Geforce fx overclock: Overclocking Geforce FX 5200 | Overclock.net

I had mine VC of FORSA GeForce FX 5200 and it is capable clock in 310/475 but i decrease it to 305/455 due to heat generation from VC stock fan. I prefer to keep it in stock of 250/400. I used TaiwanTech PowerStrip v3.14… a very nice third-party utility!

Lucan

P.S. 3DMARK2001 SE Professional Score = 12,503 :ekkk:

Extreme overclocking GeForce FX 5200 (WinFast A340 TDH) from Leadtek.

This time for extreme overclocking, I chose
GeForce FX 5200 (WinFast A340 TDH) from Leadtek. I liked this video card because it has memory from Samsung with an access time of 4 ns. The nominal core/memory frequencies are 250/400 MHz, the company itself provides an opportunity to increase the memory frequency by 100 MHz, and it would be foolish not to use it. I was also interested in the fact that there is a temperature sensor between the processor and the cooler, and three wires go from the cooler to the board, which means it has a tachometer. The CD supplied with the video card contains the WinFox II utility — Leadtek’s own control panel. A description of all the functions of the control panel would take up a lot of space, but among others there is dynamic monitoring of temperature and voltages on the processor and memory, fan speed, and overclocking functions. Otherwise, the card is not much different from its similar ones: support for AGP 8x and DirectX 9, connectors for connecting external DVI and VGA monitors and TV-out (S-video).

In a stylish box with a corporate design, along with the board, I found: description, 3 CDs with toys and drivers, S-video cable, RCA cable, S-video — RCA adapter, DVI — D-Sub adapter.

First I tested our GeForce FX5200 at nominal core/memory frequencies of 250/400 MHz. Then, using the usual method, using Riva Tuner v2.0RC14, we managed to get new results: 300 MHz for the core and 520 MHz for memory. As tests, I ran 3DMark 2003 v.3.3.0 and 3DMark 2001 SE with default quality settings, Quake3 Arena (demo FOUR) and UT 2003 (flyby inferno) three times. All tests were carried out with a resolution of 1024×768 and 1600×1200. The test stand was assembled from the following components:

• Motherboard — Abit NF7, BIOS version — 18.
• Processor — AMD Athlon XP 1700+, operating at 2241 MHz (166×13. 5).
• Memory — 2 x 512 MB DDR 333 modules from Samsung.
• Hard — IBM DTLA 305020.
• Cooler — Thermaltake Volcano 7.
• Thermal grease — KPT-8.
• OS — Microsoft Windows XP (Detonator 45.23).

After all the usual means had been used, I decided to increase the voltage on the processor and memory. The main thing: find a voltage stabilizer. On this board, it is assembled on the APW 7060 chip. It is dual, designed specifically for video cards and regulates 1.5 V on the processor and 2.5 V on the memory chips.

By changing the parameters of the feedback resistor divider, you can change the supply voltage to the processor and memory. You can control these voltages with a tester on the legs of the corresponding electrolytic capacitors.

Now let’s figure out who is responsible for what: the 6th leg regulates the voltage on the memory, the 9th on the processor. We solder two variable resistors (I have 22 kOhm each) between these legs and the nearest grounded points on the board. Before soldering, we set the maximum resistance value on the resistors. Well, then everything is very simple, set the voltage above the nominal value by 10% and run the tests.

recommendations

With the new voltages (1.7 V on the processor and 2.9 V on the memory chips), we managed to overclock the processor to 375 MHz and the memory to 580 MHz. But the very first test 3DMark 2003 v.3.3.0 (1024×768) showed a lower result than with normal overclocking, despite the fact that the frequencies increased significantly. Something was slowing down the card from the inside. It seems that the increase in voltage caused increased heating of the memory chips and the processor.

I had to change the overclocking method. Having chosen Quake3 Arena (demo FOUR) as the shortest of the tests, I first increased the frequency on the processor, as soon as the test started to fail, I gradually increased the voltage. The same operations had to be carried out with memory. Each time, increasing the frequency by 10 MHz, he ran the test and, getting a new result, compared it with the previous one. And so, when once again, the resulting value turned out to be less than the previous one, I stopped. The new overclocking results for the GeForce FX5200 turned out to be more modest — core frequency 340 MHz, Ugpu = 1.57 V., memory 580 MHz, U mem = 2.75 V., but the performance of the video card has noticeably increased.

Not so long ago I tested the GeForce FX 5200 and GeForce 4 MX440-8X from ASUS, it was planned that these products would be reworked and tested, but, as usual, they were not at hand at the right time. Why two identical video cards, let’s consider only the GeForce FX 5200 for now, showed such different results, I still find it difficult to answer. The platform on which testing was carried out remained the same. The new version of Detonator 45.23 may have affected the performance of Leadtek’s product, but not to the same extent. I had to change the BIOS version of the motherboard, since the FSB frequency was floating with the previous version, which Doors4ever wrote about in detail recently. In fact, there were more questions than answers.