Trying to figure out if a PLASMA/Neon display/driver is broken (74 series logic)

[akasaka]

UPDATE:

I am happy to announce the panel has been started up successfully and has been put to some good use! Check this video out to see my creation:



So I recently acquired a holy huge VFD matrix display. This thing weighs quite a few kilos and is much larger than my hand.



Naturally given how I love those kinds of displays, I wanted to get it running. It came with a digital board, so I thought, can't be hard, huh? It has a TEST switch, and a 8251 UART, so it'd be a matter of dumping the ROM and finding out the protocol, then stuffing an arduino on the other end and showing weather or something.



So I get my IC puller, take the ROM out, drop it into my MSX, start the dumper tool... and I can only read 00s. Well, if I erased it somehow, it would be FFs, so something is wrong in another way, right? The ROM also gets warm, so it's not like it's not connected.

Well, fat chance — I measure 27 Ohms on the power connector of the logic board, and it just causes my PSU to go into protection. It's also lacquer coated to add insult to injury — can't even pull the mask ROMs to dump the fonts, not to mention repairing it, though not like there is much point in that given the ROM is gone.

Thus, I turn my attention to the VFD module itself. It comes with this kind of a sweep/framebuffer/whatevery board:



The L1-2-3-4 plug is a standard Molex trapezoid you'd find on an HDD, and the 5V/Ground positions matched that (checking against the ICs Vcc/Vss pins). So I assumed the other side would be 12V and plugged in a standard PC power supply.
It seems to have started up fine on the 5V side, but the consumption on the 12V side was suspiciously low, at around 5mA.

Probing around with a scope shows that at least something in here is alive — the 74LS257 multiplexing the input bus gets switched with a few kHz every now and then, the 2114 SRAM is getting some motion on the data lines (which looks weird, but whatever for now) and the #WE pin. However, the display itself is dead silent. The diodes in the top right corner which, I assume, should drive the cathodes on the display, always have 5V on their negative side (and 0V on the positive side).

I tried shorting the 2114 data outputs to ground or Vcc, nothing. Then I tried shorting the negative side of those diodes to ground, still nothing.

Out of ideas, I tried slowly brushing the ground wire on the pins of the VFD itself, to see if it lights up directly, but still nothing.

Worthy of note I cannot see any heater glow in the display — but I can't also see any heater pins to check against. There is a pair of pins on both bottom and top side of it, though, that goes directly to power via a 1MOhm and 2MOhm resistor for each pin. But testing continuity across those pins doesn't give an expected 0 reading, but rather some more megaohms range.

The VFD itself also has a few boards: two side boards with weird modules I have no idea about, and two boards full of tiny tiny transistors.





To summarize, I am inclined to think the digital side of things works fine, but I don't know the protocol to communicate with it. Before I go full force into analyzing this bunch of logic gates, I want to see if the VFD tube itself is working.

My current hypothesis is:

1. It works, I've just been plugging the wrong voltages into it (e.g. should have given the pins 12V instead of Ground)
2. It works, but disabled by the drive board on a more fundamental level than just not outputting to the cathodes. The transistor contraption in the bottom right corner does indeed control a 12V supply to a negative side of a diode, which has the positive side going somewhere to the VFD. But normally it is open, thus the negative side of the diode is pulled down to Gnd, so current from the VFD must flow if it wanted to?
3. 12V is too big for this thing and I've already destroyed it. (Because when you assume, you make an ass of u and me) — I don't think that's realistic? I don't see any boost converters on the board and a VFD of that size would most certainly not run on 5 volts.
4. The 12V rail should be negative in reference to ground. Well in such case the axial capacitor in the top right of the drive board will end up being in reverse polarity and eventually explode :-) That capacitor is also rated 1uF 250V so maybe...
5. 12V is too low for this thing and I need to go higher. Well, I tried 19V, nothing changed, still the same 5mA draw and that's it. For something even higher I'd have to invest in a bench PSU which I'd rather not do unless I know this is the way. Also I presume they wanted, given the 8080 on the original logic board, to test the VFD with a PC during development, hence the Molex power connector? So I would expect it to be 12V, but also a matching connector does not always mean matching voltages...

So the question is, am I doing something wrong and what could be missing? I feel like it's something pretty simple. Would be lovely to get this thing running and put it to use in a project I'm doing right now.

P.S. If anyone knows some easy helpful tools for tracing schematics of such things filled with 74 series logic, hints are much appreciated.

P.P.S. Also hints on how to make those imgur pictures smaller are much appreciated — I ran out of attachment space, so had to put the photos there.

Also, keywords off the PCBs for those googling in case someone has a similar board in the future:
* Morio Denki 6M06056
* Mitsubishi MA7446-01
* MD-24T-ADT (2) 8201
* MD 16101DS-CONT82 06

[pcprogrammer]

Wow that is a very cool display. 

I did not read your whole post, but I do have some remarks.

A VFD display usually works on higher voltages like 35V to make the pads light up. Usually there is also a filament for heating up the tube which runs on something like 6.3V.

Forget the logic boards for now. Just focus on the display. It is driven as a matrix, with the long transistor boards for the x direction and the two side boards for the y direction.

Try to map out a schematic on how the transistors are connected, because it looks like they are time multiplexed in some way.

With the VFD working and the matrix setup worked out you can easily use some new micro to generate the needed signals to directly drive it. Although the number of pins needed is a bit high. If correct I counted 56 connections between the driver board and the display boards. Some will be ground and maybe some also power. All in all even if it is 56 pins there are MCU's with more IO pins available. An ATmega2560 might be able to pull it off.

Character set are easy to be found. Probably 5x7 matrix with single dot spacers between the characters.

[pcprogrammer]

This time round I fully read your post, and can't see heating filament wires either, but on the right display board there are the KA1 and AK2 connections to the display that are NC on the other side of the display, but like you wrote these are connected to 2MOhms resistors to, according to you, power. The question is which power line.

There is a 74LS14 on that board which needs 5V, so at least there is 5V on the connector to the board. And ground, which leaves 8 wires. Is there another power line?

For 24 scan lines 5 bits is enough to walk through them, so why the larger number of wires?

The module is a bit of a problem, nothing to be found on it. But the diodes indicate that there are 24 lines on each side of the display. So maybe 48 pixels in the Y direction. Then there are 202 transistors on the two other boards, which is a bit of an odd number for the X pixels, and does not match with the number of bits in the 2114 memory chip. It also raises the question how they are driven in respect to the Y lines.

Also counting the Y pixels on the display shows only 16 squares. So 3 pixels per square? The X direction is harder to count but looks to tell the same story.

The transistors seem to have L5 as marking and could be 2SC1623 NPN transistors with a Vce rating of 50V. This gives an indication of what the maximum on the display can be. The way they are configured shows a connection to ground or a negative voltage. Anyway this looks to point to a positive voltage on the Y lines.

A simple test could be done with a power supply that can provide up to lets say 45 volts, and connect one Y line to the plus and one X line to the ground, and slowly turn up the voltage to see if a pixel lights up. But only when it is clear that the voltage won't damage the special module. Since there is no ground connection between the boards on the display it should be safe though.

I love this kind of quests. Keep me in the loop. 

Success,

Peter

[akasaka]

Hi Peter, thanks for having a look at this.

Good catch with AK1-AK2, I didn't notice them! However they are both connected to "+12V" via 2MOhm resistors, not different poles of power. There is also no continuity between them inside.

The ones I mentioned being connected to power are the two rightmost pins on each of the transistor boards. Upon further checking both go via 1MOhm to ground, so maybe that's something else.

To answer your other question, there is +5V, "+12V" (let's call it like that for now), and Gnd, coming to the boards.

The transistors look more like "L-S" to me than "L-5"? I took a closeup photo:



They are connected as follows (using SOT23 numbering of pins):
1. (Base?) Common across the same row of adjacent groups in a "block of groups"
2. (Emitter?) Common across a group of 12, but not between groups
3. (Collector?) Not common, goes to a resistor array, and then to the VFD itself. The resistors of the array go to the drive board, where they end up on a 33k resistor in series with a [ceramic capacitor + zener diode in parallel] and then to ground.

There is also one transistor on the right board, which is a C1473 — which is apparently 2SC1473, an NPN with Vce of up to 200V.

I have also taken a closeup of the squares, if it's 2x2 pixels in a square that is some really really high resolution VFD, I've never seen such before.



Inre: another microcontroller, I'd rather keep using this 74 series board as it seems to have all that's needed for the multiplex drive and all that complex stuff. I know if I have to make my own PCB I'll never finish the project :-)

And the character set would be a nearly-full set of japanese kanji symbols, so that's not really easy to come by either.

I'm about to buy some cheap-o bench supply for starters on Amazon, there is a 30V 10A version and a 60V 5A — I presume the latter would be more helpful? It also has a USB output, so I can probably strap a USB cable to power the 5V side of the circuit later down the line.



Edit: I've found a list of LS marked components in SOT23 casing that are transistors:

* Sanyo 2SA1839: PNP 15V, 100mA, 200mW, B=200..600, 600MHz — can't see any dates online, but some pretty old looking datasheet exists: https://pdf1.alldatasheet.com/datasheet-pdf/view/38653/SANYO/2SA1839.html
* Siemens BF770A: NPN SATV-IF, 20V, 50mA, 6GHz — dated 1999, so I think that's too new?

So it may be too dangerous to apply a high voltage after all

[inse]

What’s also puzzling is the CN1 connector on the controller board with that fat resistor and the optocoupler (sensing the voltage?)
Regarding the overload issue: I doubt a resistance measurement on the 5V line makes much sense, the 27Ohms would only make 180mA - causing your supply to go into protection??
If the current draw was beyond reasonable value, I would check if anything gets hot.
An adjustable bench supply with current meter is indispensable here.
The ‘high voltage‘ generator has always been close to the display in the modules I have seen so far.
No transformer is to be seen on your display.
This strange Mitsubishi thing MA7446, what could it be for?
Any chance that it’s a converter module - how high is it?

[pcprogrammer]

Yes you might be correct that the L5 is LS instead. I was wondering what the additional pins close to the transistors where, but a resistor array sounds correct. Either a pullup or pulldown on each transistor.

The close up of the display is very helpful in telling the resolution. Four pixels with a common rail per square. How many squares are there in the X direction?

The test with applying a voltage between two pins could still be useful, but only after drawing up a schematic, and start with only 1V and slowly increase it.

I think I get the setup with the three connectors to the transistors. The 12 transistors in each array have their base connected to the same one in the other arrays and the connector on the right of the bottom board. Only thing is that seems to be one wire short. Then the 17 arrays have either the emitter or the collector tied together per array and to the other two connectors. So 17 wires + 12 wires is 29 wires leaving 7 for power.

But why the 24 connections on each of the far sides of the display?

Still intriguing. 

See if you can draw up the  left and right driver board treating the MA7446 as a black box.

Edit: The 5V is for sure due to the TTL IC's, but assuming that the other one is 12V due to the connector being the same as what is used on older HDD's and even older 8" and 5.25" floppy drives might be a stretch.

[pcprogrammer]

What’s also puzzling is the CN1 connector on the controller board with that fat resistor and the optocoupler (sensing the voltage?)

Might be a 50 or 60Hz detection from mains voltage.

[akasaka]

What’s also puzzling is the CN1 connector on the controller board with that fat resistor and the optocoupler (sensing the voltage?)
Regarding the overload issue: I doubt a resistance measurement on the 5V line makes much sense, the 27Ohms would only make 180mA - causing your supply to go into protection??

The resistor and optocoupler are part of the 8251 UART's RX-TX circuit. Think like MIDI — you have a resistor to set the current in the receiver's optocoupler for Tx, and expect them to do the same in your optocoupler for Rx.

Basically the 8080 CPU communicates with the host CPU using this UART, and also the display panel using the 16-pin connector. The remaining 16-pin connector is an input, so I guess it was a bus to connect multiple displays, or an input rig.

The 27 Ohms might have been 2.7, I don't remember exactly now. It does make the beeper on my DMM beep, that's for sure. I'll consider investing into a bench supply once financially things get easier (i.e. payday, which is about 2 weeks or so :-))

Either way, discard the controller board completely for now. I tried melting and removing the conformal coating to no avail, but it's so hard it can only be scraped off, so probably I'll just toss the board once I'm done with the project, or take it apart. Given the ROM with the software is gone, there is zero use for it anyway (unless someone wants to write 8080 assembly in this day and age? :-))

This strange Mitsubishi thing MA7446, what could it be for?
Any chance that it’s a converter module - how high is it?

It's a pretty thick module, I'd say 8-9mm tall. The pins sticking out look like a ceramic board you'd often find in modules in old CRT TVs or VCRs. But I'd expect a boost converter to also have an inductor and capacitors? Unless they stuck it all inside.

The close up of the display is very helpful in telling the resolution. Four pixels with a common rail per square. How many squares are there in the X direction?

By using a ruler I've approximated 3 squares every 1cm, which gives a total of 97 squares — 194 pixels horizontal. Combined with the previous 16 squares — 32 pixels vertical, that makes it 194*32 = 6208 bits, or 776 bytes, which is more than the 512 bytes in the 2114 SRAM. So I guess they are grouped by 4 (97*16 = 1552 bits, 194 bytes)? That would make sense.

See if you can draw up the  left and right driver board treating the MA7446 as a black box.

The issue is I'd also have to treat the VFD as a black box — I don't know where it's cathodes, anodes, or the heater, is...

The 5V is for sure due to the TTL IC's, but assuming that the other one is 12V due to the connector being the same as what is used on older HDD's and even older 8" and 5.25" floppy drives might be a stretch.

I mean, given the new LS marking instead of L5, that leaves us with the more likely 2SA1839, which is rated up to 15V. So this shouldn't really go too high.


For the time being I've messed a bit more with the drive board full of logic. I made a small jig to quickly mess with the 8 input bits and 5 other inputs, as well as LEDs for the 2 outputs it has. That's when I noticed the second nibble of the input is never read, despite there being a square wave on the 74LS257's select pin. The said square wave is actually never reaching 5V, looking kinda "squashed" with the top being around 2.5V or so.

Additionally I've noticed the SN74107 in position 107-1 is going pretty toasty. There's no switching on it's pins whatsoever, so why? Like, this is 74 series, which is pretty warm overall, but that specific chip gets so hot it's hard to keep my finger on it! I wonder if the same issue that killed the ROM and the drive board also happened to mess up something in here?


Looking up the manufacturer's name, most of their projects seem to be related to the Shinkansen, so this might be quite a historical piece we have here! Unfortunately, this time it doesn't mean there's already some web1.0-looking blog of a train guy who's already reversed all the schematics and protocols — we're on our own here :/

[pcprogrammer]

I took a dive in my hardware archive to take close up pictures of my VFD's and they look very different so I'm thinking yours is something else, but what that is the question.

On mine there is a pad with a grid above it and above that the heating filament. Each character has its own grid and a 5x7 dot matrix underneath it with a separate underline of 5x1 dots. Another display has a 5x20 dot matrix.

Could yours just be some LED's?

[akasaka]

I don't think there would be any reason to stuff LEDs in a vacuum sealed glass tube... Especially in the early 80s

It's probably just got a grid painted on the top glass to make it more like "pixels" when viewed from afar

[PA0PBZ]

I love a good mystery   Looking at the boards with all the transistors I noticed a few things:

 - The upper board has 94 transistors, the bottom board 96
 - There is something going on at the bottom right on the upper board, what components are hidden there? Are these the same in the upper right of the lower board?
 - The bottom right of the lower board seems to have been reworked

Also subscribed by posting this!

[pcprogrammer]

I love a good mystery   Looking at the boards with all the transistors I noticed a few things:

 - The upper board has 94 transistors, the bottom board 96
 - There is something going on at the bottom right on the upper board, what components are hidden there? Are these the same in the upper right of the lower board?
 - The bottom right of the lower board seems to have been reworked

Also subscribed by posting this!

You miss counted. The top board has 106 transistors. 8 x 12 and 1 x 10. The bottom one only has the 8 x 12. The two boards are interconnected in the middle.

I don't think there would be any reason to stuff LEDs in a vacuum sealed glass tube... Especially in the early 80s

It's probably just got a grid painted on the top glass to make it more like "pixels" when viewed from afar

True and that is why I have been searching a bit further and think it might be a plasma display. But need to look a bit deeper to see if it really can be that based on the layout seen in your close up picture.

[PA0PBZ]

You miss counted. The top board has 106 transistors. 8 x 12 and 1 x 10. The bottom one only has the 8 x 12. The two boards are interconnected in the middle.

 

[pcprogrammer]

Not sure about the plasma either, but did find this DMD type and someone who worked on a problematic one. On the picture of the front of the illuminated panel it looks similar.

It does work with relatively high voltages, but on yours there is no connection for multiple voltages as described in the article. Maybe the modules can generate these voltages but then still what would the supply voltage be.

There is not a lot to be found online about this thing.

[factory]

No grid or filaments can be seen, there is a random 1uF 250V electrolytic capacitor on one of the logic boards. Could this not be some form of ancient gas plasma display?

David

[pcprogrammer]

As I already dug the display out of the hardware stash, I thought "what the heck lets see if it works".

And it did. Connected to a Siemens SAB80535 board I developed long long ago running Forth in an EPROM without a sticker being buried in a box for at least 20 years. Had to resolder a voltage regulator and figure out the connections between the display, controller board and power supply to make sure all the grounds where in the right place and the correct voltage supplied to both the display and the controller.

The display showed a startup sequence and then "XACT DONGLE ON" which reminded me that I used it to emulate a dongle for running old Xilinx XACT software. 

[akasaka]

Hmm, if this is a plasma that would explain the lack of a heater. This could have a connection to how most LED displays in Japanese trains are orange, being an homage to their plasma predecessors? (even though sadly they're being replaced by LCDs nowadays) But in that case transistors capped at 15V make no sense, neither the usage of PC standard molex connector. Then again, when did the PC caome around? In 1984+, after the Macintosh? Given the ROM was labeled 1983, they might have not yet known what the connector would mean for us here in the future indeed.

But also given this was likely in an automotive setting, I don't expect the other power rail to be less than 12 and more than 24 volts.

I'm too drunk today and I have to run an event tomorrow, so maybe next week I'll get some schematic drawing done.

That being said I have a board off a 1970s cash register which has a whoppin' Panaplex display, and a bunch of similar looking Mitsubishi modules... which I couldn't get working either... let me dig it out and compare next week as well.

[akasaka]

OK, as promised, here is the schematic of left and right boards.

As usual, "+12V" is just a name for whatever the other power rail is.

Hopefully this is readable enough. The transistor on the right board is a bit confusing. But I've also omitted all the ceramic capacitors and all that. Maybe I missed something after the capacitor C3...

[inse]

The plasma idea seems to make sense, Mitsubishi Electric is focusing on high power/high voltage drivers - so MA7446 might be some sort of high voltage switching hybrid?
Are you sure about the transistor circuit in your schematic?
It receives a signal from the drive board, double inverts it and sends it back??

[akasaka]

Yes, I have a correction for that part. Will upload the new photo once I'm back home.
The collector also goes to each display line, via a ceramic cap, to the point between the diode and resistor.

Same thing happens on the left board with pin 1, but without the transistor — it's just pin 1 to all the ceramic caps to all the points between the diode and resistor from each one.

Could it be some quick blanking thing, in case we're talking plasma, to prevent stuck pixels?

[pcprogrammer]

The four lines coming from the drive board into the inverters and then to the MA7446 points to 16 pixels controlled per side, but why then only 8 outputs to drive the display. And why do they connect via diodes to three lines of the display?

It seems to have some very complex way of driving the display, but the direction of the diodes at least show where the positive voltage has to be.

The feedback from the modules to the drive board is also interesting. What is it that this signals back that you would not be able to generate otherwise?

This might be the reason why the drive board does not work properly without it. At least I assume that you did some of your test on it without the display connected.

Would be great to see if you actually succeed in getting it to work.

Edit: It is hard to find data on this type of gas discharge displays, but I found this one on ebay which seems to have an anode per character and several flat cathodes for the dots. There does not appear to be a heating filament. Works of of 5V but I suspect the black thing underneath the PCB to be a DC/DC converter to generate the high voltages.

Also found a PDF about how they work. It shows the possible usage of subsidiary anodes, which could be the AK, KA and the other connections via the 2M Ohm resistors to the positive rail.

The four pixels with the raised small bar in the middle could be cathode / anode constructs. But it is all guessing. 

[akasaka]

I have ordered a bench power supply that goes up to 60V from Amazon, as well as a step up module that goes from 45 to 400V supposedly. Let's see if it runs off double digit voltages, and if it needs triple digits — I'll measure the current and, if applicable, go buy a smaller module from AliExpress to build into the actual enclosure. They seem to only go up to 15-20mA, which I doubt is sufficient for such a large panel. Not to mention the delivery times from Ali...

[pcprogrammer]

Not to mention the delivery times from Ali...

Yes these are always good for a big delay in a project. 

I'm also waiting for components to arrive so I can continue my own remote control project.

[pcprogrammer]

I have ordered a bench power supply that goes up to 60V from Amazon, as well as a step up module that goes from 45 to 400V supposedly. Let's see if it runs off double digit voltages, and if it needs triple digits

Have you followed the for now +12V from the drive board to all the components connected to it? Knowing this can provide insight in what the maximum voltage could be and if it even is a positive voltage in respect to ground.

[akasaka]

Yes these are always good for a big delay in a project. 

And as someone who seems to have ADHD, I know that I'm obsessed with getting this to run only while the parts for this are splurged all over my apartment's floor. The moment I clean it up into a box or something — like that Nakamichi I was posting on this forum a year or so ago — I will forget about it's existence for an eternity until I need to free up space for something else and be like "Oh... Well that was one stylish waste of money"

Have you followed the for now +12V from the drive board to all the components connected to it? Knowing this can provide insight in what the maximum voltage could be and if it even is a positive voltage in respect to ground.

I did to some extent. While I don't have any idea on how to identify glass diodes, there is one huge polar EL capacitor across that line, which explains the polarity quite clearly (I don't think they designed it to explode :-), and it's also rated at 250V. The ceramics in that area also have no underline under the value code, which means they are rated at 500V (all the smaller ones have an underline). There is also a 2SC1473 on the drive board in there, the purpose of which I don't yet fully understand, but it's also rated at 250V. So I think if not 40-60V then 100ish volts might be that probably. The logic ICs that input into the display's side boards I expect to be isolated by the Mitsubishi module, as we've discovered in the schematic above.