rectangles (top/middle/bottom). You place the sensor over the desired target, and the piLagTester measures the monitor's response starting from the moment the frame of video data is sent over the Pi's HDMI port. This is plotted in the graph, which is sideways for space reasons. The lag tester measures two thresholds. In red it shows how long it takes for the monitor to start displaying t...
Read more

A $5 TV Input Lag tester using a Raspberry Pi Zero

Image
High speed monitors are all the rage now, with quoted response times  of 1-4ms. But does that mean that 1-4ms after you make an input (say, jump) you will see that on your screen? No. Often a much bigger issue is  Input lag. This is   the delay between your computer / game console sending a video frame to your display and your display starting to actually show it, and it ranges from 4ms to 80ms. The problem with measuring input lag is that you can't, at least with everyday hardware. The solution for hardware review sites is to purchase dedicated lag testers such as the leo bodnar tester,  timesleuth (both ~$100), or the pi lag tester pro ($40) or use homegrown methods utilizing an oscilloscope ("free" if you already own a $400+ oscilloscope and perhaps a $50k+ engineering degree). Here I introduce a $5 solution (or, "free" if you own the required equipment). What's required: A Raspberry Pi Zero ($5 + S&H from several sources). This single b...
Read more