NEC LD8121 VFD Tube pinout

[sergey]

Hello everyone,

It happened so i've got NEC LD8121 tube salvaged from an old piece of gear and would like to try lighting it up.

There're some specs in ebay about this puppies which provides voltages, current and impulse characteristics , but i'm failing to find actual pinout of this tube.

One old russian book about indicators says IV-27 is an analogue for LD8121, but pinout wouldn't match fr sure (IV does have 11/15 pins, LD does have 13pins on each side).

Maybe it happened so someone does have pinout? Or maybe there's a technique to figure it out?

Thanks in advance!

P.S. Attached a photo of this guy just in case someone is interested how it looks like

[c4757p]

Either probe around, or just look, to figure out which pins go the filament. (The two which have a low resistance - around 50 to 1k ohms - between them.) Apply about 2V to these, then group all the other pins together and apply about 12V to them. You can play with the voltages to get a reasonable brightness (you'll probably end up with 5V on the filament and about 20V on the anodes - be careful with the filament, don't run it glowing hot). Then just start removing pins one at a time from the anode voltage and seeing what shuts off.

Make sure you check that there are only two pins for the filament - you wouldn't want to get one of the tubes with four filament pins and stick 12V up it.

[PA0PBZ]

then group all the other pins together and apply about 12V to them.

Aren't the segments multiplexed in some way? I can't see how you control close to 100 segments with only 24 (26-2) connections otherwise. And if they are multiplexed will you get anything by just connecting them all to 12 volt?
Not criticizing your advice, just wondering...

[sergey]

Either probe around, or just look, to figure out which pins go the filament. (The two which have a low resistance - around 50 to 1k ohms - between them.) Apply about 5V to these, then group all the other pins together and apply about 12V to them. You can play with the voltages to get a reasonable brightness (you'll probably end up with 5V on the filament and about 20V on the anodes - be careful with the filament, don't run it glowing hot). Then just start removing pins one at a time from the anode voltage and seeing what shuts off.

Make sure you check that there are only two pins for the filament - you wouldn't want to get one of the tubes with four filament pins and stick 12V up it.

Thanks for the quick reply

Just to clarify, do i need to use 1/16 duty cycle with 40us PW or for such a quick test 5/12V DC is not gonna to make any damage?

[sergey]

Aren't the segments multiplexed in some way? I can't see how you control close to 100 segments with only 24 (26-2) connections otherwise. And if they are multiplexed will you get anything by just connecting them all to 12 volt?
Not criticizing your advice, just wondering...

They are actually multiplexed. Not sure about this articular device, but for IV-27 you've basically got 7 pins for segments and 12 pins for digit selection.

But i guess it's just more fun for figuring out segments/digits and powering all up is just gonna to make all segments to highlight?

[c4757p]

Yep, they're multiplexed. You've got anode pins, which go to the segments, and grid pins, which go to the digits. Both of these are active high, so by grouping them together you're turning on all segments of all digits. Then when you remove a pin, either one segment of every digit will shut off, or one entire digit will shut off.

No, you don't need any oscillation. In proper use, you'll want to use AC (either from a transformer, if this is the 70s, or with a couple switching transistors if this is 2013), because a voltage gradient across the heater will cause one end to appear brighter than the other.

[sergey]

Thanks for the info, c4757p

Hope to give it a try over the weekend. And my plan to share the findings here as soon as i get them

[PA0PBZ]

Yep, they're multiplexed. You've got anode pins, which go to the segments, and grid pins, which go to the digits. Both of these are active high, so by grouping them together you're turning on all segments of all digits. Then when you remove a pin, either one segment of every digit will shut off, or one entire digit will shut off.

Ah, both active high, I see. It's a bit counterintuitive, my first guess would be that to activate something you need a potential across it, so one side more positive than the other side. Oh well, guess I have some googling to do...

[sergey]

Ah, both active high, I see. It's a bit counterintuitive, my first guess would be that to activate something you need a potential across it, so one side more positive than the other side. Oh well, guess I have some googling to do...

If i understand the physics correct, there is a potential between anode and cathode. Grid which are used for multiplexing, acts as AND between serment and placement to higlight a given segment of a given number.

So that's not actually so much counterintuitive

[c4757p]

Precisely. A positive potential is applied to the anode, and the electric field between anode and cathode causes electrons to flow out of the cathode via thermionic emission. If the grid is held at a potential equal to that of the cathode, then the field between them is zero, blocking the "view" of the cathode from the anode. The only field that will be "seen" by the anode is the one between anode and grid, and because the grid is not heated, no current flows. If you hold the grid at a potential near the anode, it becomes like an anode and current will flow to it, and because it is thin and sparse, it will not block most of the current flow to the real anode.

If the grid is left floating, like in this test, then it could be at a high positive potential, allowing current to flow, but it will quickly attract the negative electrons flowing past it and equalize to zero.

A VFD, as well as a CRT, is just a vacuum triode with some phosphor material painted onto the anodes. You could probably amplify a signal with one....

[c4757p]

Quick update before you get started: I just realized I recommended starting with 5V on the filament. No!!! There are a good few (especially older) VFDs that take about 2-2.5V - you'll pop it with 5. Start around 2V, 12V on the anodes, and if it doesn't work satisfactorily, increase the heater voltage gently and keep an eye on the filaments - go up until they just start to glow a tiny bit, then back off the voltage a hair so that they're just about to glow.

If they are indeed equivalent to IV-27, this website claims 5.15V at 160mA.

[sergey]

Hah good to know this before doing experiments

[Harvs]

If you haven't seen it, you may find something useful in the schematics I posted a little while back.
https://www.eevblog.com/forum/projects/'jelly-bean'-vfd-driver/

[sergey]

If you haven't seen it, you may find something useful in the schematics I posted a little while back.
https://www.eevblog.com/forum/projects/'jelly-bean'-vfd-driver/

Indeed didn't see it yet. And schematics seems really useful for the young VFD players like me. Thanks

[sergey]

Just quick update.

Unfortunately, was stuck with real job for quite a while and then my DC supply decided to go kapuuut. Fixing it now and hope to come with pinouts soon

Sorry if someone was interested in such a schematic..

[sergey]

Ok, finally got it! (was rather busy month at work, but that's another story).

So, used multimeter to find filament which was rather easy (one connection i saw, other one i just bruteforced it's a pair of legs on separate sides of the tube with ~45 Ohms between them.

Applied 3.3v across them, and tube drew around 26mA on filament. I didn't actually see gloving of the filament, so perhaps it could be increased to 5v as per IV-27 datasheet, but 3.3v volts worked fine for my purposes at this point.

On the anodes i've applied 12v and this lighted the tube up (it was drawing about 9mA on the anodes).

The only issue so far was that 13th segment didn't light up evenly -- it was darker comparing to the rest of segments. Hence the question: is it because of not enough power on the filament/grids or it's about physics of rather oldie tube?

Attaching a photo of the lighted tube (for those who're interested). And also attaching pinout of the puppy. Sorry for the crudity of the scheme, didn't have time to build it for scale or paint it! (and i do hope i didn't screw up when was drawing scheme on the computer).

[Halbelunge]

Hey Sergey,

you have news with the LD8121 or need more informations?
I have bought me one of these and have found some more informations for this tube.

with regards
Benjamin

[artag]

This is a bit of an old thread, but I came across it looking for info on the same tube (probably from the same vendor, still with stock !)

You mention a dim 13th digit. I believe this is because you're using DC to drive the filament (you don't say, but you don't say AC either). This means that one end of the filament is 3.3V more positive than the other, resulting in lower acceleration voltage at that end of the tube.

The ideal way to run the filament is with AC, using a centre-tapped transformer so both ends are varying positive and negative with respect to the most negative grid or anode voltage. It's possible to get away with DC if you can choose the right bias, but AC is better.