Tutorial: Connecting a Wirz SLI-OEM to a CFAH* Series Display

Disclaimer:

Use this tutorial at your own risk. If you let the magic blue smoke out of your PC, LCD, or SLI-OEM that is your responsibility.

Some people have asked us what was involved in connecting one of our CFAH* series character displays to a Wirz SLI-OEM "serial backpack". We thought we would find out.

First up is to have the SLI-OEM Manual on hand. You will also need a resistor, soldering equipment, some electrical tape, some double-sided foam tape, and perhaps some ribbon cable and connectors, as well as the SLI-OEM and your favorite Crystalfontz CFAH* Series LCD.

Here is what you get from Wirz:



Here is a back view of the SLI-OEM:



There is some tape covering the switches on the SLI-OEM. Might as well take that off. Don't be afraid . . .



There were some ugly tabs on the SLI-OEM I received. They are from the way the PCB's are processed with several of them on a panel at once.



You can nip those off . . .



to leave a nice smooth edge.



Here is the display (CFAH2002A-RMC-JP) that will be connected to the SLI-OEM.



Wirz supplies a 16x1 "Berg strip". That works un-modified for some displays, but for others we really need an 8x2--so the handy diagonal cutters help us out again . . .



Now we can use the two halves of the Berg strip to do a test fit of the display with the SLI-OEM.



The tabs that hold the bezel on the display look like they could short to some of the unused connectors on the SLI-OEM:



A couple strips of electrical tape should prevent any trouble:



Next up would be adding the connector that will supply power and communications from the host (probably a PC in most cases). Just place the connector . . .



and solder it.



The SLI-OEM has a nearly-nifty feature on it that will allow control of the backlight using its "/OUT" pin. "Nearly nifty" because you can control the backlight, but you have to do some hacking around to add the necessary current-limiting resistor. Wirz should have just put two holes in their board where you could add your own current-limiting resistor. Oh well, it gets covered up in the end so we will just scab a resistor into place.

Now the math end of things. We need to calculate the size of the resistor.

R=(5v - Vled) / Iled

You need to look up the current and voltage of the backlight from your LCD's data sheet. The CFAH2002A-RMC-JP's data sheet gives these values:


Iled (normal) = 0.28 amps (280mA)
Iled (maximum) = 0.56 amps (560mA)

Vled (normal) = 0.42 volts
Vled (maximum) = 0.46 volts

So:

Rmax = (5 - 4.2)/0.28 = 2.8 ohms (normal)
Rmin = (5 - 4.6)/0.56 = 0.7 ohms (maximum brightness)

I happened to have a 2.4 ohm, 1 watt resistor handy. You can have one too if you go to http://www.digikey.com, type "BC2.4W-1CT-ND" into the "Parts Search" box and click on the "Go" button. The actual current through the backlight for the 2.4 ohm resistor we selected will be:

Iled = (5 - 4.2)/2.4 = 333mA

Now we need to make sure that the power dissipated (as heat) in the resistor will not create any of that nasty blue smoke.

Presistor = Vresistor * Iresistor = (5 - 4.2) * 0.333 = 0.267 watts.

I like to "derate" a resistor's power rating by at least 50%. This keeps the resistor a bit cooler. So for this case a 1/2 watt resistor would probably be fine, and the 1 watt resistor that we selected will be in great shape.

So now it is time to put the resistor on the SLI-OEM in such a way that the SLI-OEM's /OUT pin can be used to switch the backlight on and off. First form the resistor's leads around the pins shown here . . .



then solder them:



Now we need to jumper the Vcc (+5v) pin to the V+ (LED "A") to complete the circuit. Use one of the cut-off resistor leads to form a jumper:



Then solder it:



Now we need to place some double-sided foam tape (which, of course you have lying around, so Wirz did not include it in the kit) in such a way that it will support the SLI-OEM and keep things from banging around. This is a little tricky, since the resistor is in the way, but I think you can handle it:



Now it is time to put the 2x16 berg strip between the LCD and the SLI-OEM, remove the remaining cover from the foam tape and solder the SLI-OEM to the LCD. From the LCD side . .



and also the SLI-OEM side:



Now you need to come up with a cable. I used a Digi-Key part number C1AXG-1036G-ND that I happened to have lying around:



Be sure to line up Pin 1 of the connector on the SLI-OEM (the mark is covered by the connector supplied with the SLI-OEM, but it is right by the "J1" part designator) with the red stripe of the cable.

Now if we prep the end of the cable as shown we will have access to the signals we need:



The Vcc (+5v) can be connected to a red wire of the PC's power supply, the Ground can be connected to the black wire of a PC's power supply.

To connect the RS-232 I used a connector I had lying around (Digi-Key part number CFM09G-ND), and connected the "PC Tx / LCD Rx" to pin 3, and the "Ground" to pin 5:



Now it is time to set the dip switches (19200 baud, 20x2 shown):



Set the contrast:



I had to set the contrast all the way counter-clockwise to get a good display.

And power it up:



So, how does it stack up?



Here is a view of the completed project. The LCD is right-side-up, the SLI-OEM ends up being up-side-down with this display.



Closing:

I have heard of some power-up problems with the SLI-OEM. In my short experience, if power is applied in one sharp step it seems to come up OK, but if the power glitches around the SLI-OEM can get into a weird state.

The SLI-OEM detects the "sense" of the RS-232 line by sampling the state of the PC Tx/LCD Rx line at power up. This is great if your system can guarantee that the PC Tx/LCD Rx line will be valid before power is applied to the Wirz, but if you power it up then connect the serial cable it seems to guess wrong every time. The sense would be better implemented as a jumper or switch. I guess an idle-state detection routine would also work (Any time the PC Tx/LCD Rx line is in a given state for the time span of 10 characters, then that state is the defined as the idle state and the sense of the line would be re-set based on that new idle state).

It would also be nice if it displayed something at power up. The only thing shown is a non-blinking underscore cursor, which was pretty hard to see when the contrast has not yet been adjusted.

The contrast should have more range. You should ba able to adjust the display from all the way light to all way black--IMO. This would probably mean making a negative voltage generator on the SLI-OEM, which would only take a few passive components and a toggling port pin . . .

Other than the above caveats, SLI-OEM does seem to work as it should.

Thanks to Mark Rouleau for letting us use his SLI-OEM as the guinea pig for this tutorial.

Is there a reason for connecting the current limiting resistor
between the LED and ground rather than between the
supply and the LED as shown on the Wirz documentation?

Were you actually able to turn the backlighting off in
software? (We weren't.)

Tom P.

No. As long as the current flows from Vcc through the current-limiting resistor and backlight, and then to the /out pin, it should work.

The only difference is that the resistor and the jumper would change places.

We were able to switch the backlight on and off with the /OUT commands (decimal 17 & decimal 18).

Comments From Wirz

I received an e-mail from the folks over at Wirz. Here are their comments:

1)  We've had better luck using double stick tape as an insulator between the two boards rather than electrical tape.  I always just put two pieces the full length of the LCD making sure to cover the metal tabs of the glass holder.

2)  The idea of putting a spot for an onboard current limiting resistor was discussed thoroughly when we did the design.  The problem is that our goal was to simply pass power through the connector to the pins 15 and 16 of LCD header.  We wanted to support EL back lighting as well as LED back lighting.  Some manufacturers hook pin 15 to cathode some hook pin 15 to the anode.  I actually wanted to make it so that !OUT would work right off the bat by just installing one resistor but then this would require 4 resistor pads to handle all of the combinations and it would be trickery to handle EL support.  Rather then attach the resistor to the pins as you did in the app note, you may find it easier to simply put the resistor in-line with the ribbon cable lines.  We just solder each end of the resistor and cover the whole thing with some heat shrink and call it done.

3) One of things I wanted to bring up is that you mentioned that some customers might use it inside a PC.  In this case, the user would want to add some power supply conditioning.  I've seen PC's have several hunderd millivolts of ripple on the 5V supply.  We only have .1uF cap across the supply which would only be effective at high frequencies.  A 10 uF and .1uF in parallel will do a much better job.

4) Concerning the power supply coming up, we have had a few customers who mistakenly attributed problems in their circuit to the SLI-OEM.  In all the cases we were able to help then work through the problems.  Its usually result of poor power supply decoupling or incorrect handling of the start up condition.

5)  This is the third Rev of SLI, the first was SLI, then SLI-OEM Rev A and finally SLI-OEM Rev B. We've always used the date polarity detection scheme in all the revs and I haven't had any customer who haven't been about to get it to work.  For a PC, it works without any extra work as the RS-232 port is in its marking state on boot up.   A micro controller should be able to easily drive it into the correct state before the detection is done.  Other cases could use a pull up/down but I can't think of any who has had to do this.  Its actually works quite well in our opinion, we would have changed it in one of the revs otherwise.  And it comes down to what dip position would you trade for it?  Four of the serial speeds is really the only option I guess.

I want to thank the folks over at Wirz for taking time to look over this tutorial and offering their comments.

You should ba able to adjust the display from all the way light to all way black--IMO. This would probably mean making a negative voltage generator on the SLI-OEM, which would only take a few passive components and a toggling port pin . . .

Click to expand...

Hey crystalfontz would you mind going into a little more detail with this idea? I'd be more than happy to let you use my SLI-OEM and CFAH2004A-RMC-JP as a guinea pigs. please Also, is there a way to use a pot to control the backlight, especially when the contrast can be fully adjustable? It would suck to open up my case everytime i want to adjust the contrast.

BTW, awesome tutorial!

The easiest way to do the contrast would be to put a 10K pot on the Vo pin of the LCD--just do not connect that pin to the SLI-OEM at all.

Three pins on a pot:

* middle to Vo (pin 3 on the LCD)
* one to +5v (pin 2 on the LCD)
* one to ground (pin 1 on the LCD) {or to -5v if it is available}

For the backlight, you could do the same thing, +5v to one end of Rlimit, the other end of Rlimit to the center of a 100 ohm pot, one of the ends of pot to the LCD's "A" backlight terminal, LCD's "K" terminal to ground.

For both the backlight and contrast, do not connect the pins to the SLI-OEM at all (LCD's V0, A, & K).

To do the "negative voltage generator" would require a firmware change on the SLI-OEM. Can't wire around that one.

Ok this is all too confusing

Originally posted by VIRUS
Ok this is all too confusing

Click to expand...

That is why we make the serial LCD modules. Just plug them in

Ok here it goes i have the CFAH1602B-TMC-JP

On the SLI-OEM SW1 and SW2

1 row
1 row split
2 rows
4 rows

What should this be set at, DIP switch 1 is set to on, 3 is set to on, 5 is set to on, 2,4,6,7 and 8 are set to off.Is this right for my LCD

I happened to have a 2.4 ohm, 1 watt resistor handy. You can have one too if you go to http://www.digikey.com, type "BC2.4W-1CT-ND" into the "Parts Search" box and click on the "Go" button. The actual current through the backlight for the 2.4 ohm resistor we selected will be:

Iled = (5 - 4.2)/2.4 = 333mA

Now we need to make sure that the power dissipated (as heat) in the resistor will not create any of that nasty blue smoke.

Presistor = Vresistor * Iresistor = (5 - 4.2) * 0.333 = 0.267 watts.

I like to "derate" a resistor's power rating by at least 50%. This keeps the resistor a bit cooler. So for this case a 1/2 watt resistor would probably be fine, and the 1 watt resistor that we selected will be in great shape.

Click to expand...

This is the part that confuses me.

I know this is a pretty old thread, but i've read so many good things about this SLI-OEM backpack, but i've been told by wirz that they no longer make / sell this !


Does anyone know of where i can get one or one very much like this one ??