#3. Use a current regulator, and set the current at the recommended spec value; too high and you wear-out the tube. Too low and you risk cathode poisoning. A single anode resistor per-tube is OK, but as the anode voltage varies, so does the current (no such problem with a current regulator, though). You can mitigate this by using a higher anode supply voltage and larger anode resistor, at the expense of more wasted energy. As tubes age, their striking voltage may increase, so having a higher anode voltage will help mitigate this. *Understood, is this implemented a current regulator per tube or is this a single current regulator? If single current regulator, how do you account for possibly having varying current with different digit activation? or am I missing an operating principle here?*
*Single regulator per-tube, on the anode. Except for the RZ568m, I've only seen tubes spec the same current for each numeral. Generally I use an HV5530 to drive the cathodes, so only the anode-side can be current-regulated. I use a PMOS transistor with a resistor between source and the + anode supply, drain goes to the nixie-tube anode, and the gate goes to a trimpot to adjust the gate voltage (about 8 volts negative w/ respect to the anode supply). Assuming the Vgs(on) of the PMOS transistor is constant, the current is limited by the voltage-drop across the resistor.* *With my B7971 clock, I have a current-regulator on the anode, as well as for each cathode because the datasheet sets different currents for the various segments. That works out to 128 current regulators for the 8-tube clock, but given the ridiculous cost of a replacement tube it's worth the expense.* #4. PIR sensor. Turn off the tubes if nobody is there to watch them. *This will definitely be part of the overall clock design (also incorporating a Noritake Itron 20 character 5x7 VFD). I plan on having a smoked plexiglass case housing the electronics and VFD while mounting the nixie tubes on top. Which leads on to our next point...* *Be sure the PIR sensor isn't covered by anything; even clear-plastic will block infrared.* #5. Protect the tubes inside a case, and make sure there is enough ventilation so the heat doesn't build-up inside. On my later designs I have a thermal sensor (sometimes several) so that software can monitor critical temperatures and shutdown if things get too warm. *If the tubes are always going to be high up in the air (on a dedicated wall shelf), is casing them necessary? Could not casing them cause convection heat to rise around them and risk stressing the glass? The thermal sensor is a good idea, for the inside of my case I will certainly implement this, probably in multiple spots as you suggested. Do people ever put thermocouples on the tubes themselves to monitor heat or could you see value in it? I might consider that in my design.* *You never know when some flying object might hit a tube. Putting holes in the bottom of the case and on the top (or on the back-side at the top) will usually get enough convection. I have 1 clock with a fan that can be turned-on by software if one of the thermal sensors is out-of-range, but it hasn't been triggering. Nixies dont get hot, but power transformers and heatsinks can so I JB-weld an I2C thermal sensor on the critical ones and have software monitor the temperature, and shutdown or turn-on a fan if needed. If you have a DS3231, it has a built-in thermal sensor. I also have one in the Raspberry Pi CPU. The rest are on an I2C chain.* #6. Depoisoning routine to exercise all cathodes. *This is not the first time I have read of depoisoning (a common feature on most bonafide driver boards I have seen). I have also seen there are different marketed techniques for depoisoning: slot machine, strobe, etc...* *What are the principles behind cathode poisoning and the theory of prevention? If this has already been documented somewhere please give me strength on my journey you send me on.* *Basically, unused cathodes can accumulate debris over time if they are not used, ie they get poisoned. My IN-18 clock cycles digits thru their unused numerals for 1 hour per night. Tubes that glow 0-9 evenly, such as the units-seconds and unit minutes, are simply shutoff for depoisoning. Others that go 0-5 all day are cycled 6-9. Lastly, the ones that are static (day, month, year) are cycled 0-9.* *My B7971 clock does a segment walk whenever it's not displaying the time (no PIR activity). Each segment (15 per tube) runs about 200msec, then repeats that on each of the 8 tubes sequentially. It works out to about a 1% duty cycle.* On Wednesday, September 18, 2024 at 1:33:00 PM UTC-7 Chachi88 wrote: > First off, thank you guys very much for your input here, I will address > them most recent to oldest. > > Peter, > I apologize I cannot view this link I do not use facebook. I also cannot > find a link to Richard Scales B-8091 clock PCB that you mention. > > Nicholas Stock, > Thanks a lot for showing me this website, I had come across it also, but > was discouraged by the "sold out" notification for this board. Is there > any way I can message the seller, you think? I really like all the > features this board has already implemented for me...but maybe greg's first > comment is calling my name... > > gregebert, > Thanks for the very informative points, I have addressed my responses and > followup questions individually below in *Italics.* > > #1. I dont like sockets because they put stress on the tube pins. Some of > them grip the pins very hard and it's difficult to insert or remove the > tubes, which adds to the risk of bending or breaking the pins or putting > stress on the glass. Instead, I use socket pins soldered into the PCB, and > the force to insert/remove tubes is very low. > *Point taken, I will look at my sockets carefully before considering using > them. Otherwise I will probably borrow your idea, cheers.* > > #2. Direct drive. There's no need to multiplex individual tubes, and > there's no cost benefit when tubes are worth hundreds of dollars apiece and > the drive electronics is at most a few dollars. Multiplexing requires > higher current, and that degrades the tube's lifetime. > *This is what I have understood as well, thanks for confirming.* > > #3. Use a current regulator, and set the current at the recommended spec > value; too high and you wear-out the tube. Too low and you risk cathode > poisoning. A single anode resistor per-tube is OK, but as the anode voltage > varies, so does the current (no such problem with a current regulator, > though). You can mitigate this by using a higher anode supply voltage and > larger anode resistor, at the expense of more wasted energy. As tubes age, > their striking voltage may increase, so having a higher anode voltage will > help mitigate this. > *Understood, is this implemented a current regulator per tube or is this a > single current regulator? If single current regulator, how do you account > for possibly having varying current with different digit activation? or am > I missing an operating principle here?* > > #4. PIR sensor. Turn off the tubes if nobody is there to watch them. > *This will definitely be part of the overall clock design (also > incorporating a Noritake Itron 20 character 5x7 VFD). I plan on having a > smoked plexiglass case housing the electronics and VFD while mounting the > nixie tubes on top. Which leads on to our next point...* > > #5. Protect the tubes inside a case, and make sure there is enough > ventilation so the heat doesn't build-up inside. On my later designs I have > a thermal sensor (sometimes several) so that software can monitor critical > temperatures and shutdown if things get too warm. > *If the tubes are always going to be high up in the air (on a dedicated > wall shelf), is casing them necessary? Could not casing them cause > convection heat to rise around them and risk stressing the glass? The > thermal sensor is a good idea, for the inside of my case I will certainly > implement this, probably in multiple spots as you suggested. Do people > ever put thermocouples on the tubes themselves to monitor heat or could you > see value in it? I might consider that in my design.* > > #6. Depoisoning routine to exercise all cathodes. > *This is not the first time I have read of depoisoning (a common feature > on most bonafide driver boards I have seen). I have also seen there are > different marketed techniques for depoisoning: slot machine, strobe, etc...* > *What are the principles behind cathode poisoning and the theory of > prevention? If this has already been documented somewhere please give me > strength on my journey you send me on.* > > #7. Backlighting (or base lighting). I havn't done this on any of my > clocks, and I've never had problems with tubes not firing-up. Others have > reported problems without backlighting. It can have aesthetic value, though > I'm not fond of it. > *I tend to agree with you, for me, this was considered aesthetically not > necessary, but if there is value in ensuring operation, I will plan on > incorporating this in the design even if it is left partially implemented > in hardware or neglected in software...* > > That should cover the necessities for the tubes; there are always lots of > other features that can be done, especially if the clock has open-source > software (or you develop it on your own). > > *Thank you again sincerely for expounding this much on this, I in the end, > I am hoping to find a clock/driver board with open source software or just > raw signal input pins exposed that way I can incorporate it in with my VFD > with minimal additional effort expounded on the nixie driver block (because > it has already been done probably way better than I would do it on a first > pass). I can then address it, animate it, etc. with a microcontroller of > my own...* > > > On Wednesday, September 18, 2024 at 11:05:11 AM UTC-4 Nicholas Stock wrote: > >> Pete also has a remote driver system that will work with B8091s... >> >> https://www.pvelectronics.co.uk/index.php?main_page=index&cPath=43 >> >> Nick >> >> Sent from my iPhone >> >> On Sep 18, 2024, at 07:16, Peter Doroba <[email protected]> wrote: >> >> >> >> This is the clock I made using Richard Scales B-8091 clock PCB. >> >> https://www.facebook.com/groups/142414439207058/search/?q=b-8091 >> >> On Wednesday, September 18, 2024 at 12:41:14 AM UTC-4 gregebert wrote: >> >>> Those are very valuable tubes, so if you are willing to design your own >>> driver and the PC board, I would do that. I've made several clocks and >>> never had a design problem that affected the tubes. Even if you find a >>> board to purchase, see if any of the following apply: >>> >>> Some of my suggestions: >>> #1. I dont like sockets because they put stress on the tube pins. Some >>> of them grip the pins very hard and it's difficult to insert or remove the >>> tubes, which adds to the risk of bending or breaking the pins or putting >>> stress on the glass. Instead, I use socket pins soldered into the PCB, and >>> the force to insert/remove tubes is very low. >>> >>> #2. Direct drive. There's no need to multiplex individual tubes, and >>> there's no cost benefit when tubes are worth hundreds of dollars apiece and >>> the drive electronics is at most a few dollars. Multiplexing requires >>> higher current, and that degrades the tube's lifetime. >>> >>> #3. Use a current regulator, and set the current at the recommended spec >>> value; too high and you wear-out the tube. Too low and you risk cathode >>> poisoning. A single anode resistor per-tube is OK, but as the anode voltage >>> varies, so does the current (no such problem with a current regulator, >>> though). You can mitigate this by using a higher anode supply voltage and >>> larger anode resistor, at the expense of more wasted energy. As tubes age, >>> their striking voltage may increase, so having a higher anode voltage will >>> help mitigate this. >>> >>> #4. PIR sensor. Turn off the tubes if nobody is there to watch them. >>> >>> #5. Protect the tubes inside a case, and make sure there is enough >>> ventilation so the heat doesn't build-up inside. On my later designs I have >>> a thermal sensor (sometimes several) so that software can monitor critical >>> temperatures and shutdown if things get too warm. >>> >>> #6. Depoisoning routine to exercise all cathodes. >>> >>> #7. Backlighting (or base lighting). I havn't done this on any of my >>> clocks, and I've never had problems with tubes not firing-up. Others have >>> reported problems without backlighting. It can have aesthetic value, though >>> I'm not fond of it. >>> >>> That should cover the necessities for the tubes; there are always lots >>> of other features that can be done, especially if the clock has open-source >>> software (or you develop it on your own). >>> >>> On Tuesday, September 17, 2024 at 8:17:28 PM UTC-7 Chachi88 wrote: >>> >>>> I am climbing stairs from the foot of the mountain to sit at the feet >>>> of all you nixie gurus. >>>> >>>> What is the current state of the art for remote driver boards? I am >>>> aware some of the older drivers are not as reliable or can cause the tubes >>>> to degrade quicker. Some of the new options I am seeing have "cathode >>>> poisoning prevention" I have recently come upon qty 6 of the NL8091's and >>>> their original sockets and socket mounting plate, which I wish to reuse >>>> for >>>> a clock. Are there any off the shelf boards even capable of driving this >>>> tube? I was looking at a board on ebay that comes with an IR remote but >>>> it >>>> seems too good to be true, I would be willing to spend 5 times as much for >>>> a board that was properly documented and put together and had some sort of >>>> pedigree... >>>> >>>> Thanks in advance for any advice or direction you can give. >>>> >>> -- >> You received this message because you are subscribed to the Google Groups >> "neonixie-l" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to [email protected]. >> To view this discussion on the web, visit >> https://groups.google.com/d/msgid/neonixie-l/bac36066-03e6-4a6e-a83a-69020b75572fn%40googlegroups.com >> >> <https://groups.google.com/d/msgid/neonixie-l/bac36066-03e6-4a6e-a83a-69020b75572fn%40googlegroups.com?utm_medium=email&utm_source=footer> >> . >> >> -- You received this message because you are subscribed to the Google Groups "neonixie-l" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web, visit https://groups.google.com/d/msgid/neonixie-l/1e22d2df-53dd-4663-817a-221eb28cc459n%40googlegroups.com.
