Toughbook CF-30 Emissive Keyboard LED Mod

With reference to thread ' CF-30 emissive keyboard ... on the dim side' here: http://forum.notebookreview.com/threads/cf-30-emissive-keyboard-on-the-dim-side.762073/

I now have a spare Chicklet Emissive / Backlit LED keyboard to play with. I wondered if the DC-DC converter modification had actually been attempted by anyone?

I believe the standard voltage output is 5V, and was wondering what I could safely increase that too?

I'm unsure as to whether the four levels of backlight are controlled on the motherboard and manifest as differing voltage levels on the backlight cable, or whether this is controlled with the built in keyboard controller?

I can test this if no-one else has. I fancy seeing if I can double the current backlit level, which will still give me a useful range. I note that the keyboard appears to be factory sealed, so even though the previous post displays an representative image of the backlight LCD flexible PCB that realistically there is no way to gain access to that to in order to modify it in any way, at least not without some considerable 'hacking' and removing any water resistance inherent in the design.

Regards, Mark.

 

Here are the internals of a CF 19 backlit keyboard. Cf 30 is very similar.

http://forum.notebookreview.com/threads/cf19-keyboard-project.758542/

 

Mark,
You sound like you are familiar with electronics.
This schematic will help.

dropbox does not work properly...
Link and file deleted

ANY future PM requesting these files will be ignored.
ASK in the open forum. Please

 

Great, thanks. Having actually looked at that image now I can see that there is nothing fancy on that circuit board other than LEDs and resistors so the four-level control of voltage must be happening on the motherboard, so putting a DC-DC converter in between the motherboard connector and the end of the ribbon cable should allow the voltage to be stepped up to the LEDs.

I'm away this weekend but will take a look inside the CF-30 to see if there is room for the DC-DC converter I have kicking around.

 

The link doesn't work,it say it has been shut down due to high traffic

 

It does not surprise me. Best way to send a link is by PM. I see this thread has been viewed 137 times SO I wonder how many times Shawn file been downloaded

 

Short story....
Dropbox sux

The high traffic message is a catch all. They seem to block accounts for many reasons without any notice. They will not tell you what file is the problem either. Free accounts are given no support and pay accounts are not much better.
I will be looking for alternatives.

Mark has the schematics.

 

You need the file Maniac?

 

Nope,I use google drive.

 

sync.com looks pretty good.

Sorry to hijack the thread Mark.

 

No worries... I have all weekend before I am back and can make a start on the meat of the conversation. I understand the reservations with distributing schematics widespread.

 

Back on subject.
I am pretty good with electronics but, I am not an engineer so my help may be limited.

There are 2 engineers who lurk about. So please post any technical questions as Larry and Tish can answer much better than I can. Especially questions about circuitry fundamentals and design.

I will repost some of what you sent me privately so others with more edumacation can offer help..








So we have a peripheral IC generating a Pulse Width Modulated signal called BLKBD_PWM which controls a transistor driver Q127. I'm assuming VC5 is 5 Volts.


So the variation in intensity with be down to the mark-space ratio being generated on BLKBD_PWM, so assuming the switching frequency of the DC-DC converter I'm using is significantly greater than the PWM frequency boosting NEWLED_PWM via the converter should work.

 

i have been busier than a one-armed wallpaper hanger of late ... many irons in the fire and they are too slowly getting beat on .
also new "toy" that was saved from the "transfer station" : a 50" UHD 4K samsung ... a MOV and a line fuse and all is good .
i'll read through this thread around 2000 when i get back from the job .

heh ... "larry and tish" ... sounds like some locally produced radio show with a 250 W. transmitter , located in nowhere AZ .
(yep , it is a real place)

 

LED-dimming is done with PWM, so no constant voltage. KeyboardBL and LEDs on the Toughbook use the same. Why do u need more voltage on the KB? Taser-Keyboard against theft?

 

I'd like the backlight on the keyboard to be about twice as bright as it is now.

 

U can not simply overpower a LED through more voltage without killing them.
Every type of a LED needs a defined voltage and current too. PWM means the Votage and current are constant. Only the switch on-off rate vary on ur needs, the frequency is also constant and so high, ur eyes did not recognize it.
Only option IMHO is to exchange the LEDs to a better variant. But maybe the better LED is consuming more power (current) and so u kill ur mainboard....

To be exact: U can overpower a LED: with very short impulses. It is used in IR-lightbarriers or fiberoptics.

 

Thanks for the input. I guess the first step is to see what the voltage, frequency and mark space ratio is when the LEDs are being driven at full brightness. This will give an idea of how hard they potentially can be driven by over-voltage.

 

LEDs-off means no on-pulse and LEDs-full-on means no-off-pulse. The PWM u can only measure between on and off in dimming mode. So there is only a 5V maximum voltage (with at the moment unkown current). If u use a stepup buck to get maybe 6V, it could work, but will kill the LEDs in a short time. To find out buy a cheap white 5V-SMD-LED and try it first without the TB. U will see how the color of the LED is changing and how fast it dies.

 

Another thought is to take a lesson from a CF52GUN.
Use a standard keyboard and put the leds in the lid shining down on to the keyboard.

 

Buy a few cheap rubber keyboard to dissect and bench test the mods on. Guys here may have a few torn ones to donate.

 

I certainly concede that this a more complex problem than I initially thought. I think a lot depends on how hard the LEDs are being driven based on their maximum rating. On the one hand it wouldn't make a lot of sense to put high-power LEDs in and then current limit then with resistors that will just eat power and warm up, but on the other hand I can't help feeling at the current maximum brightness that there isn't some headroom to be tapped

Or should I just leave it? So far most of my purchases for this CF-30 have been a bit lame - I bought a 'new old stock' battery that turned out to have 600+ charge cycles, I got sent a DIMM module in a paper envelope that was (unsurprisingly) dead on arrival, the cheap Chinese replacement stylus has scratched the screen (I should have seen that one coming) and I got my GOBI1000 running after I had bought a GOBI2000 which I have yet to try but am not sure it is going to bring a lot to the party.

I did buy a duplicate chicklet keyboard that I could take the hit on and destroy in experimenting with the backlight but part of me thinks I should just be happy from now on!

Educational discussion though, that's always a win!

Regards, Mark

 

If I look at the purchases as being for educational purpose, then I do not get so annoyed when the item is not what I expected.

Someone sent me a DIMM in a paper envelope once also. I love giving honest feedback....

600 cycles does not mean the battery is bad. Cycle count does not really matter much.

Get a screen protector!

Consider the CF 30 as your learning tool for when you buy the CF 31. Which you will buy eventually..

 

Well - as it happens I was looking at either a CF-31 or a CF-30 when I got this. The CF-31 was BNIB, Core i5, base model, normal keyboard (IIRC) £500 (an absolutely steal if you have the cash in other words), or this CF-30 Mk3 with no HD (caddy included), backlit, 2GB, GPS, WWAN, £150 - second hand, long hours, but immaculate condition, oh but bios locked (but only impacting the wifi switch and bluetooth - neither of which is an issue). I *had* to get the WWAN working just so I could deploy that cool aerial with purpose of course, and there is something really quite satisfying about sitting in a car with 1:25000 ordnance survey maps rattling past with all their glorious detail.

I worried about the 1024x768 display being too small which is partly why I was reticent to drop £500 on the CF-31, but somehow it doesn't seem to get in the way as much as the 1366x768 display on the Lenovo X220 - go figure! That laptop would be perfect in a lot of ways with a better resolution display - I've put up with that for four years but at times it has driven me crackers. I think Win10 is probably helping with intelligent scaling, although I can't be sure.

I've just checked my bank account earlier - and regardless of how much I might lament passing up the CF-31 my bank account doesn't. What I have to be careful of is starting to 'level up' kit with eBay purchases. Getting an additional 2GB was a no-brainer for the cost, I had installed a spare 320GB HDD that I have replaced with a 240GB SDD (they are useful whatever, so not specific to the CF-30 if it goes south tomorrow), but the rest of my purchases probably shouldn't have happened. When you start totting things up, it makes no sense spending another £100.

Must... stay... away... from... eBay

Cheers, Mark.

 

Oh, and just to add, regardless of wanting to 'pimp' I am treating the CF-30 as the fifth member of the family, it is a seriously cool piece of kit. Can't believe it's been so long in the purchase...

I got serious kudos from a couple of young lads the other day - it's not often you make an impression with a laptop. I think it was the aerial that did it

 

funny to read and it is like a mirror to me

 

Yeh...Good luck with that.

 

I may have a go with the GOBI 2000 later, it is Fujitsu branded but I have the FUJI firmware/software. May be a case of backing up the Win10 partition first. I think I got the DIP switch settings, or something else random, wrong last time.

 

Just having a look now - according to BOB Johnson GOBI 1000 is 2G only, GOBI 2000 is 3g?

https://www.bobjohnson.com/gobi-2000-wwan-3g-mobile-data-card/
https://www.bobjohnson.com/gobi-1000-wwan-2g-mobile-data-card/

 

One board member got a Fujitsu Gobi to work in a Toughbook.

I "think" you can swap the 1000 for the 2000 without changing jumper//dip switches.

 

Woohoo - that's the Fujitsu Gobi 2000 installed! As it happens I did have to change the DIP switches (at least I followed a photo on this website), looking at the sledge with the DIP switches top right they are now set to down, up, up, down.

I used the Fujitsu UK software. Windows 10 then recognises the Gobi 2000 as an HSPDA modem.

Grand. Now I need to get myself out somewhere with a 3G signal to test!

 

yeppers ...

i would leave the backlight on the keyboard alone before you engineer the magic smoke right out of it .
resistors will not cut it when it comes to current limiting of this type of situation ...
leds drop in resistance as they warm up and to boot , they are current driven devices .
this means that they can go into the thermal runaway mode ... just like bipolar transistors .
for this situation an adjustable current regulator circuit is the best bet ...
and with a 5V max voltage , this means that the leds are singles hooked in parallel (figure a little over 3 volts for a white led) ...
and this is why current sensing PWM was used to control the brightness (the efficiency factor being secondary) .

regular tungsten filament lamps have a positive temperature coefficient ... they increase in resistance as the temperature of the filament increases (increase in voltage across and current through the filament) .
this is exactly opposite of leds ... and where most people fall short because they have not dealt with it directly .
(a florescent lamp is an example of a negative coefficient device)


a regular dmm will not accurately measure the total current due to the pulsed nature of the source .
the total max current can be measured accurately with a 'scope and shunting resistor (use ohms laws to figure out the max current from the voltage drop across the resistor) .

(the following is the plan of attack i would use)
however , if you want to experiment with parallel leds get some white leds that will mechanically fit into the keyboard (aint no good otherwise) .
back-of-the-napkin or desktop scribblin's ;
figure 40-60mA max current , this means the leds should be good for around 100mA (headroom ... tish can chime in on this aspect) . this should be plenty bright enough .
hook 'em in paralell and drive them with a current pass transistor (mosfet or bipolar ... your choice) driven by the PWM from the CF-31 (a buffer/level shift transistor or opto-isolator would be a good thing here for isolation and safety) .
the advantage with this type of circuit is that it is easier to measure the total current drawn ...
hook the ammeter in series with the source and bypass/filter with small value electrolytic cap and subtract about 5% from the total current indicated (inefficiencies due to switching transition) .
this will suffice and you will not need a 'scope to measure the current .

 

"I got serious kudos from a couple of young lads the other day - it's not often you make an impression with a laptop."

you sure can make an impression ... along with contusions and a concussion .

 

PWM full off and PWM full on are the two states, whithout any pulse. So it is quite easy to measure the current at full on with a normal DMM, I think

 

Thanks for the very useful info - I will read up some more.
Physically accessing the LEDs involves cutting into the seal, so I need to think long and hard as to how much I want this brighter. I like the idea of an opto isolator as I am very good at teasing magic smoke out of components, it is a talent that comes very naturally!

Regards, Mark.

 

On a road trip this weekend - sure is a luxury being able to 'compute' in broad daylight

 

Well, I decided in my infinite wisdom to swap my keyboard out to see if the replacement one I had bought worked so I could sell it as 'sold tested' on eBay rather than unknown condition. In the process of removing the existing keyboard I managed to crack the plastic just below the right alt key - by pushing the keyboard up and wedging my finger nail in the front of the keyboard and pushing - rather than loosening the adhesive tape first by gently lifting the keyboard. So be aware - if you are attempting to slide the keyboard up - lift it first so you know the adhesive has unstuck, as that gray plastic surround on these chicklet keyboards is surprisingly thin and brittle.

To add insult to injury when I installed the replacement keyboard I found that all keys worked except 'P' - so I've had to Araldite the cracked plastic back in place and re-install the original in worse condition.

Sometimes you should listen to your gut feeling and not bother in the first place, which is what my gut was telling me. If it ain't broke don't fix it!

The repair job is probably not noticeable on first glance, but I know it is there. I should have just gone to bed early!

 

I feel your pain.
Been there done that on various devices.

At least the one with the non working P is a great candidate for dissection.

WooHoo! Grab the screwdriver and let the solderin begin!!!!

 

scenario : three broadcast engineers (out of tech school for 2 years , 15 years experience (myself) and 40 years experience) that managed to put a lightning damaged 20KW BCFM online after 8.5 hours of work .
this was accomplished with two runs down the mountain in the wee hours for parts (the backup transmitter at the studios was identical and a couple of three phase 440 breakers grenaded) .
got the thing up and running ... not perfect but close enough .
the young guy kept futzing with it , trying to get the grid drive peaked and the plate drive dipped on the final (a little to the reactive side is best for stability ... tried to tell the young'un that) .
suddenly there was a "kerbam !" and *the lights went out in georgia* ...
it was pitch black in the shack ...
then the senior engineer said in a calm voice ... "yessir , he done engineered it right off the air" .

 

Rather than installing a DC/DC converter to increase the voltage you might want to try to change the resistors that are next to each LED to a slightly lower value. That will increase the current to the LED's.

Current through the LED's is a function of Ohm's law. E/(I*R) and E*I=W (E is volts, I is current in Amps, R is resistance in ohms, W is watts). LED's are highly efficient devices so most of their energy goes out as light.

When you overload an LED it starts putting out energy as heat.. heat kills LED's and they die a spectacular death.

You have to be cautious doing this, if you increase the current too much you run the risk of burning out the driver transistor that supplies power. Also higher current will shorten the life of the LED's. Usually they have a bit of margin to them.

The resistors look like black grains of rice and they are next to each LED. Use an ohmmeter and measure across the resistor. Whatever the value is, you can decrease that by a value of 20-30% (usually) without getting in to trouble. That will make it noticeably brighter and the brightness will work across all of the illumination ranges.

(FYI, if you wanted to be really neat you could replace the LED's with different color LED's to make the keyboard stand out. Since you would be that far on to the flex anyway that might be fun to do.)

You need to be careful doing this; if it is a PWM (pulse-width modulation) driver then the LED brightness is controlled by "flickering" the LED really fast at a duty-cycle that varies up to 100% (full bright). If you fry that transistor driver it very well may be on the main circuit board and that will not be fun to replace.

Those resistors play an important role, there must be some resistance on LED's otherwise they try to draw an incredibly high amount of current and the LED's will pop and burn out (they smell bad too when that happens).
-----------------------------
The DC/DC converter "would" work but it is unnecessary complexity; As I mentioned ohms law; a DC/DC converter attempts to make the LED's brighter by increasing "E" (voltage), A resistor change decreases (R) resistance and does the same thing with with only replacing very tiny resistors on the flex board. Good engineers prefer the most straightforward approach and to not "Rube Goldberg" something (to make something ridiculously complex for little benefit).

 

To change the color of the LEDs, anyone have the specs on what LEDs I'd need to get?

 

I thought someone posted something about them.. Maybe it was a guess.
You may need to get a sacrificial keyboard to dissect. I have a thread on CF 19 keyboards here. Find it for some helpful info.

At one time I had a CF 30 keyboard here with a bad key. Wonder if I still have it? You may need to remind me to look.
Maybe post a want ad in the buy sell trade forum. Jeff might have a bad one.

 

i wonder where tishers is at ?
she has not been around (or posted ?) of late .

i have noticed that the light on these keyboards is uneven ... to the point of not being able to read some of the keys in low light conditions .
the eyes adapt to the ambient and the screen and parts of the keyboard are bright and this leave parts of the keyboard in the dark (i'll take the pun) .
putting more leds on the board is part of the answer to this problem with the other problem being control of the brightness of the leds .
using the existing PWM circuitry , as was pointed out , is a risky undertaking (this i know from personal first hand experience) .
the possible solution would involve using a CdS cell , phototransistor or photodiode and the appropriate PWM support circuitry to *sniff* the brightness of a single original led to control the *brightness* of the add-on leds .
admittedly , this is a seemingly complex solution , however , the other simpler solutions would involve no adjustment (fixed brightness) or a separate manual control that does not "track" with the brightness of the original leds .

there are some neat little 3 channel (RGB) led controllers available ... this would allow one to control the color of the keyboard backlight :
https://encrypted.google.com/search?q=RGB+led+controller&ie=utf-8&oe=utf-8
there are so many that sifting through to find the right one for this sort of application (somewhat specific or niche) is time consuming .
(i get side-tracked by the offerings ... oooooh , shiny !)

i have seen the lenovo keyboard that glows a red color ...
it is bright and very evenly illuminated .
so , it has been done on a commercial level ... it can be done on a hobbyist level .

 

In my past life I had occasion to glow a red color..I THOUGHT that I was bright and illuminating.
I was on an enthusiast level, although I was so proficient at it I would have gone commercial if I could.

It's good to be sober.

I will take a quick look for that old keyboard. Maybe I already sent it to Larry.

 

/ me reaches over onto top of filing cabinet (it's good to have long arms) ... finds aforementioned keyboard .

yeppers ... got it !
or not ... this one has no missing keys and was sent along with/fits the '19 .
now i am curious ... i can't recall a dud keyboard being sent to me .
you did send a bigger keyboard ... it was for a 30/31 and was for swap out if needed/wanted .

i agree ... it is good to be sober ... sure beats being dead .
"i am an **shole by choice , if i want to be , or a decent person , if i want to be ... the choice/responsibility is mine"

 

I have a few extra keyboards to pull apart... I'd love to mod it for green LEDs... Possibly red...

 

That would look so cool. My Dell 1710 has at least a dozen light on it that you can change to about 10 different colors. Not sure how it works but it sure would look good on a Toughbook

 

I'd luv to somehow change the LEDs and switch the F12 to a Color change VS brightness. I can do the resoldering, but the rest is Way above my paygrade! ;-)

I have a few extra boards I'm going to tear apart here soon to get an idea of what they are like inside and see if I can't find out more about the LEDs in them. I have an old Modem with some red/green micro LEDs that might just be suitable as donors.

 

I like your idea.
You will probably need to source LED's from somewhere like Mouser. The LED's in the keyboards are SMD's.
Larry CWB32 can probably help with the switching circuit.

 

Just get the PWM-Signal from the power LED. Build a simple 2 Mosfet circuit, which is used to take the signal from the Power LED and switch the power of 5V onto others LEDs, like ur Kbd.

------

If u want just colors, it could be easier for u to use something like this: http://www.lilyleds.com/9305/5mm-RGB-Color-Changing-LED---Fast-Speed
Or maybe just use 2 colors 2wire LED and get a lot of colors by fast changing polarity in pwm-style with the help of a arduino stamp. In more simple way u will find the idea here: http://www.instructables.com/id/The-RedGreen-LED-Guide/

 

there are many ways to accomplish correction/improvement of keyboard lighting ...
the primary concern is that of "killing" the final "step" in the PWM drive signal .
many of the fellows that read these forums (including the googlers) have very little knowledge of the infernal workings of solid state devices (and even less about *hollow state*) .
it is pretty easy to *pop* the final signal drive mosfet or bipolar transistor .
by using the methods i mentioned , this will isolate the PWM drive signal from extraneous attempts at magic smoke generation and custom built animosity/frustration generators .
an opto isolator (with a mosfet drive output) could be used across the original leds ... this could be used to provide the drive to the mosfet circuit that onirakkiss mentioned .

i never noticed ...
does the power led dim with the keyboard brightness ?
on one hand it would make sense , on the other , if the keyboard is fully dimmed and the machine is used in bright ambient light ...