LCD Inverter Test - Laptop Testing for Backlight or Inverter Failure

Order book for $24.95 or printable ebook for $13.95

The Laptop Repair Workbook

Laptop Shopping

Computer Repair with Diagnostic Flowcharts

Copyright 2010 by Morris Rosenthal

All Rights Reserved

Copyright 2010 by Morris Rosenthal -All Rights Reserved contact info

The printable eBook version of The Laptop Repair Workbook is available for download anywhere in the world for $13.95.
I've been looking for an easy way to test for live laptop inverters for a couple years and I finally found a cheap, non-invasive method. The funny thing is I'd just ordered up a couple PC modding CCFL lamps with inverters to do a page about testing inverters with cheap replacement lamps. I'm not sure that would have worked given the impedance differences and the way inverters have to go through a timed sequence of voltage ramp up and down to strike and hold the plasma. Since the impedance drops when the tube lights and the plasma conducts, it's quite a bit more complicated than simply providing an RF power source. But as I was taking apart my old Toshiba screen today to expose the inverter leads for testing, it occurred to me to try the new Cen-Tech meter I picked up at Harbor Freight a couple weeks ago for $20. To the right I'm showing the zero (well, 10Hz is well within 1% of zero on a 20KHz scale) reading with no power to the inverter. Khz test reading zero from unpowered inverter
Field strengthens as probes near live inverter Inverters put out fairly high voltage, in the 500 V to 700 V range, and a very low radio frequency, between 35 KHz and 60 KHz on data sheets I've looked up. That's something you could easily pick up with a spectrum analyzer and a probe for either the electric or magnet field component, but the last spectrum analyzer I worked with cost around $30,000, so it's a bit out of the reach of the home consumer. Now, the neat thing about the Cen-Tech meter is it comes with a Hz measurement. It's limited to 20 KHz, after which the display will simply show a "1" for over scale. Rather than equipping the meter with a special probe, I just held the two standard probes a fraction of an inch apart, and the son-of-a-gun picked up the cyclic field for the live screen almost an inch away from the inverter output, as shown to the left.

The weak signal results in a lower than reality frequency reading, and as I moved the probes close to the inverter output, it simply went off scale. I'd try to be more technical, but the instructions that came with the meter were so vague about its capabilities and what it's supposed to read that I'm just assuming here. The important thing is that the inverter test worked, and at no point am I touching the probes to exposed wires or terminals. This beats the heck out of a test I saw a guy recommending where he sets a multimeter to high voltage and shorts out the live inverter. He reported he could get a transient reading before the inverter shut itself down, and with luck, it would still work after rebooting a few times. The inverter is designed for an RF impedance, not to drive into a simple DC resistive load like a multimeter, and I wouldn't be surprised if that voltage test has ruined more than a few inverters. Overscale reading as inverter output to backlight is above 20KHz
Finding live inverter inside laptop lid With my cyclic field test, I don't see how it can do any damage, unless you're careless with the probes and short out something in the screen. Then it occurred to me that the test could be truly non-invasive since laptop screens are so thin. To the left, I'm showing my newer Toshiba which I normally run plugged into a 19" Samsung monitor so I can see what I'm typing. A little hunting around on the outside with the spread probes and the meter picked up the field to the left. A further small move and to the lower left, you see the "1" indicating the field is off scale, higher than the 20KHz maximum frequency the meter can handle. Below, just as a proof, I'm holding the probes in the same spot with the laptop live, but the image diverted to the Samsung. Hunted around forever, no reading. So, this is probably the best use I've gotten from that MSEE I earned in the RF/Radar concentration 16 years ago!
Strong field (overscale) from working inverter No inverter output to backlight when using external  monitor
Of course, into every test procedure some rain must fall, and when I went out and tested some other random laptops, as well as a simple CCFL tube and inverter for modding, my $20 meter failed to register anything! So I borrowed a better meter from my neighbor, a Fluke 110 true RMS meter. The Fluke specs show it's rated to 50KHz, which turned out to be critical in the inverter test application. As the measurment to the right shows, the Toshiba I'd originally tested has an inverter frequency of around 33 KHz. The reading varies a little with the exact positioning of the probes, the air gap, and the noise on the leads, but something in the sub 40 KHz was clear. That's why my cheap meter that is spec'd to 20 KHz was able to pick up the field, even though it was over range, it was still within an octave. But higher frequencies are just filtered out or unmeasurable. Strong field (overscale) from working inverter
Strong field (overscale) from working inverter When I tested a Dell sub-notebook and my neighbors industrial rated laptop, I couldn't pick up anything on my meter. Using his Fluke, you can see that the inverter frequency was picked up as nearly 68 KHz. By this point, I'd taken to using a piece of paper to make-up the air gap, rather than trying to hold the probes apart. So I'm guessing the Fluke, which sells for a little less than $100 in the aftermarket, will cover all laptop inverter testing applications. The model 110 has been replaced by the model 115 which goes for around $110 new. If I have some time I'll design a little probe to function as an antenna, rather than just using the test probes, which involves a lot of fiddling around.
Of course, there are numerous things that can go wrong with laptop displays, so try to jump into testing the inverter before trying the obvious solutions. The first step, assuming you have power, etc, is to just plug in an external monitor and see if you can get that lit up. Newer laptops will usually autosense the presence of an analog monitor and shift the display without you having to use the function key (Fn-F5 on my Toshiba). If the external monitor works, you know that your video processor is good, that the laptop is booting into the OS, that the only problem is you can't see the screen. Wiring harness problems are common with laptop LCD failures because the screen is a moving part. The cables may fail in the hinge, or the constant movement and lid flexing may cause the signal connector to work loose of the LCD screen. It's also possible for the video connector to lift off the motherboard, especially if you're a heavy typer:-) When the external screen works, the cabling all appears good, and a very faint image is apparent on the laptop screen, you know that the backlight isn't lighting up. Not only is inverter failure more common than backlight (CCFL lamp) failure, but inverters are easier to replace. And now that it's winter, keep in mind that temperature has an affect on the voltage required to get the backlight to strike, so if the laptop has been sitting in a freezing car for a while and the screen doesn't light up, don't rush to take it apart. Give it a couple hours to warm up, but don't do anything whacky like sticking it in an oven or on a radiator!
The printable eBook version of The Laptop Repair Workbook is available for download anywhere in the world for $13.95.