Having had experience of the russian VFDs, the anodes may even need to be pulled as far as -15V to get a proper switch-off. I seem to recall that the heater was run from a centre-tapped secondary, and the centre was pulled to -1.5V to get the display even.
I have the same displays and they work fine if you center tap the AC filament voltage and have pull down resistors on the segments to a negative voltage
Чем больше дисплей, тем сложнее в производстве он и тем больше проблем с ним. Это классика, особенно в советском союзе. Например, дисплей ГИП-10000 имел еще более серьезные проблемы, чем этот ИЛЦ1-1\7. Than bigger the display than it more complex to fabricate and it have more issues. It's classics, especially in the Soviet Union. For example the GDCCI-10K(GIP-10000) had even more serious issues than that ILC1-1\7. But you doing very nice. Really. Well done comrad, Fran. Keep doing something interesting stuff next. That's goddamn cool and interesting. Also thanks for inspiring, now I wanna made my smaller ILC clocks. (and sorry for my eng, hehe)
try a small reverse voltage on the off segments, also do the off segments still glow if they are left open circuit? run the heater from an isolated supply
You can't run the heater from an isolated supply, because the heater is the electron source. Without a potential difference between the heater and the segments, they won't light, because there is no acceleration force on the electron cloud.
The heater AC supply should be grounded at the transformer centre tap so that each end of the filament alternates between +2.5 and -2.5V relative to ground. If it's grounded at just one end, the other end will alternate +/-5V, which may cause uneven brightness or segments to light when they shouldn't
@Fran Blanche I suspect the ghost illumination is from feedback through the other segments due to the transistor outputs being floating, try a hard turn off by tying them to the same common supply rail through a second set of transistors.
Definitely need the filament voltage to be grounded at a center tap to stop the ghosting. For really small VFD you can sometimes ground one end of the filament. On some of the really big displays you not only need to have the filament voltage using a center tap, but you might have to float the center tap a few volts positive.
FYI - The VFD displays that I have worked with suggest that the filament is powered by a center tapped secondary and the center tap is grounded. If you ground one of the filament pins to ground, you end up with uneven illumination.
As others have said, negative bias to turn segments off. Also try turning one segment on at a time. I don’t know if you can tie all the same segments together between digits and controlling each digit by changing cathode voltage. You could control all six digits with 13 lines. I’ve seen similar circuits and code for LED displays. The main difference would be going from positive to negative bias instead from high to low on each pin. The code would cycle through each of the seven segments turning on the cathode for each digit that wanted to light that segment.
Try one of those two things: 1) decrease the grid voltage + increase the segment voltage instead, or 2) apply negative voltage to a segment in respect to grid and see if it shuts off.
Hey, don't give up just yet! Perhaps some clever biasing of the off segments could do the trick, perhaps driving them negative with regards to the filament? Also you can have a four digit clock, even though six digits are cooler, you can make do.
The wonderful thing about vacuum tubes: If they don't want to work at the specified voltages you can MAKE them work at alternate voltages. At least for a while...
It can be better because of lower voltage difference on the heater ends. But I think if the transformer is insulated from the rest of the circuit and you apply cathode voltage by resistors to both ends of the heater, it should work too. Heater impedance is much lower (its wire) than A-K and G-K impedance. I would also try to put grid voltage lower in OFF mode, not to ground, but to about -5V. It can be also done just by putting cathode to the +5V level.
Fran, I love the effort you take to drive these devices into operation that most people would never get to see otherwise. Most people would only ever get to see just the cold device, and just fantasize about how it may look and probably espouse an opinion about it. Thanks for your most excellent efforts!!!
There are lots of great comments about negative for off segments and the gradient being intentional for using DC for the filaments. These are strange/quirky displays and probably do require strange/quirky circuits to drive them. Run your tests first in a quick messy way to prove the concepts before building your final circuit board.
Tubes do not work like transistors; Grid needs a negative relative to the heater, note that the heater dc voltage varies along the length of the tube. Reverse the AC to the filament and the top traces will fail to turn off proving this idea.
Why not use shift registers and PWM the OE for brightness control. Saves a bunch of pins, infinitely expandable and with a little bit of extra code just as easy to use as digitalwrites. I’ve made a large clock using LED strips (3 die plastic encased modules) for each segment, to drive them I have them wired common anode and using ULN2003’s, CD4511’s and a shift register per 2 digits, 6 digits and 8 additional LEDs for status indicators and the colons. As for the segments on that VFD, you may have to bias some or all of the segments or the heater possibly with a negative voltage
You have to have the filament voltage center tapped to the cathode voltage. If you don't one end of the filament will have a higher voltage to the anode creating dark on one end and bright on the other. Also the open collector drivers are allowing the leakage glow. A totem pole drive will force the segments off.
Better luck on the next project. That’s a bummer to put so much work into an idea to have it go south like that. But it did lead to my new favorite Fran sound bite! “Obey Me!”, lol. Lots of luck on the next one.
Tube may be a bit gassy. Also might need to drive the grid negative to fully turn off display. Try driving the display with complementary mosfets, one to connect to B+ when ON and one to pull to ground when OFF.
Fran, try cleaning your driver board from flux, I had a similar problem with my IV-22 VFD clock. The segments draw so little current that a voltage next to a pin can supply enough current to the next pin on the driver board. I was able to increase the isolation between each pin, and the ghosting segments was solved. Worth a quick try!
Old school EE trained here. Looks like one of two things: uneven getter fire, vacuum leak during mfg. Its obvious there is contamination. Looks like u got mfg defect from QC rejects.
Thank you for publishing this project Fran. You have demonstrated a valuable lesson in engineering, that sometimes with the best and careful efforts the results are not what we are striving for.
How did you wire up the filament? It should be driven with a center tapped transformer. The filament should be biased above ground by about 5-10 volts by connecting a zener diode between the center tap and ground.
I feel like it’s a kid’s show when they want you to shout the answer at the tv. “The voltmeter is doing it. It’s grounding the segments it’s connected to.”
When you have a filament you have to manage how the voltage drop goes along its length. You can minimize its effect by center tapping the filament voltge to halve the offset. You also can raise the center tap to effectively negative bias the grids.
Hey Fran! I remember that the VFD tubes needs a negative voltage on the grid to turn segments off. Furthermore why don't you use SOCKETS to quickly diconnect-reconnect them?
High impedance on anodes, capacitive coupling from filament takes them positive so they turn on where the grid is closer. Also filament power should be effectively centre tapped, and then you need to make sure the off anodes are more negative than the filament at all times .. perhaps connect filament ' centre tap' to a positive supply.
Oh wow... I love VFDs, from the standpoint of a glassblowing and neon bender. Looks like they epoxied on the processing/evacuation tubulation. There must be a small drilled hole under there, but it's hard to see.
With all due respect, Fran, you're trying to use some cheap third party electronic circuits with this tube and you are quick to assume that the tube is low quality. I have lived in the Soviet Union and we were surrounded by all kinds of soviet made electronic clocks installed at factories, airports, train and metro stations and they all worked just fine with proper soviet electronics inside.
Some dark glass, or foil might surpress the light coming through. And some calibrated tuned PWM values per segment might make it an evenly lit when segment is on.
Having grown up in USSR I have seen these massive VFD 7 segment displays everywhere (schools, factories, etc). They were usually mounted in hallways to show the time, temperature and shift/class details. They were bright and clear (no ghosting segments). Perhaps they need negative bias on the segments to turn them off?
I would love a clock made from these. It would not even have to be presented in any fancy way. The could be wrapped in a frame made of recycled 2x4s, for all I care. It reminds me that I have a bunch of VFD tubes laying around - not so big and cool as these, but the real deal. I should get after a clock project, maybe?
I wonder if there's a transparent conductor that can hold up to a Neon discharge. I think a hybrid of Nixie and VFD that's fully transparent would be really cool!
these tubes (indicators) are actually made just for wall mounted clocks ) (and that's why they are so cheaply made. usually they work ok but they are very fragile so yours must be damaged) another iconic russian clock tube is "ИВ-26" which are long thin tubes with one row of 7 round dots (spacing between dots within one tube is equal to spacing between similar dots in separate tubes put next to each other). those tubes are made for combining them in 7xN dot matrix.
Like others are saying, I'd try driving the voltage the other way. I might also carefully try palm of hand persuasive percussion. After spending so much time on it, I would build the clock anyway and just learn to appreciate the imperfections--think of it as retro or rustic or something.
Are the transistors getting enough base current? You may have to set them up in a darlington array as those microcontrollers can usually only source 20mA.
Can't help but imagine a Frankenstein's clock with all manner of mismatched VFs, nixies, 7-segment LED, and whatever else comes out of the woodwork. While innocently telling the time, the casual observer might mistake it for some mad scientist's countdown clock to armageddon.
Fran if this is NOS then you would be wise to preheat the unit for a while also when you closeuped the lower portion of the tube the separation between the segments did not look sufficient...maybe each element needs a bleed resistor?
Fran, I believe you should build the clock anyway - despite the wonky appearance when operating. It's still gonna look really cool. Consider the historical value and backstory. This set of poorly operating giant (post-Soviet-era) VFD tubes somehow managed to get through production despite the collapse of the USSR. Of course they're wonky!
Is it possible to put a little negative voltage on the segments that should be off to stop the bleedover? Maybe it's a common, negative, positive thing.
As other folks have said, i bet there is a way to brute force the digits to be off. I also wonder if relays really would be better for these behemoths. a 4 digit clock would still be really cool.
Try reverse biasing them, ever so slightly? Can you shift the grid with a thump? Maybe are design to be used upside down, or with a smoked glass filter?
![Seungmin "그렇게, 천천히, 우리(As we are)" | [Stray Kids : SKZ-PLAYER]](http://i.ytimg.com/vi/kAzmhLHePqU/mqdefault.jpg)
Hello there. It is actually recomended to pull anodes to about -0.3 ~ -1V to turn them complitely off.
Hopefully she sees this!
Having had experience of the russian VFDs, the anodes may even need to be pulled as far as -15V to get a proper switch-off.
I seem to recall that the heater was run from a centre-tapped secondary, and the centre was pulled to -1.5V to get the display even.
That was my first thought as well.
That is not only recommended, it's a must. Otherwise this ghosting is unavoidable, when multiplexing it's even more visible.
So that's why my CD40xx based driver was ghosting, I suspected that something was wrong._
I have the same displays and they work fine if you center tap the AC filament voltage and have pull down resistors on the segments to a negative voltage
someome else mentioned the negative voltage too.
i think thistube deserves a whole nother in depth video.
cool beans.
Чем больше дисплей, тем сложнее в производстве он и тем больше проблем с ним. Это классика, особенно в советском союзе. Например, дисплей ГИП-10000 имел еще более серьезные проблемы, чем этот ИЛЦ1-1\7.
Than bigger the display than it more complex to fabricate and it have more issues. It's classics, especially in the Soviet Union. For example the GDCCI-10K(GIP-10000) had even more serious issues than that ILC1-1\7. But you doing very nice. Really. Well done comrad, Fran. Keep doing something interesting stuff next. That's goddamn cool and interesting. Also thanks for inspiring, now I wanna made my smaller ILC clocks. (and sorry for my eng, hehe)
I was going to bring up bias voltages, but the comments already explained it better than I could have LOL.
Even though this didn't go to plan; I think they are still kinda pretty.
non isolated heater,
bottom edge of heater is 5v higher than top of heater and its acting like a drive
In fact it should be lower - heater is cathode, not anode.
@@xsc1000
i didnt see how she hooked it up, but i still suspect its not ac and isolated
came back to watch again after the new one - thanks for showing rhis one and that one. Showing the process
try a small reverse voltage on the off segments, also do the off segments still glow if they are left open circuit?
run the heater from an isolated supply
The display is probably better with relays? You could use transistors on the first level, controlling relays on the second level.
I'd say to ground the OFF segments properly.
You can't run the heater from an isolated supply, because the heater is the electron source. Without a potential difference between the heater and the segments, they won't light, because there is no acceleration force on the electron cloud.
The heater AC supply should be grounded at the transformer centre tap so that each end of the filament alternates between +2.5 and -2.5V relative to ground. If it's grounded at just one end, the other end will alternate +/-5V, which may cause uneven brightness or segments to light when they shouldn't
@@g0hjq exactly what I was thinking.
Regardless of the planned aim, the work of art (tube) was so incredible to view, and actually see operating :-)
@Fran Blanche I suspect the ghost illumination is from feedback through the other segments due to the transistor outputs being floating, try a hard turn off by tying them to the same common supply rail through a second set of transistors.
Definitely need the filament voltage to be grounded at a center tap to stop the ghosting. For really small VFD you can sometimes ground one end of the filament. On some of the really big displays you not only need to have the filament voltage using a center tap, but you might have to float the center tap a few volts positive.
Some quick research turns up that some VFDs are made with a height gradient because the filaments are supposed to be run on DC.
Love your videos on anything electronic and glowy! Fascinating.
FYI - The VFD displays that I have worked with suggest that the filament is powered by a center tapped secondary and the center tap is grounded. If you ground one of the filament pins to ground, you end up with uneven illumination.
As others have said, negative bias to turn segments off. Also try turning one segment on at a time. I don’t know if you can tie all the same segments together between digits and controlling each digit by changing cathode voltage. You could control all six digits with 13 lines.
I’ve seen similar circuits and code for LED displays. The main difference would be going from positive to negative bias instead from high to low on each pin.
The code would cycle through each of the seven segments turning on the cathode for each digit that wanted to light that segment.
Fantastic..allways wanted too see these displays/ lite up..GreatJob...BeHappy😃
Fran not being satisfied and demanding "Obey Me" was the energy inspiration i needed for today.
I get the exact same feeling when working with legacy plc hardware
weird fetish but ok
:)
Try one of those two things: 1) decrease the grid voltage + increase the segment voltage instead, or 2) apply negative voltage to a segment in respect to grid and see if it shuts off.
Hey, don't give up just yet! Perhaps some clever biasing of the off segments could do the trick, perhaps driving them negative with regards to the filament?
Also you can have a four digit clock, even though six digits are cooler, you can make do.
The wonderful thing about vacuum tubes: If they don't want to work at the specified voltages you can MAKE them work at alternate voltages. At least for a while...
Think you need a center tap transformer to bias the cathode. Verify the cut-off voltage for the bias. This is easy to correct with a zener diode.
It can be better because of lower voltage difference on the heater ends. But I think if the transformer is insulated from the rest of the circuit and you apply cathode voltage by resistors to both ends of the heater, it should work too. Heater impedance is much lower (its wire) than A-K and G-K impedance. I would also try to put grid voltage lower in OFF mode, not to ground, but to about -5V. It can be also done just by putting cathode to the +5V level.
Try moving a powerful magnet over the front and back of the display. Maybe also check the gas content?
Fran,
I love the effort you take to drive these devices into operation that most people would never get to see otherwise.
Most people would only ever get to see just the cold device, and just fantasize about how it may look and probably espouse an opinion about it.
Thanks for your most excellent efforts!!!
Alas. You did what you could. Thanks for all the cool things you share.
I would love to see the calculator the Soviets used those in!
There are lots of great comments about negative for off segments and the gradient being intentional for using DC for the filaments. These are strange/quirky displays and probably do require strange/quirky circuits to drive them. Run your tests first in a quick messy way to prove the concepts before building your final circuit board.
Tubes do not work like transistors; Grid needs a negative relative to the heater, note that the heater dc voltage varies along the length of the tube. Reverse the AC to the filament and the top traces will fail to turn off proving this idea.
Dammit!! That would have looked so cool!!!
Why not use shift registers and PWM the OE for brightness control. Saves a bunch of pins, infinitely expandable and with a little bit of extra code just as easy to use as digitalwrites. I’ve made a large clock using LED strips (3 die plastic encased modules) for each segment, to drive them I have them wired common anode and using ULN2003’s, CD4511’s and a shift register per 2 digits, 6 digits and 8 additional LEDs for status indicators and the colons.
As for the segments on that VFD, you may have to bias some or all of the segments or the heater possibly with a negative voltage
You can try that with your version.
You have to have the filament voltage center tapped to the cathode voltage. If you don't one end of the filament will have a higher voltage to the anode creating dark on one end and bright on the other. Also the open collector drivers are allowing the leakage glow. A totem pole drive will force the segments off.
I know a lot about driving these displays from the way they are used in VCR's. There 3 voltage points to hit plus the filament current drive.
Better luck on the next project. That’s a bummer to put so much work into an idea to have it go south like that. But it did lead to my new favorite Fran sound bite! “Obey Me!”, lol. Lots of luck on the next one.
**Beautiful** vacuum tubes! In perfect condition too!
Week down the tubes! I see what you did there.
Tube may be a bit gassy. Also might need to drive the grid negative to fully turn off display.
Try driving the display with complementary mosfets, one to connect to B+ when ON and one to pull to ground when OFF.
Fran, try cleaning your driver board from flux, I had a similar problem with my IV-22 VFD clock. The segments draw so little current that a voltage next to a pin can supply enough current to the next pin on the driver board. I was able to increase the isolation between each pin, and the ghosting segments was solved. Worth a quick try!
Absolutely beautiful objects! Wow.
I love your channel! So glad I found it! I loved analog and 60s, 70 , 80s, tech was so durable….
Are you grounding the center tap of the AC heater supply ? looks like you have DC across the heater. This will cause it to be brighter at one end.
I think she tried both AC and DC.
DC across the heater is bad. Grounding the center tap removes any differential caused by the bias supply across the heater resistance.
It seems like it. Hope she see it
amazing tech even if they act funky. Love the fact you can break down how these work
Old school EE trained here. Looks like one of two things: uneven getter fire, vacuum leak during mfg.
Its obvious there is contamination. Looks like u got mfg defect from QC rejects.
Wow they look more like glassbricks 😉.
Such cool LEDs
Are you actively driving the segments positive to turn them off, or just grounding them to turn them on ?
i second this, theres current going somewhere to make those glow. there could be leakage from the perfboard or solder flux as well
Isn’t that logic for Nixie? I think VFD is positive grid and segments for on, and negative voltage for segment off.
I agree... these VFDs are alot like driving IGBTs and MOSFETs in which the gates must be pulled in both directions ... and never left floating.
Had the same issue when I got some soviet vfd-s, pullup resistors fixed it. Also running at too high pwm/switching frequency causes weird effects.
yes, I was going to say the same, off segments shorted to grid
Thank you for publishing this project Fran. You have demonstrated a valuable lesson in engineering, that sometimes with the best and careful efforts the results are not what we are striving for.
Yes, with experimental error ruled out by careful measurement, including within out-of-spec regimes.
How did you wire up the filament?
It should be driven with a center tapped transformer. The filament should be biased above ground by about 5-10 volts by connecting a zener diode between the center tap and ground.
I feel like it’s a kid’s show when they want you to shout the answer at the tv. “The voltmeter is doing it. It’s grounding the segments it’s connected to.”
That's a bummer, but this was a great video! Would love to see a tear-down of one of these!
They look amazing. I'd love some. Not sure what I'd use them for but I'd love them nonetheless.
When you have a filament you have to manage how the voltage drop goes along its length. You can minimize its effect by center tapping the filament voltge to halve the offset. You also can raise the center tap to effectively negative bias the grids.
Hey Fran! I remember that the VFD tubes needs a negative voltage on the grid to turn segments off. Furthermore why don't you use SOCKETS to quickly diconnect-reconnect them?
Or at least some short wires tacked onto the leads, which would make swapping in/out a lot easier than the proto board plated hole system.
How are you applying the DC filament bias?, to one end of the filament? it should be via a centre tap on the AC filament supply.
Still a cool video, even though the project didn't work. Keep up the great content! Always enjoyable.
Maybe will work as an LCD: need inverse voltages some times to work properly.
is really neat to see though. i love to see stuff like this.
High impedance on anodes, capacitive coupling from filament takes them positive so they turn on where the grid is closer. Also filament power should be effectively centre tapped, and then you need to make sure the off anodes are more negative than the filament at all times .. perhaps connect filament ' centre tap' to a positive supply.
Love this displays. No need to google for a wiring diagram, it's already part of the design :-)
with these big plates you need to negatively bias the anodes to turn them off.
Gosh, gotta love those mammoth displays!!!! ⚡🤘🏻🤖💕
Did you try to apply a reverse voltage to turn off the segment?
I just come here to hear you saying "it´s Fran. Again. At the bench". It makes my day.
This episode has a bonus “Obey me!”
Oh wow... I love VFDs, from the standpoint of a glassblowing and neon bender. Looks like they epoxied on the processing/evacuation tubulation. There must be a small drilled hole under there, but it's hard to see.
Beautiful!
With all due respect, Fran, you're trying to use some cheap third party electronic circuits with this tube and you are quick to assume that the tube is low quality. I have lived in the Soviet Union and we were surrounded by all kinds of soviet made electronic clocks installed at factories, airports, train and metro stations and they all worked just fine with proper soviet electronics inside.
Very likely the best thumbnail ever.
Nice meter!
just to cut down on the variables, did you try driving one seg at a time with just power supplies and alligator clips?
Ooooh. NICE!
Some dark glass, or foil might surpress the light coming through. And some calibrated tuned PWM values per segment might make it an evenly lit when segment is on.
Perhaps try resistor networks on each segment to fine tune on/off voltages?
Having grown up in USSR I have seen these massive VFD 7 segment displays everywhere (schools, factories, etc). They were usually mounted in hallways to show the time, temperature and shift/class details. They were bright and clear (no ghosting segments). Perhaps they need negative bias on the segments to turn them off?
I could listen to you forever. If I ever got to do Desert Island Discs I think your RUclips channel would be my luxury item.
I would love a clock made from these. It would not even have to be presented in any fancy way. The could be wrapped in a frame made of recycled 2x4s, for all I care. It reminds me that I have a bunch of VFD tubes laying around - not so big and cool as these, but the real deal. I should get after a clock project, maybe?
Have you tried using an isolated supply for the heater?
Frantastic show!
I wonder if there's a transparent conductor that can hold up to a Neon discharge. I think a hybrid of Nixie and VFD that's fully transparent would be really cool!
these tubes (indicators) are actually made just for wall mounted clocks ) (and that's why they are so cheaply made. usually they work ok but they are very fragile so yours must be damaged)
another iconic russian clock tube is "ИВ-26" which are long thin tubes with one row of 7 round dots (spacing between dots within one tube is equal to spacing between similar dots in separate tubes put next to each other). those tubes are made for combining them in 7xN dot matrix.
Are you letting the off segments float?
OOOMMGG they're so big and prettyy
I love the thumbnail! 😆
I think you could make some pretty authentic looking ones with electroluminescent segments. It would take some work though
Have you considered a DC filament supply ?
Those are gorgeous :o
This is what i subbed for, big ol' chonker tubes getting the workover
Like others are saying, I'd try driving the voltage the other way. I might also carefully try palm of hand persuasive percussion. After spending so much time on it, I would build the clock anyway and just learn to appreciate the imperfections--think of it as retro or rustic or something.
It makes me wonder if a pull up/down resistor would help. Something to pull the emitter voltage to the grid when it's not being used.
The feels.
Shunt unused segments to ground?
Are the transistors getting enough base current? You may have to set them up in a darlington array as those microcontrollers can usually only source 20mA.
@Fran Blanche ... That point in the video with the glare made me wonder if a coat rack and a dark cloth behind the camera might help next time
Can't help but imagine a Frankenstein's clock with all manner of mismatched VFs, nixies, 7-segment LED, and whatever else comes out of the woodwork. While innocently telling the time, the casual observer might mistake it for some mad scientist's countdown clock to armageddon.
Fran if this is NOS then you would be wise to preheat the unit for a while also when you closeuped the lower portion of the tube the separation between the segments did not look sufficient...maybe each element needs a bleed resistor?
You're gonna work it out, Fran, don't give up! Even if these tubes are not well made, I already love them.
Fran, I believe you should build the clock anyway - despite the wonky appearance when operating. It's still gonna look really cool. Consider the historical value and backstory. This set of poorly operating giant (post-Soviet-era) VFD tubes somehow managed to get through production despite the collapse of the USSR. Of course they're wonky!
Where did you source these?
Is it possible to put a little negative voltage on the segments that should be off to stop the bleedover? Maybe it's a common, negative, positive thing.
Chin up Fran! Was still great to watch, and still a wonderful demonstration of a really cool display technology, even if it had its faults!
As other folks have said, i bet there is a way to brute force the digits to be off. I also wonder if relays really would be better for these behemoths. a 4 digit clock would still be really cool.
Aren't you supposed to pull negative voltage to stop the ghosting? 🤔
Maybe try DC for the filament?
If that is a display, are the driver transistors the size of soup cans?
I wonder if turning the first display upside down changed the evenness (maybe the grid is loose and is affected by gravity)?
Try reverse biasing them, ever so slightly?
Can you shift the grid with a thump?
Maybe are design to be used upside down, or with a smoked glass filter?
Haven't hear ATMEL repeated like that since I worked there.