Vacuum Tube Computer P.05 - Vacuum Fluorescent Displays

Подписаться 88 тыс.

Просмотров 14 тыс.

50% 1

UPDATE ABOUT PLATE BRIGHTNESS WITH DC FILAMENT:
My description of why the brightness shifts across the length of the VFD is incorrect, though the brightness does indeed shift. I believe it actually shifts in the opposite direction, with the plates above the negative end being brighter and the plates above the positive end being dimmer.
However, this is not due to a temperature shift in the filament. Instead, at the ground side of the filament, the potential difference between the plate and the filament is at its greatest. However, at the positive side of the filament, the potential of the filament in relation to the plate is actually elevated above ground due to the resistance of the filament. This means that there are fewer electrons present that can be boiled off and results in dimmer plates.
===================================================
In this episode, we take a deep dive into Vacuum Fluorescent Displays (VFD) and try to figure out exactly how they work and what you can do with them. You can even do some stuff they weren’t designed for, like audio amplification!
Check out Dalibor Farny’s channel here: / daliborfarny
Check out TannerTech’s channel here: / @tannertech
VFD Picture 1: / 885656792440729601
VFD Picture 2: en.wikipedia.org/wiki/Vacuum_...
Also, we’re now on Discord, Reddit and Twitter, so join us!
Discord: / discord
Reddit: / usagielectric
Twitter: / usagielectric
Thanks for watching!

Наука

Опубликовано:

5 авг 2024

Ссылка:

Скачать:

Готовим ссылку...

Добавить в:

Мой плейлист

Посмотреть позже

Комментарии : 57

@simoncopar2512 2 года назад

Potential gradient along the filament isn't what determines the heating - the current is constant along the entire wire (ignoring the evaporated electrons). What *could* impact the intensity is, that the potential difference to the anode is slightly different (24V vs. 23V). However, the main reason for the dimming at the edges is probably nonuniform electric field between the cathode and the anode -- edge effects, especially if the rest of the conductors (connectors) are nearby the end of the element and attract the electrons instead of the anode.

@UsagiElectric 2 года назад

I actually did a ton of reading about brightness and gradient along the filament after I put this video live (a little late...). From what I read from multiple VFD manufacturers, is that I believe it actually shifts in the opposite direction as to what I state in the video, with the plates above the negative end being brighter and the plates above the positive end being dimmer. This isn't due to a temperature shift in the filament. Instead, at the ground side of the filament, the potential difference between the plate and the filament is at its greatest, and at the positive side of the filament, the potential of the filament in relation to the plate is actually elevated above ground due to the resistance of the filament. This means that there are fewer electrons present that can be boiled off and results in dimmer plates.

@schmitzvonschmitzen2870 3 года назад

Another great video about vaccum tubes, thank you! Your problem with oddly spread illuminessensce is probably not a problem of the heater itself, rather than using DC to power it. Applying 1V DC to the heater means putting one side of the heater on a more positive potential than the other. This is the reason why AC is mostly used for powering VFD heaters. By using AC this phenomenen will alter so faster you couldn´t see it.

@UsagiElectric 3 года назад

Thank you! Yup, I actually totally got that part of the video wrong. I had an awesome conversation with a viewer over on my Discord channel where we worked through trying to figure out exactly what was causing this. Since (ideally) the power supply for the filament is separate from the primary supply, the filament supply doesn't affect the brightness at all actually. However, the filament itself is a resistor, so when the electrons from the primary supply boil off, there's a stronger negative potential at one end compared to the other. The electrons from the primary supply have to flow through the resistance of the filament to the far end, which means not all of the electrons make it there, reducing the number of electrons that actually boil off and strike the plate. To combat this with DC filaments, the filament itself was often installed at an angle, with the positive side placed physically closer to the plates. VFDs are weird, but amazing pieces of technology! Here's a few good links (though the explanations on DC Filaments is a bit lacking): www.noritake-elec.com/technology/general-technical-information/vfd-operation www.vwlowen.co.uk/arduino/vfd/VFD-Application-Note.pdf

@edgeeffect 2 года назад

I came to this channel post-haircut. These older videos still come as a bit of a shock. .... and now I know why my VFDs like AC. :)

@jonka1 2 года назад

The dim lower anode is likely to be a lack of emissive coating on the ends of the filament. The slight alteration of brightness end to end is due to the bias effects of the DC potential with respect to the anodes.

@jorgeferreira6727 2 года назад

The effect of reduced brightness seems to affect both ends of the VFD, being stronger at the negative side. My explanation for it has 2 components. 1 - The electron cloud formed around the filament is not homogeneous. Its, naturally, less dense at the ends of the filament than in the middle section. 2 - The electron cloud is slightly shifted to the positive terminal due to the small electric field created around the filament by the difference in potencial at the ends, even at only 1 volt it exists. Given the above, both dots at the ends show lower brightness than the ones in the middle, with the one at the negative end being even less bright than the one at the positive end.

@BrainDamageBBQ 11 месяцев назад

And the TI Disk Controller AND the TI Thermal Printer? Wow... Nice.

@weirdboyjim 3 года назад

Enjoying this series. Optimizing the circuit from binary to 7-segment is an interesting challenge.

@UsagiElectric 3 года назад

Thank you so much! Binary to 7-segment is crazy difficult, especially if you're going to decimal instead of hexadecimal. Having EEPROMs makes it a lot more manageable, but there is something very cool about doing it with the fundamental building blocks. Also, I'm up to episode 38 in your 8-bit series, I'm loving it so far! I watch one to two episodes everyday on the couch with the bunny.

@weirdboyjim 3 года назад

@@UsagiElectric Oh yeah, the Eeprom trick is for too high level for the likes of us! ;-) I actually made a 7 segment decoder in my diode matrix sub series. I wouldn't try and do a decimal display in raw circuits, I wrote a divide for my processor to handle decimal display. If you really want to do that, look into the algorithm "double dabble" for converting binary to bcd, you can serialize that to do with fairly simple circuits.

@UsagiElectric 3 года назад

@@weirdboyjim I actually recently watched the first part of the diode matrix sub-series, and using the pin headers on the perf board is a brilliant idea! Going to decimal is definitely not something that I think is worth the amount of additional circuitry required, but a demultiplexer and diode matrix would work really well for a hexadecimal display, which is something I'll probably look into in the future if I get to the point where the binary is starting to melt my brain. (And your sub-series is going to be an excellent resource!) That double dabble algorithm is brilliant, but oh man it requires a lot of steps! Still, if I ever build something that has an 8-bit output register, that could be a quite interesting solution. The circuitry for that operation could almost be built entirely separately, so it does the shifts and adds simultaneously while the rest of the CPU does work on something else. Of course, doing all that with tubes makes it exponentially more difficult than it should be, haha.

@michaelturner4457 3 года назад

@@UsagiElectric The Apollo Guidance Computer did binary to 7-segment with just using NOR gates. Although I've not seen a schematic.

@UsagiElectric 3 года назад

@@michaelturner4457 Interesting! I did a bit of looking through some of the logic diagrams, but there's so much it's hard to know what you're looking at (especially without an intimate familiarity to the AGC). But, it seems like they often would use SR flip flops to store data and then decode that using Q and inverse Q outputs. Which makes sense. If you're using just NOR gates, that's by far the most economical way to do it! Lots of great diagrams here: klabs.org/history/ech/agc_schematics/

@AgentOrange96 3 года назад

4:42 So a VFD is just a diode that emits light? A "Light Emitting Diode" if you will? That's really funny but really cool. Basically, like there are tube diodes and semiconductor diodes, there are tube LEDs and semiconductor LEDs!

@UsagiElectric 3 года назад

Well, you're not wrong! In fairness, most VFDs have a grid, making them Light Emitting Triodes, but in the case of the IV-25 VFD bar dot display I use here, it's just 7 little diodes with a common cathode. So in that case, it actually is indeed a Light Emitting Diode, haha.

@AgentOrange96 3 года назад

@@UsagiElectric Fair point about the triodes. That made me wonder if Light Emitting Transistors exist, and apparently they do! en.m.wikipedia.org/wiki/Light-emitting_transistor

@zyeborm Год назад

@@AgentOrange96 all transistors are light emitting if you operate them wrong enough.

@W1RMD Год назад

Awesome video! Perhaps the first VFD has more burn in those top and bottom segments due to age and previous use. Also I use LM7448's and UDN2981AT. Seems like there's better, easier options out there, but this works good. I'm curious about multiplexing these tubes and reducing down to one 7448 and one driver IC. Just having fun here. The wife bought me a couple of IV-12 VFD tubes and I'm having a blast designing a 59 second timer to go with my giant led clock. I love your unusual use of tubes!

@michaelturner4457 3 года назад

13:55 I got a nixie tube watch that has two ИВ-15(IV-15) in it, that looks awesome. :) And I just subscribed to your channel.

@UsagiElectric 3 года назад

Thank you so much! Those IV-15 VFDs are just awesome looking, I can'6t wait to get my hands on the ones I ordered!

@stamasd8500 2 года назад

You are actually also undervolting the VFD's plates. The datasheet states that the plate voltage in normal use should be 25-30V. So even 24V is under the nominal voltage required. And the 17-18V you're providing even more so.

@therealchayd 2 года назад

There are some even tinier 1-bit VFDs in the form of the CV6094 (used on the British Elliott 803 and Marconi TAC computers)

@michaelturner4457 2 года назад

I got a nixie watch that has two IV15 VFDs as auxiliary lighting, instead of LEDs. and they do look awesome.

@robot797 3 года назад

the grid is also used to multiplex the displays most devices coupled all grids together and that means that you need to multiplex to get them to show diferent things

@weirdboyjim 3 года назад

That was my thought as well.

@michaelardai9703 3 года назад

That's exactly how multi-digit VFDs (like the IV-18 or IV-27) work

@UsagiElectric 3 года назад

I can't believe I forgot to mention that! Especially considering that I not only mentioned it in the "In a Minute" episode on VFDs I did, I also drew a diagram showing the principle!

@geneh9210 3 месяца назад

Konichiwa, I sent you a message over on Hackaday, but I'm not sure how often you check that out. I missed a couple of messages to me :) Anyway, I picked up some LD8035 VFD tubes and TD62783AP driver ICs in Osaka about 18 months ago. I finally had a chance to use them in a clock project. The project is nearly complete, but I have one issue. I'm using 12VDC for the ICs and an ESP32 via the ICs to control the plates and grids. The time is correctly displayed on my clock, but the plate segments don't appear to turn off fully...a little residual ghosting can be seen on the digits. I mistakenly used 4.7k pull-down resistors on that lines between the driver ICs and the VFD tubes for the grid and plate connections. I meant to use 47k resistors like a similar project that I found. Do you think that could be causing the ghosting? (lower value resistor is too strong of a pull-down on the lines?) That's the only real difference I can tell in my hardware setup. Thank you for any ideas or recommendations on troubleshooting.

@hmbrt12 3 года назад

Beauuuuuuutiful!!!👌🏻⚡🤖

@UsagiElectric 3 года назад

Thank you!

@RobertBaruch 3 года назад

Did you say at one point you'd be willing to use semiconductor diodes? Because you can create a 4x7 matrix ROM out of diodes.

@UsagiElectric 3 года назад

Yup! I do indeed use semiconductor diodes, which is bordering on not quite period correct (though IBM did use Germanium diodes in quite a lot of their later tube computers), but it means that a demultiplexer and diode matrix is by far the easiest way to convert binary to hexadecimal.

@erlannderrantem6972 Год назад

Actually the whole concept of using a vfd as a triode is used by Korg as the so called NuTube 6p1. They use it in some newer vox amps, in a specially designed tubescreamer and as a suboscillator/saturator in a Korg volca bass synthesizer.

@Bleats_Sinodai 2 года назад

Those NuTubes that KORG launched a few years ago are basically just VFDs being used as tiny triodes.

@UsagiElectric 2 года назад

Yup, those are really, really cool! Though, interestingly, you can use just about any VFD that has a grid (the tall ones with the dots don't actually have a grid) as a little amplifier too. A VFD out of an old VCR could totally be a low voltage pre-amplifier tube!

@michaelardai9703 3 года назад

The Noritake/Korg Nutube 6P1 is a VFD that was designed to be an audio headphone amp. As for the binary to 7-segment decoder, can you do something with lightbulbs and photoresistors like HP used to do for nixie decoders?

@UsagiElectric 3 года назад

I really want to get my hands on a Nutube one of these days, I think they're such an awesome little VFD. But, they're pretty compact and have two directly heated triodes, so you could theoretically stand them on edge and stack them together to build some pretty compact triode logic. I could probably get 8 to 12 of those in the same space as a 4 standard dual triodes, which would be pretty wild! Oooh, I haven't heard about that type of Nixie decoder. Everytime I search for Nixie decoder information, it's always done with ICs or transistors of some type. Do you have any links to the photoresistor method? That sounds really interesting!

@michaelardai9703 3 года назад

@@UsagiElectric Look at the HP AC4J or similar counter modules. They use dual triodes as flip-flops for the counter with neon bulbs as the outputs. Those neons shine on photoresistors; multiple neons on one resistor forms an OR, multiple resistors in series an AND. www.decodesystems.com/nixie.html has a description

@UsagiElectric 3 года назад

@@michaelardai9703 That's insanely brilliant! Those HP engineers were on another level. Though, I'm having a hard time wrapping my head around 1242 BCD, which is what is being used here: www.decodesystems.com/hp-decimal-display-1-m2.gif 1100 is actually six instead of 12, which threw me for a loop. But, then again, it just shows how much of a pain it is to go from Binary to Decimal. Even the engineers back in the day were having to come up with really strange and inventive ways to make it work.

@kylejones8392 3 года назад

@@UsagiElectric Honestly, it would be way cheaper to use subminiature triodes since those Korg tubes are like $50 each on their own and their current handling ability is almost nothing. I have one in a headphone amp. Something like the IV-15 or DM-160 might work? They were mostly used as indicators for computer output but might work otherwise as at least the DM-160 saw use in a portable battery radio. Another option is something like the 6BF7 which is a subminiature dual triode similar to a 12AU7 but without the dissipation of its larger brother and doesn't use directly heated cathodes so power supply is a bit simpler.

@UsagiElectric 3 года назад

@@kylejones8392 Oh yeah, those Korg nutubes are quite expensive, I was more thinking if someone had a whole box of them they were tossing out, haha. Interestingly, the IV-15 VFDs are the ones I've ordered to use as my indicators! They're just gorgeous looking little tubes and I can't wait to get my hands on them. I would love to get my hands on some subminis someday and see how they compare to their bigger counterparts when using at low voltage like I am. With just 24V and a 10k plate resistor, the amount of electrons that can flow is extremely limited, so subminis may be right in their element!

@frankowalker4662 3 года назад

I have a big old calculator that uses 7 segment VFD's. Though I have'nt plugged it in for years, so I don't know if it still works.

@UsagiElectric 3 года назад

I know that on a fundamental level, LEDs are better, but VFDs have such a nice glow to them that I wish they were still being used in calculators, clocks, etc. I really gotta find a good use for those 7 segment VFDs one of these days!

@trevorvanbremen4718 2 года назад

I suspect a slight issue with your analogy at around 08:30 (One end of the heater being HOT with the other being LESS HOT when using a DC filament supply). The filament has a constant current passing through it and a constant resistance per unit length of filament so the actual temperature should be largely constant along its length. However... The potential difference (a.k.a. voltage) between the filament and it's associated array of plates WILL differ between the two ends and it's THIS aspect that will cause the plate emission to be lower. At the LEFT end of the filament, the voltage between filament and plate will be 1 volt (filament drive voltage) MORE than it is at the RIGHT end of the filament. Edit: As for driving the filament from an AC source, my understanding is that this 'works' if the AC source is a center-tapped source with the MID POINT of the AC source being the common / ground. Therefore, the long term AVERAGE plate-to-cathode (a.k.a. plate-to-filament) voltage will remain pretty constant. The mid point of the filament will always be at a 'virtual ground' potential. Ar the extreme ends of the filament (left and right), the plate-to-cathode voltage swings equally above and below that of the mid point thereby producing a broadly similar illumination level.

@Lee_Adamson_OCF 3 года назад

Implement your VFD driver in a GAL for low-chip solution.

@UsagiElectric 3 года назад

Interestingly, one of the uses I'm looking into using VFDs for now is not as indicators, but actually as memory! I'm thinking it can be a nice, compact way (compared to vacuum tubes) to implement a type of memory that can be accessed by my tube computer. Here's an early proof of concept test on the breadboard: imgur.com/gallery/yz4yWDY

@irgski 2 года назад

use the vf 7-seg tube but only display “0” or “1”…should be a pretty simple decode algorithm.

@UsagiElectric 2 года назад

I actually did exactly that in a future video! (The last few minutes of this one: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-AB84_vbH_e8.html )

@AndersNielsenAA Год назад

You’re absolutely sure you wont make a 74LS47 using tubes? 😅

@RWBHere 6 месяцев назад

That tube has 8 fluirescing plates. There is a decimal point.

@nepaliangusyoung Год назад

System IC

@ropersonline 2 года назад

Not gonna lie, every time I hear you pronounce "quite" as "quiet", it gets me a little. It's an odd quirk.

@UsagiElectric 2 года назад

Oooh, going for an old video! It's been a while since I've gotten this comment on my pronunciation of "quite". Check out some newer videos, I think I've gotten a lot better about it!

@ropersonline 2 года назад

@@UsagiElectric No worries, it's just something that got me, so of course I had to comment on it. Don't mind me; it boosts engagement. ;-P I've gone back to watching the whole vacuum tube computer series in chronological order. I think I'm all caught up on the Centurion series. You might encounter more of this as new subscribers discover your impressive backlog.