The tungsten can be brazed with a TIG torch. (Using a flame is hopeless!) Nickel wire can be used as the brazing metal. The welder I have is a Primeweld TIG225 which is an excellent low cost unit that includes everything you need except the argon tank.
@@daliborfarny Oh my! I watch from here and think of how great you are doing! Be careful to not get too big too quick. Keep out of debt! Let me know if there are tech problems you are having, such as the current limit for the cathode cleaning machine. I could have solved that for you immediately! (free!)
I just love watching your journey and your passion for your work, I’d love to be able to work at your castle, keep it going fantastic. Regards Matt from east coast of Australia.
That vacuum relay looks to be brazed. I'd try nickel or silver solder, there is very little that can't be brazed or soldered together with them. I am surprised you went with inductive arc current limiting, I figured some high-side resistive current limiting would do the trick? Unless the anode capacitance is high enough that it alone can supply enough energy to zap the BE junction when it gets into breakdown? With a hundred ohm emitter resistor I am guessing the current limit is around 40-50 mA and with say a 5 volts BE breakdown you must be seeing > 100 mA current surges during the arcing. The transistor must be getting into second breakdown or something to even be still conducting the arc current for very long once the emitter rises above cut-off, what is its VCE rating? Adding a series diode in the emitter may protect the BE junction (more reverse breakdown than a few volts of the BE junction at least), but then the emitter resistance would need to be changed to reset the normal current limiting point. Kinda surprised you aren't killing something in the base drive circuit if the BE junction is failing, it must have a fairly stiff drive or a lot of capacitance to be able to sink damaging amounts of fault current? It might actually be the transistor capacitances that are letting it kill itself, the miller capacitance in particular... I think a better topology would be to have another small NPN (like a humble 2N3904) steal the base drive from the the HV transistor's base, fed itself from the emitter resistor, that way it need only generate a diode drop of voltage on the emitter before limiting - much less than the diode drop less than logic level, this does reduce the voltage reversal required to reverse bias the BE junction considerably but decreases the emitter sense resistor to about 13 ohms which will drop only a volt or two under fault conditions and the feedback transistor should be very fast in clamping the base drive, discharging any current that feeds though the larger device's significant capacitances.
I wonder if for the burn in you could use RF. I think that might reduce the chance of arcing, and plasma cleaning is a pretty common technique for vacuum chambers.
Unfortunately not. You will face the same issues with arcing. The change in polarity alone (at MHz "speed") will not stop the arc development process. The RF generators are used in applications where the target is a non-conductive material (plastic, glass). For nixie tubes DC power supply seems to be a good choice. I would say lowering the pressure will reduce the arcing tendency, but I am not sure about mechanical limitation of a tube. But maybe for that kind of test you can use thicker glass. Please also localize where you have arcing. Maybe you can also improve wires insulation (and use enamel wires?)
You could add an overcurrent shutdown circuit to the test rig. That could shutdown the output in under a microsecond when the arc is detected. That would prevent the wires from being burned off.
Has anyone who worked on these when they were originally being made reached out to you? We're loosing people from back then quickly so I am really glad you're documenting the art.
dumb question - wouldn't a reducing atmosphere (hydrogen?) easily eat off all the oxides? Vacuum, hydrogen, burn-in, vacuum again, working gas mix, seal off.
Gotta admit that these huge Nixies are absolutely beautiful! Nice drivers too. Makes me wonder if you could use thyratrons for controlling the cathodes in the aging device. That would be pretty cool :). That vacuum relay is loveliness too! There's soooooooooo much stuff I could and would love to learn working with you :)
I find your lack of a freewheel diode on that inductor disturbing. Also ultrasonic soldering can definitely do tungsten, though I can’t say how structural it is.
Add a 15W 230V pygmy lamp in series with the anode supply. This will limit the current that can flow into an arc, and thus reduce the transistor damage. You can use an indicator lamp, which has a nice long squirrel cage filament, so it will rapidly heat up and limit current.
wouldn't separating the wires with some kind of insulator prevent the arcing? Or perhaps creating 5 connection hubs (it looks like you have one, I'm sorry if I'm not getting the terminology right) but distributing the concentration of bare wires brings the possibility of arcs down. perhaps a new way to create the seal with the wires and glass, more of a circular method than a straight line... I mean traditional methods are great but perhaps some innovation or experimentation is required? perhaps you have already solved this and I'm just screaming into the void
I made some linear laser drivers which could drive nixies with a transistor swap. I had some logic and an op amp that'd control current and trigger a failsafe shutoff if current got too high. Tell me if you guys need help.
Please read about "arc management" functionality in DC or RF plasma power supplies (like in TRUMPF or Advanced Energy products). In an overcurrent state the switch off has to happen in 10-20 us, so it has to be pretty quick. Then you have to wait 20-50 us and you can restart the voltage source.
Dalibor, thanks for such an interesting and long video. Can ultrasonic sound remove oxides from the surface inside the bulb? Also, stay with Ukraine, do not support russia with your products, please. Stay Safe!
The current source solution is perfect. I use it widely if need to reduce the power. May be, will be useful a pre-ionisation electrode. The very small current through it (150G resistors) significantly reduce the ignition voltage, and allow to use low voltage switches (80V). Possibly, there isno condition for the arc in this case.
Difficult to doubtful. The glass would cause a lot of problems. Also, they cost 3 legs, two eyebrows, and half a bread loaf to buy or rent. Source, we rent one two months a year to clean rusted architectural features that cannot be removed from their existing location (heritage, structure, rarity or state) in order to preserve them. In two years we'll have enough dough rolling and enough jobs with it to justify spending. And the biggest problem is there's a lot of chinese fakes out there, that barely do the job, but really are a waste of time, and can even be dangerous.
Interesting idea - the glass would distort the light, the digits are obstructing each other, the area on the bottom of the digits is not accessible. These are the most important reasons why this would be hard..
@17:10 about the oxidization of the digits during sealing, why don't you flood the tube from the back with nitrogen to avoid that? You could probably redirect the nitrogen from your oxygen concentrator rig..,
Love the video! Regarding the oxidation issue you're having, have you considered flushing the tube with forming gas? It's a mixture of Argon or Nitrogen with 10% Hydrogen and works as a reducing agent without being a fire hazard. As far as I have seen you have only one access port to the tube so a constant flow would be hard to do, but a few mBar of positive pressure inside, coming from that port, might still work out when sealing the front window. Good luck either way!
I wonder if it could be part of the bake out process even? They still have a tubulation for gas access, it'd need to be done through a set of fill and drain cycles though.
It may be possible to use an Op-Amp based overcurrent detection circuit on the tube ager. Depends on how quickly and how high the current spikes during an arc.
You must make a super-mega-top de luxe with numbers in pure gold . The visual aspect will be incredible and surely some customer want it , with the price !
How complicated would it be to do a tube with a stylized clock face that is basically just two rings of tick marks. There would be 24 electrodes, so more than a numerical tube, but each of the electrodes is much smaller and perhaps easier on the aging process. I'd bet you sell a lot of single tube clocks like that and you could do a Kickstarter to gauge interest and collect some startup capital for the project...
You can braze almost everything even stainless steel and tungsten carbide with high silver percentage material and flux, try searching in hvacr shops, they use this brazing material a lot.
Yes, but we kind of struggled with it - it was difficult to keep plasma where we wanted it. But it is a backup solution, we still have the plasma setup prepared..
@@daliborfarny hmm yeah plasma can be unstable like that. i'd be tempted to try somthing like a few MHz at high voltage fead into each number in tern and allow it to capasitively discharge to whereever it wants. it could be guided by external electrodes on the glass if needed, somewhat like a plasma ball. the other internal structures do complicate it a bit and limit your options though
@@daliborfarny To get predictable results, you would need to calculate the electric field strength between elements (ANSYS Maxwell software or equivalent). Additionally the geometry of the elements would have to be designed so as not to exceed the specified field strength in order to minimize the arcs. But this is a higher-tech topic.
Might be possible to burn off oxides using a specific wavelength laser that sweeps back and forth shot in from the side of the tube at a slight angle as it moves up and down so it can reach all the digits. I also wonder about controlled inductive heating of the digits to a low heat, perhaps that would also excite the neon and help it scrub the digits without the risk of arcing. Then, there is always the possibility of flushing the tube with a constant flow of nitrogen while the face is being sealed so there is no oxygen to react with the digits.
Can you ramp the current and voltage up slowly or does it just start full that might help and if you just put low value high power resistors in line with each digit it would at least save your transistors
Likely better than the inductor, is using lots of resistors around those output transistors -- a sudden short at the output can indeed pull up on the whole thing, so, transiently, base and emitter voltages end up near B+. A series resistor to the collector limits that action, while series base resistance limits the current there, and so on. (Emitter is already shown of course.) This sort of crude current limiting, keeps transient currents within ratings, until the transistor/circuit can stabilize back to nominal currents. Which should take on the order of 100s ns for the transistor itself, and a few µs if there's an op-amp servoing its current. Slower than this, I would consider a design problem, and a faster op-amp or more likely, better choice of compensation component values, should be chosen. Don't forget a zener clamp on the base voltage source, or from emitter to sense amp if applicable, to absorb transient currents from their respective resistors/pins. The biggest contra-indication for the inductor is twofold: 1. Yes, it acts as a lowpass filter, slowing the risetime from the arc; but it stores up energy in the process, released immediately after the transistor's current stabilizes, causing its collector voltage to spike, to perhaps 2 * B+ or more. 2. Even the transistor alone, let alone in a feedback loop with an op-amp, can oscillate due to the LC circuit formed by this and the transistor's capacitance, and say inductance of the base circuit, etc. (Series base resistance also helps by killing AC gain.) Usually an R+C (Zobel) network is needed across the transistor (C-E or C-GND), with C a few times larger than Ccb, and R = sqrt(L/C) or thereabouts. Then adjust compensation for best step response (if applicable). This all also assumes adequate thermal management, which is definitely a concern with what looks like top-side-heatsinked DPAKs (there is very little heat flow through thick plastic!). A protective circuit to decrease or remove current flow some time after a fault occurs (excessive high collector voltage?) is mandatory. This could be a foldback circuit, an ADC scanning the outputs and just turns off whichever channel for some ~ms (I guess these are programmable with an MCU of some sort in charge of everything), etc. For testing purposes, I would just as well suggest using TO-220s (or larger) with adequate heatsinking; make it able to handle the load continuously. And, at these voltages, MOSFETs will be better choices than BJTs (likely with little or no change necessary to the circuit, aside from compensating for the somewhat higher Vgs(th)). Keep up the good work, the tubes are beautiful!
Hi! One question, in the minute 6:35, what’s the difference between the tubes on the left and the right, obviously the right are bigger but do you use the tubes on the right to make the “ Puri Nixie Clock | 6? I’m asking because I want it but I don’t know which tubes that one uses and I really would like to have the big one on it. Thanks in advance for your help.
Hi Felix, thank you for being interested in our nixie clocks. All the tubes have the same size, this is "fish eye" effect of the lens I used for the camera - I didn't realize it made such a visible difference to the tubes. We make only one model of a nixie tube (R|Z568M, shortly R), which we use for all our clocks.
Wow!!! It was pretty difficult to solve the board somehow causing spark and conflict with the circuit on the nixie tube. I love that you were trying to focus on these oversize BEAUTIFUL tubes. Great work hard! Thumbs up! Cheers!
Ultrasonic weld and you don't need solder. Could try spot welding too. Switchmode constant current power supply is easy to make for electrode ageing. Plasma etching with the right gas will remove oxide. Can make a simple RF induction plasma driver.
These videos must be an enormous amount of work atop your normal work, but this viewer definitely appreciates it. 👍️ I specifically appreciate the detailed breakdowns you give of the operational and failure modes. That helps this Nixie neophyte to follow what's going on.
two things to consider for the T joint. longer jig to help center the initial bubble by holding where you have it and an extension to the end to guarantee position of the center and repeatability. The second is that to copper ( or other metal ) plate anything you need a very dangerous chemical in the metal salt solution to make even coats. I do not know where you would be able to get cyanide, it seems to act like a catalyst for electroplating, you do not need much, roughly 1 gram per 500 grams ( five hundred ) of the metal salt ( as measured dry ), more cyanide for some metals. The solution will warm up due to the current in the solution so using a fume hood is always recommended. Great work, keep going!
Fascinating and beautiful, as always, thank you. I love reading the comments, especially from people suggesting solutions - it’s not something I know anything about so I’ve no idea how viable the suggestions are, but it’s great to read supportive comments from people trying to help.
Unrelated to the issues, but for the design of digits on the big tubes, it would be really neat to see a design of digits that have a fake isometric "3D" effect on the design of the numbers. Or maybe just some more "modern" style fonts, stuff you would have never seen in a nixie tube. Designer style hehe.
Maybe you're ramping the voltages up too fast? some of these things need to have the voltage applied slowly until it reaches the right amount. Unfortunately that would mean extending the burn time you'll need to have then run.
No, slowing down the voltage won't change anything. The arc discharge is a result of the high voltage applied to the conductors in a high vacuum. Obviously, no arc will occur during the slow voltage ramp. As I said, this is not the source of the problem.
Hey man! If you're looking for a solution to slow down the current spike use an inductor in series. Inductors don't like the current to be rapidly changed through them and do anything to smooth it. In addition to the inductor use what's called RCD snubber to protect your transistors from burning out
@@daliborfarny Because you could achieve very high levels in success and perfection by simple budget hi-tech engineering work than me.. Your progress results made my work useless and trash . I couldn't reach 1/10 th of what you did so far even by small budget biz ...
@@ahmedalshalchi success is relative, I am not talking about all the million frustrating problems on our side (only the technical ones) - we all have them, no worries!
The tungsten can be brazed with a TIG torch. (Using a flame is hopeless!) Nickel wire can be used as the brazing metal. The welder I have is a Primeweld TIG225 which is an excellent low cost unit that includes everything you need except the argon tank.
I wondered how long it takes to find an commend from you on this channel.
@@gekkedirkie I keep an eye on Dalibor in case he gets in trouble! :)
Hi Ron! Nice to see you :-) Good idea with brazing with TIG torch, never thought about it. We have TIG set up, so we can try it.
@@glasslinger I am in one big trouble since I started with nixie tubes :-)
@@daliborfarny Oh my! I watch from here and think of how great you are doing! Be careful to not get too big too quick. Keep out of debt! Let me know if there are tech problems you are having, such as the current limit for the cathode cleaning machine. I could have solved that for you immediately! (free!)
Man, the way you guys work and attack the engineering problems is so inspiring.
Nebylo by lepší, kdyby se ty číslice při tom stárnutí přepínaly pomalejc? :)
I just love watching your journey and your passion for your work, I’d love to be able to work at your castle, keep it going fantastic. Regards
Matt from east coast of Australia.
🙏
That vacuum relay looks to be brazed. I'd try nickel or silver solder, there is very little that can't be brazed or soldered together with them.
I am surprised you went with inductive arc current limiting, I figured some high-side resistive current limiting would do the trick? Unless the anode capacitance is high enough that it alone can supply enough energy to zap the BE junction when it gets into breakdown? With a hundred ohm emitter resistor I am guessing the current limit is around 40-50 mA and with say a 5 volts BE breakdown you must be seeing > 100 mA current surges during the arcing. The transistor must be getting into second breakdown or something to even be still conducting the arc current for very long once the emitter rises above cut-off, what is its VCE rating? Adding a series diode in the emitter may protect the BE junction (more reverse breakdown than a few volts of the BE junction at least), but then the emitter resistance would need to be changed to reset the normal current limiting point. Kinda surprised you aren't killing something in the base drive circuit if the BE junction is failing, it must have a fairly stiff drive or a lot of capacitance to be able to sink damaging amounts of fault current? It might actually be the transistor capacitances that are letting it kill itself, the miller capacitance in particular... I think a better topology would be to have another small NPN (like a humble 2N3904) steal the base drive from the the HV transistor's base, fed itself from the emitter resistor, that way it need only generate a diode drop of voltage on the emitter before limiting - much less than the diode drop less than logic level, this does reduce the voltage reversal required to reverse bias the BE junction considerably but decreases the emitter sense resistor to about 13 ohms which will drop only a volt or two under fault conditions and the feedback transistor should be very fast in clamping the base drive, discharging any current that feeds though the larger device's significant capacitances.
I wonder if for the burn in you could use RF. I think that might reduce the chance of arcing, and plasma cleaning is a pretty common technique for vacuum chambers.
Unfortunately not. You will face the same issues with arcing. The change in polarity alone (at MHz "speed") will not stop the arc development process. The RF generators are used in applications where the target is a non-conductive material (plastic, glass). For nixie tubes DC power supply seems to be a good choice. I would say lowering the pressure will reduce the arcing tendency, but I am not sure about mechanical limitation of a tube. But maybe for that kind of test you can use thicker glass. Please also localize where you have arcing. Maybe you can also improve wires insulation (and use enamel wires?)
Haven't seen that rheostat since elementary school... good memories. Dobra prace sefe
It can dissipate quite a lot of heat, that's why we use it :-)
What about flushing the tube with gass (argon) while sealing the frontwindow (like tig-welders do) ?
This is my question also
We do that - I will try more approaches in the next sealing..
@@daliborfarny Strage, it should not be able to oxidize without oxygene ... maybe you need more flow.
Now that's what I call quality control! - also nice you've introduced some of your staff. How many employees does DF have now?
Just amazing :D Czech People Czech Engineering !
You could add an overcurrent shutdown circuit to the test rig. That could shutdown the output in under a microsecond when the arc is detected. That would prevent the wires from being burned off.
I know the H-tubes are for a project, but a 4 digit Htube clock would be 🔥🔥
Love the tiny fun moments such as when a piece fall on the floor, oups ;) Thanks for these videos ! It is amazing to follow the progress of the team !
18:41 Pretty sus. jokes aside, great video!
Has anyone who worked on these when they were originally being made reached out to you? We're loosing people from back then quickly so I am really glad you're documenting the art.
dumb question - wouldn't a reducing atmosphere (hydrogen?) easily eat off all the oxides? Vacuum, hydrogen, burn-in, vacuum again, working gas mix, seal off.
I need more neon in my life
Gotta admit that these huge Nixies are absolutely beautiful! Nice drivers too.
Makes me wonder if you could use thyratrons for controlling the cathodes in the aging device. That would be pretty cool :).
That vacuum relay is loveliness too!
There's soooooooooo much stuff I could and would love to learn working with you :)
18:42 seems a bit sus to me...
Hey, can you explain a bit what you using for control of the board from the app and make sequences?
I find your lack of a freewheel diode on that inductor disturbing.
Also ultrasonic soldering can definitely do tungsten, though I can’t say how structural it is.
The inductor is not the only component there.. It is additional small circuit..
Add a 15W 230V pygmy lamp in series with the anode supply. This will limit the current that can flow into an arc, and thus reduce the transistor damage. You can use an indicator lamp, which has a nice long squirrel cage filament, so it will rapidly heat up and limit current.
wouldn't separating the wires with some kind of insulator prevent the arcing? Or perhaps creating 5 connection hubs (it looks like you have one, I'm sorry if I'm not getting the terminology right) but distributing the concentration of bare wires brings the possibility of arcs down. perhaps a new way to create the seal with the wires and glass, more of a circular method than a straight line... I mean traditional methods are great but perhaps some innovation or experimentation is required? perhaps you have already solved this and I'm just screaming into the void
I made some linear laser drivers which could drive nixies with a transistor swap. I had some logic and an op amp that'd control current and trigger a failsafe shutoff if current got too high. Tell me if you guys need help.
Please read about "arc management" functionality in DC or RF plasma power supplies (like in TRUMPF or Advanced Energy products). In an overcurrent state the switch off has to happen in 10-20 us, so it has to be pretty quick. Then you have to wait 20-50 us and you can restart the voltage source.
Dalibor, thanks for such an interesting and long video.
Can ultrasonic sound remove oxides from the surface inside the bulb?
Also, stay with Ukraine, do not support russia with your products, please.
Stay Safe!
Neko 9:27 9:32 9:37
The current source solution is perfect. I use it widely if need to reduce the power.
May be, will be useful a pre-ionisation electrode. The very small current through it (150G resistors) significantly reduce the ignition voltage, and allow to use low voltage switches (80V). Possibly, there isno condition for the arc in this case.
You know those videos of a laser hand tool removing rust from steel? Could you use something like that on a sealed tube?
Difficult to doubtful. The glass would cause a lot of problems. Also, they cost 3 legs, two eyebrows, and half a bread loaf to buy or rent. Source, we rent one two months a year to clean rusted architectural features that cannot be removed from their existing location (heritage, structure, rarity or state) in order to preserve them. In two years we'll have enough dough rolling and enough jobs with it to justify spending. And the biggest problem is there's a lot of chinese fakes out there, that barely do the job, but really are a waste of time, and can even be dangerous.
Interesting idea - the glass would distort the light, the digits are obstructing each other, the area on the bottom of the digits is not accessible. These are the most important reasons why this would be hard..
@17:10 about the oxidization of the digits during sealing, why don't you flood the tube from the back with nitrogen to avoid that? You could probably redirect the nitrogen from your oxygen concentrator rig..,
Love the video!
Regarding the oxidation issue you're having, have you considered flushing the tube with forming gas?
It's a mixture of Argon or Nitrogen with 10% Hydrogen and works as a reducing agent without being a fire hazard.
As far as I have seen you have only one access port to the tube so a constant flow would be hard to do, but a few mBar of positive pressure inside, coming from that port, might still work out when sealing the front window.
Good luck either way!
I wonder if it could be part of the bake out process even? They still have a tubulation for gas access, it'd need to be done through a set of fill and drain cycles though.
or they can add some hydrogen to the argon. The hydrogen will eventually escape by itself, most likely during the aging.
Když je práce zábavou. Práce je to určitě náročná a precizní.
It may be possible to use an Op-Amp based overcurrent detection circuit on the tube ager. Depends on how quickly and how high the current spikes during an arc.
I'm not trying to be negative, but why? Nostalgia and the basic beauty and simplicity? Or is there something bigger?
You must make a super-mega-top de luxe with numbers in pure gold . The visual aspect will be incredible and surely some customer want it , with the price !
How complicated would it be to do a tube with a stylized clock face that is basically just two rings of tick marks. There would be 24 electrodes, so more than a numerical tube, but each of the electrodes is much smaller and perhaps easier on the aging process. I'd bet you sell a lot of single tube clocks like that and you could do a Kickstarter to gauge interest and collect some startup capital for the project...
You can braze almost everything even stainless steel and tungsten carbide with high silver percentage material and flux, try searching in hvacr shops, they use this brazing material a lot.
What do you do with all of the failed H-tubes? I would love to own one as a curio.
Might it work to have a high frequency arc detection circuit that interrupts the current for a short time when an arc is detected?
whoa whoa... flux capacitor!? TORR TVR-1!?
Have you thought about useing RF to create a uniform plasma in the tube to do the ageing and avoid the need to drive the tube so hard and risk arcing?
This is a great idea.
Yes, but we kind of struggled with it - it was difficult to keep plasma where we wanted it. But it is a backup solution, we still have the plasma setup prepared..
@@daliborfarny hmm yeah plasma can be unstable like that. i'd be tempted to try somthing like a few MHz at high voltage fead into each number in tern and allow it to capasitively discharge to whereever it wants. it could be guided by external electrodes on the glass if needed, somewhat like a plasma ball. the other internal structures do complicate it a bit and limit your options though
@@daliborfarny To get predictable results, you would need to calculate the electric field strength between elements (ANSYS Maxwell software or equivalent). Additionally the geometry of the elements would have to be designed so as not to exceed the specified field strength in order to minimize the arcs. But this is a higher-tech topic.
Might be possible to burn off oxides using a specific wavelength laser that sweeps back and forth shot in from the side of the tube at a slight angle as it moves up and down so it can reach all the digits. I also wonder about controlled inductive heating of the digits to a low heat, perhaps that would also excite the neon and help it scrub the digits without the risk of arcing.
Then, there is always the possibility of flushing the tube with a constant flow of nitrogen while the face is being sealed so there is no oxygen to react with the digits.
It's really nice to see some updates from you again and being able to follow along with the progress.
Did you consider treating heavily oxidized spots with a laser ?
I have heard that uranium glas binds better to mettal rods/vires than normal glas so it was used in old vacum tubes, is this something you have heard?
Can you ramp the current and voltage up slowly or does it just start full that might help and if you just put low value high power resistors in line with each digit it would at least save your transistors
You should make a subscriber counter with the nixie tubes
Likely better than the inductor, is using lots of resistors around those output transistors -- a sudden short at the output can indeed pull up on the whole thing, so, transiently, base and emitter voltages end up near B+. A series resistor to the collector limits that action, while series base resistance limits the current there, and so on. (Emitter is already shown of course.) This sort of crude current limiting, keeps transient currents within ratings, until the transistor/circuit can stabilize back to nominal currents. Which should take on the order of 100s ns for the transistor itself, and a few µs if there's an op-amp servoing its current. Slower than this, I would consider a design problem, and a faster op-amp or more likely, better choice of compensation component values, should be chosen.
Don't forget a zener clamp on the base voltage source, or from emitter to sense amp if applicable, to absorb transient currents from their respective resistors/pins.
The biggest contra-indication for the inductor is twofold: 1. Yes, it acts as a lowpass filter, slowing the risetime from the arc; but it stores up energy in the process, released immediately after the transistor's current stabilizes, causing its collector voltage to spike, to perhaps 2 * B+ or more. 2. Even the transistor alone, let alone in a feedback loop with an op-amp, can oscillate due to the LC circuit formed by this and the transistor's capacitance, and say inductance of the base circuit, etc. (Series base resistance also helps by killing AC gain.) Usually an R+C (Zobel) network is needed across the transistor (C-E or C-GND), with C a few times larger than Ccb, and R = sqrt(L/C) or thereabouts. Then adjust compensation for best step response (if applicable).
This all also assumes adequate thermal management, which is definitely a concern with what looks like top-side-heatsinked DPAKs (there is very little heat flow through thick plastic!). A protective circuit to decrease or remove current flow some time after a fault occurs (excessive high collector voltage?) is mandatory. This could be a foldback circuit, an ADC scanning the outputs and just turns off whichever channel for some ~ms (I guess these are programmable with an MCU of some sort in charge of everything), etc. For testing purposes, I would just as well suggest using TO-220s (or larger) with adequate heatsinking; make it able to handle the load continuously. And, at these voltages, MOSFETs will be better choices than BJTs (likely with little or no change necessary to the circuit, aside from compensating for the somewhat higher Vgs(th)).
Keep up the good work, the tubes are beautiful!
Hi! One question, in the minute 6:35, what’s the difference between the tubes on the left and the right, obviously the right are bigger but do you use the tubes on the right to make the “ Puri Nixie Clock | 6? I’m asking because I want it but I don’t know which tubes that one uses and I really would like to have the big one on it. Thanks in advance for your help.
Hi Felix, thank you for being interested in our nixie clocks. All the tubes have the same size, this is "fish eye" effect of the lens I used for the camera - I didn't realize it made such a visible difference to the tubes. We make only one model of a nixie tube (R|Z568M, shortly R), which we use for all our clocks.
A back to the future flux capacitor relay?
Ending music: all good folks - lee money
Lovely bit of bombarding!
Use valves instead of transistors
Wow!!! It was pretty difficult to solve the board somehow causing spark and conflict with the circuit on the nixie tube. I love that you were trying to focus on these oversize BEAUTIFUL tubes. Great work hard! Thumbs up! Cheers!
I’m wondering if the H tube will also be sold separately like the R tube. I’m already looking forward to making a huge clock…
А вы не пробовали кроме неона работу ламп с другими инертными газами ? Например аргоном.
Ultrasonic weld and you don't need solder. Could try spot welding too.
Switchmode constant current power supply is easy to make for electrode ageing.
Plasma etching with the right gas will remove oxide. Can make a simple RF induction plasma driver.
These videos must be an enormous amount of work atop your normal work, but this viewer definitely appreciates it. 👍️ I specifically appreciate the detailed breakdowns you give of the operational and failure modes. That helps this Nixie neophyte to follow what's going on.
Beautiful!
two things to consider for the T joint. longer jig to help center the initial bubble by holding where you have it and an extension to the end to guarantee position of the center and repeatability. The second is that to copper ( or other metal ) plate anything you need a very dangerous chemical in the metal salt solution to make even coats. I do not know where you would be able to get cyanide, it seems to act like a catalyst for electroplating, you do not need much, roughly 1 gram per 500 grams ( five hundred ) of the metal salt ( as measured dry ), more cyanide for some metals. The solution will warm up due to the current in the solution so using a fume hood is always recommended. Great work, keep going!
Fascinating and beautiful, as always, thank you. I love reading the comments, especially from people suggesting solutions - it’s not something I know anything about so I’ve no idea how viable the suggestions are, but it’s great to read supportive comments from people trying to help.
is there no way to use like plasma cleaning after the sealing?
Yes, this is a backup solution, we have already a plasma setup for it.. We first want to try the way of preventing the oxide buildup..
Those tubes are awesome by the way
Unrelated to the issues, but for the design of digits on the big tubes, it would be really neat to see a design of digits that have a fake isometric "3D" effect on the design of the numbers.
Or maybe just some more "modern" style fonts, stuff you would have never seen in a nixie tube. Designer style hehe.
Modern circuitry in old reliable tech?
Making a comeback happen!!!
👍👍👏
Cool
Chào ⭐⭐
RIP No 8 wire :(
Fascinating!
It's so much intricate work on these beautiful clocks that you make. It's my dream to own one of these. Wish you lots of success and fame!
Good to see people reaching out to you for further production.. and required methods
у меня есть одна никси трубка ссср
I've never heard of a vacuum relay before... neat!
I have no words how beautiful all this is! For every admiration!!!
So cool! I love watching the progress on the H tubes.
Maybe you're ramping the voltages up too fast? some of these things need to have the voltage applied slowly until it reaches the right amount. Unfortunately that would mean extending the burn time you'll need to have then run.
No, slowing down the voltage won't change anything. The arc discharge is a result of the high voltage applied to the conductors in a high vacuum. Obviously, no arc will occur during the slow voltage ramp. As I said, this is not the source of the problem.
WOW !
42
Hey man! If you're looking for a solution to slow down the current spike use an inductor in series. Inductors don't like the current to be rapidly changed through them and do anything to smooth it. In addition to the inductor use what's called RCD snubber to protect your transistors from burning out
Hi! Inductor in series is what Radim implemented in the end, worked :-)
@@daliborfarny Yeah! I've got to that point in Video! Awesome!
Excellent work !!
Beautiful work
👏
Hats off for you and your team .... The more videos I see on your channel , the more hopeless I feel ....
Why hopeless?
@@daliborfarny Because you could achieve very high levels in success and perfection by simple budget hi-tech engineering work than me.. Your progress results made my work useless and trash . I couldn't reach 1/10 th of what you did so far even by small budget biz ...
@@ahmedalshalchi success is relative, I am not talking about all the million frustrating problems on our side (only the technical ones) - we all have them, no worries!
First