Overclock.net › Forums › Graphics Cards › NVIDIA › [Guide] How to test for and fix blurry/fuzzy analog video output on the GeForce 8400 GS
New Posts  All Forums:Forum Nav:

[Guide] How to test for and fix blurry/fuzzy analog video output on the GeForce 8400 GS

post #1 of 2
Thread Starter 
Background: How I discovered and fixed the problem

    My desktop PC originally came with a GeForce 7300 LE video card.  However, I needed better HD video playback and I wanted to be able to use two screens.  So when I came across a cheap GeForce 8400 GS about a year ago, I bought it and installed in the computer.  Immediately, I noticed that the picture on my screen was blurry compared to the GeForce 7300 LE, and there were very faint diagonal lines continually moving down the screen in white areas.  Additionally, I could no longer set the refresh rate of my main monitor to 60 Hz because the only option was 75 Hz, which isn't good for videos because normal video frame rate (for DVD and Blu-ray anyway) is 59.94 FPS.  Going into the nVidia control panel, I found that I could force 60 Hz using custom settings.  At 60 Hz, the screen got noticeably clearer, but it still wasn't crisp.  Over the period of a year, I got quite used to the fuzzy screen image.  Recently, I acquired a new HD camcorder (a Sony PJ-710V) and was disappointed at the lack of clarity in its videos.  However, when viewing its videos on another computer, I noticed that the same videos were perfectly clear.  That's when I realized that the problem was my computer, not the camcorder.

    As an electronics hobbyist, I knew that there were several possibilities why this video card had the problem.  One of the possibilities was that some ignorant engineer had put some components in the video signal path that would attenuate high frequencies.  So I created a test picture to confirm my hypothesis.  The picture has a row of white pixels, followed by a row that rapidly alternates black and white, followed by a row that slowly alternates black and white, followed by a row of black pixels.  Then, I got a camera and took a close-up picture of my LCD screen.  Sure enough, the fast alternating bars didn't make it all the way to black or white, and the slow alternating bars showed noticeable fading between black and white.

    On a background: I am regularly appalled at the utterly stupid things I see professional engineers design—at extra cost—into analog circuits.  Usually, improvement of the afflicted device comes at the removal, not the addition, of a dozen or more electronic components.  In other words, they could have saved money, and made a better device at the same time. Two of my favorites (Click to show)
  • A 450W PSU that used all four op-amps in an LM324 IC just to control the speed of the fan.  There were over 20 components including an expensive power transistor, an LM324 IC, resistors, capacitors, and diodes dedicated just to that fan control circuit.  In the end, the fan's RPM only varied slightly between room temperature (the PSU cold and just turned on) and the thermistor being shorted out (which represents an infinite temperature).  With the thermistor short-circuited, the PSU didn't even shut off, and the 12V fan only got about 9V.  I replaced that entire circuit with a single transistor and two resistors.  My circuit probably cost at least $5 less (with volume pricing) than their circuit, and the fan will now reach full voltage/speed before the PSU goes up in smoke.  Later, I added a diode and another resistor so the PSU will now shut off if it were to overheat.  The PSU is also much quieter now because the fan can actually slow down if the PSU is cool (it usually is).
  • The low-wattage PSU in an eMachines EL1200-07w computer that had so many bleeder resistors and aggressive AC-line filtering capacitors that the computer's idle power consumption went from 51.6W to 32.4W when the resistors were removed and the AC-line filter capacitors reduced to conservative values.  Now, the output voltages are much closer to their rated voltages, and I almost doubled the PSU's efficiency.  Not bad.  The only side-effect?  The PSU chirps quietly if it is powered on via paperclip with no load connected.  The static load of the motherboard (before it actually powers up) easily loads this power supply enough for it to be completely quiet and stable.  Now, the capacitors in the PSU will last a long time because they are no longer being cooked to death by the extremely hot bleeder resistors.  Many PSUs I have replaced capacitors in have need replacement specifically because of the heat generated by the bleeder resistors.

    So, I removed the video card to see what I would find this time.  Sure enough, between the outputs of the IC that generated the analog video signals, and the VGA (and DVI) jacks, there were inductors attenuating the high frequencies, followed by capacitors shorting out any high frequencies that managed to get through the inductors!  What were they thinking?!  Monitor signal cables are shielded, so they shouldn't have been concerned about RF emissions and complying with FCC regulations.  Again, I am appalled—the screen is blurry already, and I'm only running it at 1280x1024.  It is capable of 2560x1600—twice the horizontal resolution—which would require frequencies twice as high to pass through those inductors!

    So after verifying the signal path with a multimeter, I got out my soldering iron and removed the inductors and capacitors on the three (RGB) analog signal lines to the DVI jack.  I left the others in, because they would not affect picture sharpness, only the timing signals (which were working just fine).  Then I put a solder blob in place of the inductors, and left the capacitors' spaces open.  Then I reinstalled the video card and booted the computer.  My secondary screen, which is a small, old LCD screen connected to the DVI jack via an analog adapter, was noticeably clearer.  I hadn't previously noticed it being blurry.  So I powered down and did the same thing for the VGA port.  Upon booting back up, I was disappointed that my main screen looked the same.  But when I loaded my test picture, it was shimmering and flickering violently.  So I hit the "Auto adjust" button on my LCD screen.  In a few seconds, everything was razor sharp.  I hadn't seen such picture clarity from my main screen in a long time!  In fact, the difference was so big that I had to recalibrate ClearType because text was so sharp, thin and crisp!  Just to confirm that it was displaying correctly, I took another picture.  Every time I come to my computer even these several days later, the clarity of the screen still grabs my attention.  Additionally, after replacing a bulging 470uF capacitor on the video card, the light diagonal lines have disappeared.  Yay!


How to test your video card for high frequency attenuation

    Make sure that your monitor is running in native resolution.  Also make sure that your web browser is displaying the image 1-to-1, without scaling the image.  Run the auto-adjust feature on your monitor, and then get good digital camera and take a close-up photo of this test graphic:
 

    Do look at the picture above first, and make sure that the second row (with the fine lines) is displaying evenly and cleanly.  If it isn't, your web browser is probably scaling the image, which will render it useless for the purposes of this test.  Chances of this are high.  In that case, you may try changing the zoom setting in your browser, or downloading the picture and displaying it in a good picture viewer like IrfanView.


    Here is my "after" photo.  This is what you should get when you take a close-up photo of the above test graphic on an LCD monitor.  Notice how the black areas are totally black, matching the black row on the bottom of the photo:
 

    If your photo shows clear blacks like this one, your video card is outputting a good, crisp signal, and there is no need for modification.


    Here is my "before" photo.  Notice that the alternating lines on the second row don't go all the way to full intensity or all the way to black (as defined by the black row on the bottom).  Also, note how the slowly alternating lines on the third row fade to black over two pixels (two RGB pairs).  If the dark spots in the picture below look black to you, or if you only see one set of pixels fading to black on the third row, your monitor is not adjusted properly; try increasing the brightness or gamma settings on your monitor:
 

    If your photo looks like this, with the intensity from the light colored pixels bleeding horizontally into the black areas, your video card probably has a high frequency resolution problem and may benefit from this modification.


Disclaimer

    If you have no experience soldering, I strongly recommend that don't attempt to do any of this.  It is easy to break something, and we are working with very small parts here.  Your soldering will have to be accurate to the millimeter.  Also, the pads on the circuit board are very fragile.  You will need to be prepared to make the connections another way if a pad comes off.  Two pads came off for me while removing the offending components; but in both cases I was able to make the connection elsewhere.  In addition, you will need to know how to use and understand a multi-meter in order to verify the signal path on your video card, which I expect will look different from the photos below.
    Please note: If you're using the DVI port in digital mode, or if you're using HDMI, this mod does not apply to you.  This problem affects analog signals only, from the VGA port or using a VGA adapter on the DVI port.  If you are using a digital connection and are experiencing blurriness, the problem is most likely with your monitor, or is due to a resolution mismatch.
    Proceed at your own risk.  If you know what you're doing, and are familiar with the principles of electronics and with surface-mount soldering, chances of permanently breaking anything are very low.


How to remedy this problem on the GeForce 8400 GS

    Since my GeForce 8400 GS has a DVI port and a VGA port, there are two parts for this modification.  Please note that different manufacturers make video cards with the GeForce 8400 chipset.  Some (or many) of these will not have the same circuit layout, problem, or features that I am describing here.

    On the DVI port, you’ll need to identify the R, G, and B signal lines.  I did this by testing for continuity between analog pins on the DVI port (those are the four pins by themselves at the bottom of the DVI port) and the components you see in the center of the picture below.  The inductors are basically a "jumper" or "short" that conducts electricity with very low resistance until high frequencies try pass though.  They are also marked with the "L" prefix (L1, L2, L3) on circuit boards.  You want to find the ones that pass the R, G, and B signals, and replace them with a solder blob that will pass high frequencies (fine lines) just as well as low frequencies.  Next, find the capacitors (if your video card has them in the signal path).  Capacitors are marked with the "C" prefix (C1, C2, C3) on circuit boards.  First, verify that the capacitors are in fact doing the dastardly deed of shorting the high frequencies to ground.  One side should be connected to the signal line (be it the R, G, or B line), while the other should be connected to ground (usually the card's slot bracket is also grounded, so you can connect one side of your meter there).  If that is the case on your video card, remove the offending capacitors also.

    In the picture below, you can see my solder blobs across the inductors positions, and the spots where I removed three capacitors (compare the 3 modified groups of components on the left with the two untouched ones on the right).  Notice that on the leftmost one, a pad came off when I removed the inductor, so I had to solder to the diode (the black device).  Your video card may or may not have diodes like this, so carefully check the signal path with a meter before making any changes so that you know what the options are if you were to lose a pad.
 

    After finishing your work, make sure that there is no short-circuit between ground and any of the R, G, and B signal lines, and then give the video card a test.


    The modification for the VGA port is identical: Identify the VGA port's R, G, and B signal lines, find the inductors and short them, find the capacitors and remove them.  Here is a picture of the VGA mod on my card:


    I hope that this may prove helpful to someone else as it did for me. thumb.gif
 
Edited by Techie007 - 11/12/13 at 12:12pm
My desktop PC
(14 items)
 
  
CPUMotherboardGraphicsRAM
Intel Core i7-3770K Gigabyte P67A-D3-B3 NVIDIA GeForce 8400 GS  1x Corsair 8 GB 
Hard DriveHard DriveHard DriveOS
Kingston SV300S3 WesternDigital WD10EZEX Samsung HD154UI Windows 7 Ultimate SP1 x64 
MonitorMonitorKeyboardPower
Daewoo L947BK Gateway FPD1530 HTK-2001 Dynex DX-400WPS 
MouseAudio
Kensington K72400 Realtek ALC889 
  hide details  
Reply
My desktop PC
(14 items)
 
  
CPUMotherboardGraphicsRAM
Intel Core i7-3770K Gigabyte P67A-D3-B3 NVIDIA GeForce 8400 GS  1x Corsair 8 GB 
Hard DriveHard DriveHard DriveOS
Kingston SV300S3 WesternDigital WD10EZEX Samsung HD154UI Windows 7 Ultimate SP1 x64 
MonitorMonitorKeyboardPower
Daewoo L947BK Gateway FPD1530 HTK-2001 Dynex DX-400WPS 
MouseAudio
Kensington K72400 Realtek ALC889 
  hide details  
Reply
post #2 of 2
This would have been great 7 years ago wink.gif
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: NVIDIA
Overclock.net › Forums › Graphics Cards › NVIDIA › [Guide] How to test for and fix blurry/fuzzy analog video output on the GeForce 8400 GS