I picked up a batch of LCD panels from Goldmine. They come essentially unmarked except for a few of them have a designation of Planar systems inc. (they make whole units, and I have yet to find any data sheets on their site for anything)

I assume they are marked not because they manufactured them, but rather because they had used them as a part in a display and simply marked them with their name and sku.

Anyway, in my hands I have 5 panels. 3 of them were relatively easily to figure out. They had bilateral symmetric pin outs and essentially pin one was grounded and all the others tended to light up some segment of the displays. Of course some of the pins essentially did nothing. One 5 digit display had enough pins to be a 6 digit display and by extrapolation the unreactive pins would map to the phantom 6th digit. One of the 70 pin display had 18 non-reactive pins but I could find enough pins that would light up all the segments and dp's etc.

** HOWEVER **

The other two panels are not behaving in the same way, so I would like a pointer in the right direction with a strategy for figuring it out.

#1
has a 1-3/4" x 11/16" viewable area
has 20 pins all on one side of the panel with no pins on the opposite side
It has 4-1/2 digits with a superscript B and a dash below it in position before the first 1/2 digit. I can't seem to get to a place where one pin is common and all the rest light up a segment, so I assume this works differently. I am open to suggestions. Or a data sheet if this is a recognizeable panel.

#2
has 3-5/8" x 5/8" viewable area
it has 24 pins on one side and 6 pins in 2 sets of 3 on the opposite side. With the multimeter I have determined that each of the 3 pins on one side appears to be connected to a one on the other side. So if we called them ABC and DEF then A-F are at 0ohms as are B-E and C-D. I am not sure what this means, but it seems to be a significant piece of data.


It has 8 digits, each followed with a dp including the last. Underneath each digit is the reverse of a circumflex.
If I jump from one of the pins in either group of 3 to any single pin on the other side I either get B/C/dp, A/G/D, F/E/circumflex to light up. So I can walk across the pins this way and sight up all the segments but only in groups of 3 each pin.

My exploration technique is quite crude. I have a battery with two leads. I attach one to some pin (that I assume is com) and touch all the others one at at time until I notice something or the device releases its magic smoke. The latter is yet to happen I do have enough discrete components to actually properly run an experiment, and I have some atmel micro-controllers that i can program to do whatever needs be.

I just don't really have a good strategy or frame of reference on these panels to know where to start.

Any advise and pointers would be quite welcome.

BTW, I am new here, Hi!

-- Absinthe

 

Many of these are going to be multiplexed in some way but sometimes contain single little indicators too. If Goldmine had a clue as to what the pinouts were they could sell them for more money, but since they probably just came as bulk in a box on a pallet that could have already been through other surplus electronics places hands you're always just taking your luck on grab bag situations. In essence they figure if you can come up with one or two that work and can be put to use you got your money's worth.

I like their surprise boxes. While a great portion of the contents are of absolutely no use to me I usually find enough things in them to make sifting through one worth my while.

 

Thanks Marshallf3 --

I appreciate your taking the time to reply to this thread. This is my first time with "grabbagging" from Gold Mine. I recognize that they were 5 for $2 because they were unidentified and without driver circuits. The fact that 3 out of the 5 were easily mapped via brute force was pretty satisfying.

Although I agree with every thing you said, I still feel you failed to actually address my question. That may be my fault for being so verbose in describing what I had, and not make clear what i was looking for. Let me restate the question as this:

Strategically, how do I figure out what the pinouts are? I came up with one strategy initially, putting a battery across different pins and "see what happens." That worked for 60%, and at least identified the screen layout for the other 40%. What is the next thing to try?

Obviously, there are more than one ways to drive these things. This was the only way I could imagine, thus the technique. What other methods are there and how could I simulate one of them for verification? This is perhaps more about learning and exploring as it is about getting a useful display. This is also less than the first time I have had call to use any kind of display, so I don't have any transferable concepts from having used other displays.

I assume, that through my travels I will encounter derelict electronics begging to be recycled. I can read or test many discrete components. There must be some way to do something similar with things like this...

 

I am assuming that you mean bare LCD without driver electronics. Otherwise, I would think that:

putting a battery across different pins and "see what happens."

would, more often than not, be liable to blow something so that no method will enable you to find the pinout from that point forward.

Bare LCDs are somewhat less fragile than electronics, but lots of pins mean lots of possible combinations. Probably a linear search is a reasonable place to start, it just might take awhile.

The solution I usually use for unknown devices in general is to try to find something online that is similar, it doesn't even need to be exactly the same thing. If you know or can discover the manufacturer, you can often look at pictures on their website till you find something close (you may have already tried that). Often, families of devices have identical pinouts even if not exactly the same.

Bare LCDs are a little different. They can be made much like circuit-boards, custom for each application, so each one may be entirely different. Displays with driver electronics are more likely to be made according to some sort of standard.

If the device includes driver electronics, pinouts for similar devices, besides keeping you from blowing things, often even share the same commands/protocols.

 

Thanks DonQ --

That is exactly what I am talking about. They are bare LCD panels. With enough poking I can get all the segments to light up, but certainly not the way anyone would want.

I have searched a lot online but haven't been able to find anything similar enough to give me too much guidance. That, primarily, is why I am asking here. My hope is that by description, someone will recognize what I have. The fact that some are marked with Planar Systems, agrees with your suggestion that they are purpose built for some system, since that is what they appear to do from what I can gleen from their website.

The one that intrigues me most at present is the one having 24 pins on one side and 6 on the other. It seems that powering any one of the 24 pins while grounding any of the 6 pins will set exactly 3 segments. So for each 3 pins I power I end up making an 8 with a dp and an inverse circumflex below it. If I power from one of the 6 pins to the next one (lets call it A-B) I get all A characters. If I power the first one to the 3rd let's call it A-C I get all top 3 segments along with the bottom segment and the dp and the inverted circumflex. It just seems like I am soooo close Anyway, pointers and suggestions are welcome. I will give anything a try and be happy to report back the results.

 

I've tried before and given up almost every time on the oddballs. Sad too, as I've got a couple that I can visibly see the little pictures inside (they're automotive) and I had a place I could have used them if I could have ever figured out how to drive them. I don't think they're bad as I can get some segments to fire, others just won't regardless of what I try.

 

The other panels may be COMMON ANODE opposed to COMMON CATHODE.

So if you are poking around, you might have your ground wrong.

I find it easiest to set your multi-meter in DIODE TEST mode then hold the neg lead on 1 pin and drag the positive lead across the others. Then switch. Put the pos lead on the pin you just had the neg lead on and drag the neg lead across the pins.

Often the corner most pins are the common..be it cathode or anode.

Give it a whirl.

If it doesn't work, you can say "Oh, what a drag." (rimshot please! )

[ed]
added rimshot
[/ed]

 

Ok, I finally found some things that line up for this beast.

1st thing I found is that it is a VIM-838
I found the datasheet for that that suggests that the commons are as I had suggested and further suggests that they be connected when used on pcb. It gives the layout for each pin-common-segment relationship. Of course, that is still not what I am observing.

But I also found this:
http://www.datasheetarchive.com/Indexer/Datasheet-017/DSA00301502.html

which suggests that it needs to be used with AC. I am not sure how to pull that trick off yet, but at least it is a point in the right direction. Any idea? Suggestions for what voltage of AC?

I am excited, I am so close on this one that I am ready to call it 4 out of 5

 

Hi Absinthe,
Considering the lower pin count compared to segments, your LCDs are multiplexed. With appropriate lighting angle, you can see the multiplexing layout on some LCD.
If not, make electrical tests : LCDs are just like (very high impedance) lamps, just apply a voltage and the segment will appear. Use 3V with a 100k series resistor (to avoid undesirable smoke) and test all 2 point combinations.
Once you get the pinout, consult some documentation on LCD multiplexing. Search for example at National Semiconductor, they have made multiplexed lcd drivers for ages and even have uP with that and some very good documentation. Or maybe you should begin first knowing how multiplexing works to devise a test plan.

You can multiplex-drive LCDs with general I/Os of an Atmel (some resistors for generating the backplane signals are needed). If interested, I can tell you how.
http://forum.allaboutcircuits.com/member.php?u=93982

 

The other panels may be COMMON ANODE opposed to COMMON CATHODE.

So if you are poking around, you might have your ground wrong.

I find it easiest to set your multi-meter in DIODE TEST mode then hold the neg lead on 1 pin and drag the positive lead across the others. Then switch. Put the pos lead on the pin you just had the neg lead on and drag the neg lead across the pins.

Often the corner most pins are the common..be it cathode or anode.

Give it a whirl.

If it doesn't work, you can say "Oh, what a drag." (drumroll please! )

I tried this and I have to say it is nice. It is also probably a little nicer to the component than my 9 volt battery was.

BTW, it is actually called a "rimshot" after the joke http://instantrimshot.com/

Click that link, and they have a button that will give you rimshots all the time. So if you have speakers at work, you can entertain all your coworkers with your standup routine and punctuate it with rimshots just like on TV.

I hereby disavow any knowledge of this, and if your boss fires your for performing standup when you should be working, don't blame it on me. You do the drum roll before someone does something stupid Like attempting to do standup at work

 

which suggests that it needs to be used with AC. I am not sure how to pull that trick off yet, but at least it is a point in the right direction. Any idea? Suggestions for what voltage of AC?

LCD segments are made of electrolytes which polarizes light when a voltage is applied. So the voltage must be AC (10hz to 100hz) with a zero DC component, otherwise, the electrolytes would progressively migrate to electrodes and destroy the segment.
That's the theory. In practice, I have a 4 digit LCD which is driven with ICL 7107 (an ADC normally used with LED displays). So it's lit with DC signal and no visible degradation has occurred after more than 10 years in use (as an V indicator in an old car).
Anyway, you'll have to multiplex the LCD you bought anyway, so applying an alternating signal is no big deal and desirable.

 

which suggests that it needs to be used with AC. I am not sure how to pull that trick off yet, but at least it is a point in the right direction. Any idea? Suggestions for what voltage of AC?

Both the liquid crystal material and the alignment layer material contain ionic compounds. If an electric field of one particular polarity is applied for a long period of time, this ionic material is attracted to the surfaces and degrades the device performance. This is avoided either by applying an alternating current or by reversing the polarity of the electric field as the device is addressed (the response of the liquid crystal layer is identical, regardless of the polarity of the applied field).

This is an excerpt from the Wikipedia article - http://en.wikipedia.org/wiki/Liquid_crystal_display

The AC voltage is small, on the order of 3 volts at 60 Hz.

 

Hi Absinthe,
Considering the lower pin count compared to segments, your LCDs are multiplexed. With appropriate lighting angle, you can see the multiplexing layout on some LCD.
If not, make electrical tests : LCDs are just like (very high impedance) lamps, just apply a voltage and the segment will appear. Use 3V with a 100k series resistor (to avoid undesirable smoke) and test all 2 point combinations.
Once you get the pinout, consult some documentation on LCD multiplexing. Search for example at National Semiconductor, they have made multiplexed lcd drivers for ages and even have uP with that and some very good documentation. Or maybe you should begin first knowing how multiplexing works to devise a test plan.

You can multiplex-drive LCDs with general I/Os of an Atmel (some resistors for generating the backplane signals are needed). If interested, I can tell you how.
http://forum.allaboutcircuits.com/member.php?u=93982

Thanks!! I will lookup some multiplexing information. As of yet, I still haven't done any 12c with the atmels, but it is on my list. I do have some port expanders to play with and some other ics.

I am guessing that the amount of dc voltage I am feeding this thing is simply bleeding into all the segments on a given pin rather than just lighting up the one that I am intending.

 

The AC voltage is small, on the order of 3 volts at 60 Hz.

So I guess my immediate next project should be a 9v dc to 3v ac inverter circuit. I wonder if I just found another use for a 555

 

I am guessing that the amount of dc voltage I am feeding this thing is simply bleeding into all the segments on a given pin rather than just lighting up the one that I am intending.

That's right. On multiplexed LCD, the way to light or not a segment is determined by the voltage threshold. In general, at <1V the segment is off, at about 3V, the segment is on (for 3V LCDs but with 1.5V LCDs, divide values by 2). So if you apply 6V, the voltage will "bleed" through the backplane signal and you won't be able to discriminate segments from backplanes (that is if you apply 6V to 2 segments, they will switch on a lot of segments, not a really pinout dectection method, right ?)

 

That's right. On multiplexed LCD, the way to light or not a segment is determined by the voltage threshold. In general, at <1V the segment is off, at about 3V, the segment is on (for 3V LCDs but with 1.5V LCDs, divide values by 2). So if you apply 6V, the voltage will "bleed" through the backplane signal and you won't be able to discriminate segments from backplanes (that is if you apply 6V to 2 segments, they will switch on a lot of segments, not a really pinout dectection method, right ?)

Leaning is taking place. So how do I figure out if it is a 3 volt vs a 1.5 volt lcd?

 

How could I have forgotten the rimshot? I feel like such a has-been comic.

Well the method was good even if the joke was not.

 

Ok, so digging a bit deeper, I see that by definition a segment is "on" when it is out of phase with the common. If I simply use a ASMV to give it AC will that work? Essentially square wave pulsed DC... is that close enough? Or do I need to actual have out of sync waves? Am I over-thinking things now?

 

Leaning is taking place. So how do I figure out if it is a 3 volt vs a 1.5 volt lcd?

Start testing with 1.5V, if no segment is ON, then test with 3V, then 5V... Most probably, your LCD is 3V.
BTW, don't lean so much, you'll fall

 

If I simply use a ASMV to give it AC will that work?

Why not. But what ASMV is
You should read some doc on LCD multiplexing. Microchip and National Semiconductor have some solid & didactical ones. I can never explain better than them and it would spare you a lot of questionning & time.
Cheers.