Inverting a dot matrix LCD display in hardware?

[TERRA Operative]

I'm currently repairing a Tektronix THS720 scope and have a replacement LCD with a small problem..

It turns out the LCD is inverted to what it should be for this application.. the screen is transparent and the pixels go dark, however this scope normally has an opaque screen and an inverted graphics are sent to make a correct image.
Using a normally transparent screen means I get an inverted image.

I know the easy solution would be to rotate one of the polarisers, but the polarisers are stuck to the LCD panel and finding replacement FSTN/STN polarisers is difficult..

So, is there any relatively simple way to invert the signal coming into the LCD in hardware somehow so the image displays correctly?

[kilobyte]

These types of 320x240 pixel displays have normally 4 data lines for the pixeldata. I think it doesn't have a build in display controller so it should be enough to simply invert the 4 datalines.

[Buriedcode]

These types of 320x240 pixel displays have normally 4 data lines for the pixeldata. I think it doesn't have a build in display controller so it should be enough to simply invert the 4 datalines.

I second this.  That looks like an LM32P07 which is indeed controllerless. 12-way FFC are pretty common, as are the FFC connectors, its 5v signaling, so 74HCT04 on D3-D0 would do. Or you could go with a quad XOR gate (74HCT86) which would allow you to have control over whether or not the data is inverted - either by a 5V signal, or a jumper setting.

And yes, finding polarising film specifically for FSTN is difficult - I still haven't found a srouce for low quantities.

[TERRA Operative]

Yep, can confirm that this is an LM32P10, and it does have the 4 data lines.

I just checked my stock of TTL chips and I have a tube of 74HC04 which might do the trick for testing, but I can get TSSOP14 Toshiba 74VHCT04 chips for cheap when I want to make the final circuit.
Looks like supply to the LCD is 16VDC, so a little 5V vreg and a couple SMD caps alongside would have it sorted I think.


Now to get some flat flex cables and adapters to rig up a test.

[NiHaoMike]

Most scopes use a light trace on a dark background, maybe it's really a feature!

[Buriedcode]

Yep, can confirm that this is an LM32P10, and it does have the 4 data lines.

I just checked my stock of TTL chips and I have a tube of 74HC04 which might do the trick for testing, but I can get TSSOP14 Toshiba 74VHCT04 chips for cheap when I want to make the final circuit.
Looks like supply to the LCD is 16VDC, so a little 5V vreg and a couple SMD caps alongside would have it sorted I think.


Now to get some flat flex cables and adapters to rig up a test.

The display will be powered by 5V for logic, and a VEE line which is most likely ~-20V,  so no need for a regulator - the negative VEE is probably only capable of supplying a few mA anyway.

[TERRA Operative]

Awwwww yiss!

Rigged up a 74HC04 on a breadboard with a few FFC adapters and it's working perfectly! 


Now, next issue.... Seems this LCD needs a +17 to 21.5VDC for contrast control, whereas the original  used a -17.5 to -14VDC

So, to get the contrast control working from the scope menu, I need to invert the voltage and tweak it a little... Any ideas? Maybe an opamp, but I only have the 5VDC rail to power it..


Looks like the main chip in the LCD is a BU6132F, if I could get a datasheet I might be able to figure out something a bit easier. Anyone know where I can find one?

[Buriedcode]

Positive voltage boost is easier than negative - especially considering that is only powering a quad opamp and a few drivers on the panel (<5mA but pretty constant).

Depends on cost, size of circuit, and how much control you want to have over the voltage, being with a trimpot? or digital control?  The low current requirement means a charge pump can be used, but these are rarely regulated.  the ancient MC34063 will work, but very inefficient, noisy and probably overkill.

There are dedicated switching controllers for generating LCD voltages (MAX749?), but agian, I think these are overkill, when any boost converter capable of outputing <25V @10mA, with a trimpot on the feedback will do well. TPS61040 for example.

For quick and dirty "parts bin" options, a 555 driving a charge pump/multiplier could work for such a low current application.  Not easily regulated/adjusted though.  I'm sure there are some cheap no-brand "modules" about that will get you that voltage that can be modded to be adjustable.  The danger with these is not knowing how noisy the output will be and what they do when they're turned on/off.  I've seen SDB628 based boost converters about, that will output up to 28V and seem pretty small and efficient.

Edit: whilst invasive, if the original generated the negative VEE with a switching supply - adding a charge pump on the switching node can generate a positive voltage from that: https://www.analog.com/en/technical-articles/regulator-topologies-for-batterypowered-systems.html figure 3

[TERRA Operative]

The little adapter board on the LCD has an LM2733 boost converter that boosts the 5VDC supply to the required +17 to 21.5VDC for the contrast control, with a small trimpot to allow for adjustment.

I wonder if there is some way to control a boost converter with a variable negative voltage? I haven't got much space to play with (A bit bigger than a thumbnail) so the smaller the better.

[Buriedcode]

The little adapter board on the LCD has an LM2733 boost converter that boosts the 5VDC supply to the required +17 to 21.5VDC for the contrast control, with a small trimpot to allow for adjustment.

I wonder if there is some way to control a boost converter with a variable negative voltage? I haven't got much space to play with (A bit bigger than a thumbnail) so the smaller the better.

Ahh sorry, I misinterpreted what you were asking - thought you were using a bench supply for the new VEE and wanted a more permenant solution.

So the original equipment produces a negative VEE for the old LCD, that it can control to change contrast, but the new adapter kit creates its own fixed positive VEE for the new display, and you wish to be able to control that voltage using the negative voltage produced on the original board?

You can certainly control the output of the boost converter by playing with the feedback pin - putting a resistor in parallel with the lower resistor in the feedback path to your VEE, but it gets very tricky working out what the resistance sohuld be, and what the range of control would be.  You are probably right in that you may have to use an opamp for gain/level shifting -  that is, you can use a resistive divider to convert the negative range to a positive one, and then have to scale that to control the range of the positive boost converter.  Possibly inverting it too, since raising the -VEE would raise the positive VEE, where-as the effect should be the opposite - very easy to get confused here, but I'm sure it can be done.

One thing about messing with the feedback path of a converter - it could affect stability, and since we dont' know the startup behavior of the negative VEE, wait for that to stabilise before enabling the boost converter, so its feedback path has a solid reference.

[ajb]

Given that you need a very small amount of current, you could set the boost converter to the max voltage and then use a linear regulator made from an op-amp and pass transistor to set the final output voltage.  This requires bodging the LCD module by cutting a track or lifting a pin somewhere, but it's a bit easier to design since you don't have to worry about the stability of the boost converter or figuring out the right resistor values to mix with the existing feedback network.  It's a simple inverting amplifier with (according to the back of this envelope here) gain of 5/3 and offset of 5.35V, so just need four resistors to set gain and offset and a cap or two for compensation.  (Plus maybe a zener for a more stable offset voltage.)   It's also inherently simpler to ensure good startup behavior, since the maximum possible voltage is set by the fixed feedback of the boost converter.

[TERRA Operative]

So the original equipment produces a negative VEE for the old LCD, that it can control to change contrast, but the new adapter kit creates its own fixed positive VEE for the new display, and you wish to be able to control that voltage using the negative voltage produced on the original board?

Correct. The voltages I have available are a stable +5VDC for powering the LCD (which also powers the boost converter on the piggyback PCB on the LCD panel that has a trimpot for manually adjusting the +ve contrast control voltage) and also the user adjustable variable -ve voltage from the scope mainboard that was originally connected directly to the old LCD contrast control pin.

On the new LCD, the contrast control line from the scope main board that supplies the user variable -ve contrast control voltage is ignored and instead connected directly to that boost converter.
So if I could rig something up so the variable -ve voltage from the scope can be what controls the +ve contrast control voltage instead of using the trimpot, that would be workable.

You can certainly control the output of the boost converter by playing with the feedback pin - putting a resistor in parallel with the lower resistor in the feedback path to your VEE, but it gets very tricky working out what the resistance sohuld be, and what the range of control would be.  You are probably right in that you may have to use an opamp for gain/level shifting -  that is, you can use a resistive divider to convert the negative range to a positive one, and then have to scale that to control the range of the positive boost converter.  Possibly inverting it too, since raising the -VEE would raise the positive VEE, where-as the effect should be the opposite - very easy to get confused here, but I'm sure it can be done.

One thing about messing with the feedback path of a converter - it could affect stability, and since we dont' know the startup behavior of the negative VEE, wait for that to stabilise before enabling the boost converter, so its feedback path has a solid reference.

Yeah, the LM2733 just has the usual resistor divider to set output voltage, but if I were to use an opamp I would need supply rails and I don't have much space to fit it in... Is there a tiny 'one-chip solution' for providing say about +/-24VDC-ish rails for an opamp?

Given that you need a very small amount of current, you could set the boost converter to the max voltage and then use a linear regulator made from an op-amp and pass transistor to set the final output voltage.  This requires bodging the LCD module by cutting a track or lifting a pin somewhere, but it's a bit easier to design since you don't have to worry about the stability of the boost converter or figuring out the right resistor values to mix with the existing feedback network.  It's a simple inverting amplifier with (according to the back of this envelope here) gain of 5/3 and offset of 5.35V, so just need four resistors to set gain and offset and a cap or two for compensation.  (Plus maybe a zener for a more stable offset voltage.)   It's also inherently simpler to ensure good startup behavior, since the maximum possible voltage is set by the fixed feedback of the boost converter.

I don't want to bodge the LCD PCB if I can avoid it, beyond attaching a piggyback PCB. I might do more of these in the future so simpler is better, especially as data sheets for the chips used ont he LCD are hard to come by and can change between batches..

[ajb]

but if I were to use an opamp I would need supply rails and I don't have much space to fit it in... Is there a tiny 'one-chip solution' for providing say about +/-24VDC-ish rails for an opamp?

You can do this with only the supplies you have available, even just within the 5V rail.  If you're steering the feedback node of a boost converter, that node will be ~1-2V, and steering it with 0-5V is just a matter of selecting the right mixing resistor.  You can even do the pass transistor method with a 0-5V op amp with an additional transistor. Tapping into the old or new VEE could be helpful for providing more convenient offset voltages, but there are ways around that.

I don't want to bodge the LCD PCB if I can avoid it, beyond attaching a piggyback PCB.

Fair enough. You can do a lot in a "postage stamp" size add-on board. A whole new boost converter and a couple of op amps to implement whatever transfer function you want would be no problem. So it depends on what looks easiest to bodge 

[TERRA Operative]

You can do this with only the supplies you have available, even just within the 5V rail.  If you're steering the feedback node of a boost converter, that node will be ~1-2V, and steering it with 0-5V is just a matter of selecting the right mixing resistor.  You can even do the pass transistor method with a 0-5V op amp with an additional transistor. Tapping into the old or new VEE could be helpful for providing more convenient offset voltages, but there are ways around that.

Of course! It completely passed my mind that the feedback voltage is only a few volts... Even if I do need a negative voltage for the opamp, within the 5V supply rail I can just use something like an ICL7662 for an easy solution.

Maybe then, I can feed the variable -ve voltage from the scope through a voltage divider into the opamp (to keep it within the bounds of the 5V supply rails), invert the voltage with the opamp, then feed that into the boost converter to bump it back to the required +17 to 21.5VDC

Sound reasonable?

[Buriedcode]

Just an idea - again, with the caveat that you might want to control the "piggy back" boards boost converter - pretty sure all boost converters have an enable line, and you should power up logic before VEE on most STN LCD's anyway.

Most converters use a ~1.25V reference on the FB pin. I assume the trimpot used to adjust this is one of the resistors in the feedback path.   As you have access to the "old" negative VEE, then a resistor from this to the FB pin instead of a resistor to GND should give you want you need.  If the values of the resistors in the voltage divider on the feedback are the same then, a -17.5V from the negative VEEE should prodice ~+20V out of the boost converter.  Similarly a -14 input should produce a ~+16.5V on the boosts output. 

So the adjustment range on the original board was:

 

-17.5 to -14VDC

Making the divider the /2 by making both resistors say, 20k, then the output is  (-1*VEE)+(2*Vref) = -VEE+2.5V.   Giving a range of 16.5 to 20V.  You could adjust the values of the resistive divider to shift this slightly higher to give your +17 to +21V.

Again though, the main caveat, which I shall now put in bold:
Normally the feedback pin has a resistor to GND, which is generally a pretty stable and solid reference.  Referencing this to an adjustable source means the output of the boost converter is dependant on this, and so it should be reliable and stable before the boost converter is started.

As most negative supplies when they start up - just like boost converters - start with an output at 0V and then go negative from there. The effect of this is that if you power up the boost before the negative supply, the boost converter will output a voltage of ~2.5V (feedback pin = 1.25V, both R1 and R2 = 20k, so output is 2*1.25V = 2.5V.  And wil increase as VEE goes negative.  So it shouldn't blow anything up.

I would still add some kind of delay before powering up the boost converter.

So, yeah, you might be able to do this with a single resistor.