I really need a hand here. I am trying to control an IN-19B nixie tube with an Arduino UNO. I am using a 170v power supply with a 20K resistor on the anode and the k155id1 driver IC. I have got things working fine (cycling through characters) with the small problem that after about 10-60 seconds I will hear a small click/pop noise and multiple characters on the nixie light up, some still cycling but with other digits that are turned on obscuring them. I assume this noise is the nixie driver blowing and causing some cathodes to short to ground. I have checked the voltage I am supplying the chip with from my arduino (5V) and the placement of all jumper cables and everything seems to be wired correctly. I've also tried running the chip at a lower 3.3v in case I was somehow killing it overvolting it or something. (I know this is a bit stupid because the IC definitely runs at 5v but I was desperate at this point)

I would be eternally grateful if someone could shed some light on this situation. The graveyard of drivers on my desk is growing :(. I have contacted the eBay seller that supplied me with the drivers and they are equally confused. Below you should find the schematic I am using, relevant data sheets and an image of my setup. (I am aware the name of the driver on the schematic and in my circuit are different but from what I saw online the name is the only difference between the two)

I am hoping to use theses tubes in a project for my art A Level so you really would be saving my life if you helped. Thank you. :)

Circuit Schematic

Driver data sheet

Nixie data sheet

Overall setup view[![][4]]5


The problem appears to be solved after re positioning and replacing the resistor on the breadboard all is functioning correctly.

Repositioned resistor

efully functional nixie!

Thank you everyone for your help and very well spotted. I'll make sure to learn from mistakes like this and hope this informs others that see this.

Update 2:

Slightly worrying development. The chip I tried after the fix just did the same thing. This time it took an hour instead of 10 seconds which is definitely an improvement but still sub optimal.

Update 3:

Through sheer incompetence I misunderstood the instructions about the resistor. I believe this should work. Really sorry about that everyone. I remember thinking the pink tinge to the cathode was odd. Assumed it was just variation in tubes but with the colour difference after re positioning the resistor it was obviously a cry for help. Thank you everyone again apologies for my incompetence I will also make sure to learn from this.

Fix after misinterpritation

  • 1
    \$\begingroup\$ You should always add decoupling capacitors across the power rails of all your ICs as close as possible to the IC. Rule of thumb is 10 to 100nF. Use ceramics capacitors. This may not be your actual issue, but it does prevent a bunch of unpredictable issues from happening. \$\endgroup\$ – DKNguyen Nov 28 '19 at 18:06
  • \$\begingroup\$ Thanks for the tip, I will try adding these to the circuit tomorrow and test the results. \$\endgroup\$ – Callum Cooper Nov 28 '19 at 19:47
  • 3
    \$\begingroup\$ The resistor shown in the photo afer "Update" does nothing - it is shorted by the power bus on the edge of the breadboard. \$\endgroup\$ – Peter Bennett Nov 28 '19 at 20:49
  • \$\begingroup\$ Oh damn I get what you mean now. Sorry inexperience showing through again. Will post another image showing what I now think you mean. \$\endgroup\$ – Callum Cooper Nov 28 '19 at 21:11
  • \$\begingroup\$ Your latest fix is correct, does it work now? (btw the pink tinge is probably caused by mercury vapour in your bulb. A lot of tubes have it, as it enhances their longevity. Take care not to break it though!) \$\endgroup\$ – anrieff Nov 28 '19 at 21:32

This is a common mistake, and almost invariably for this type of fault you are allowing an illegal state on the decoder chip inputs. There is NOTHING wrong with using 170-180V for your supply but you must ABSOLUTELY ENSURE that there is always one digit in the 'ON' condition. These Nixie tubes were designed in an era when small scale logic (counters) were used. The counter designs typically did not support illegal states, so the issue was moot. With the use of MCU's proliferating today, the issue of programmed states has to be considered.

The datasheet for the 74141 shows the problem clearly:

enter image description here

DO NOT reduce your supply to 55V + strike voltage as you will have problems turning on your digits. You need to have about 20-50V across your 20k Ohm resistor to allow for. Also be aware that the strike voltage rises when the tube is cold.

Since you are driving your decoder from a microprocessor the port pins you are using are set to inputs when the MCU first starts up. the inputs to the decoder will therefore be high and ALL digits will be off. During this time you are imposing a severe over-voltage on the driver outputs. Current will flow through the zener diodes and the weakest zener will pop.....then the next....then the next and so on. The zeners fail short circuit mostly hence the digits all start to come on.

The easiest solution is to put an output pull-down resistor on your drive pins to the decoder to handle the power up situation. You have to program your MCU to never use the illegal states....ie always have one digit enabled, and don't spend time between digits in illegal states.

If you do need digits to be OFF (leading zero suppression) then you need to clamp the anode voltage for that digit.

Update: To answer a question posed, you only need to pull down one input on the decoder to ensure that you always have legal states on the input. The current requirements for A,B,C,D vary with A being the least current. Since a valid low is <= 0.8V this can be achieved by putting a 470 Ohm resister from A to ground. If the port pin is set to an input then A=0 so digit '0' will be on. When the MCU is driving correctly it will have to source 11mA which is well withing rating. You still need to ensure that when program your port pins and they are set to outputs that you handle the values to ensure you don't get an illegal state.

  • \$\begingroup\$ "don't spend time between digits in illegal states." How far do you have to go to achieve this? (i.e. do you have to go so far as to write to the entire GPIO register in a single write? instead of manipulating the GPIO pin by pin atomically?) \$\endgroup\$ – DKNguyen Nov 28 '19 at 19:32
  • \$\begingroup\$ Hmm, this passage is not correct: "There is NOTHING wrong with using 170-180V for your supply but you must ABSOLUTELY ENSURE that there is always one digit in the 'ON' condition.". I have a Nixie clock built myself, uses 74141, and it's "display=off" state is implemented by sending state 15 to all decoders. They have been surviving this for years. \$\endgroup\$ – anrieff Nov 28 '19 at 19:53
  • \$\begingroup\$ Would posting the code I am using help here? If I was to attempt this solution what value of pull down resistor would you recommend? I'm also not completely clear where exactly the resistor needs to be placed but this is likely my lack of experience showing through. I really appreciate your help. \$\endgroup\$ – Callum Cooper Nov 28 '19 at 19:57
  • \$\begingroup\$ @anrieff Not sure what you are saying since you don't explain what you anode resistance or high voltage is. I reiterate though ....there is absolutely nothing wrong with having 170-180V high voltage if the design accommodates this. We have no way to jusdge what you design is. \$\endgroup\$ – Jack Creasey Nov 28 '19 at 19:59
  • 1
    \$\begingroup\$ @DKNguyen The programming is relatively safe once you have at least one pulldown resistor (initial powerup is the real hazard). Since the digit you are about to turn of is conducting, it will continue to conduct current into the zener for a short time. In the meantime you program another digit to be on (micro seconds). If you take too long in the illegal states then you certainly can pop an output zener, but I've never had problems with sequential writes where the delay is only a few microseconds at most. \$\endgroup\$ – Jack Creasey Nov 28 '19 at 20:25

It is hard to tell from the photos, but it looks to me like you may have shorted out the 20K resistor on the breadboard, which will yield the kind of unfortunate results that you are seeing.

  • 1
    \$\begingroup\$ Good catch! As you say - hard to see but well worth looking at. \$\endgroup\$ – Dwayne Reid Nov 28 '19 at 18:42
  • \$\begingroup\$ Very well spotted. I have isolated it and tried again but had the same result unfortunately. I really appreciate the idea though. \$\endgroup\$ – Callum Cooper Nov 28 '19 at 19:50
  • 1
    \$\begingroup\$ This is quite possibly the answer. @CallumCooper, please update your question with new photos after the fix suggested here. \$\endgroup\$ – anrieff Nov 28 '19 at 19:55
  • 1
    \$\begingroup\$ It looks like I am eating my words here. After putting in a fresh chip and resistor (Last time just fresh chip) the circuit has worked for the last 3 minutes flawlessly. A new record! Great solve thank you so much. \$\endgroup\$ – Callum Cooper Nov 28 '19 at 20:11

If I recall correctly, the SN74141 is rated to handle a maximum of 60 Vdc on the output pins.

enter image description here

Find out (measure) what the voltage drop across an ionized digit is and add 55 to that value. Hopefully, that voltage is higher than minimum ignition (firing) voltage of your Nixie tube. Adjust your high-voltage supply to be somewhat less than that value.

You will then most likely need to adjust the value of the current-limit resistor to achieve the desired current.

I would guess that there are numerous articles available via Google that show how to use that chip to drive the Nixie tubes that are currently available.

  • \$\begingroup\$ This is very interesting thank you. Another answer has suggested this may cause difficulty making digits light up but if other solutions don't work this sounds like a good option. Thanks again :) \$\endgroup\$ – Callum Cooper Nov 28 '19 at 19:52
  • \$\begingroup\$ > "If I recall correctly, the SN74141 is rated to handle a maximum of 55 Vdc on the output pins. " No, this is not the case. It can handle 160 volts. \$\endgroup\$ – anrieff Nov 28 '19 at 19:53
  • \$\begingroup\$ @anrieff Absolutely NOT. The zeners on the output of the 74141 are about 60-65V. Look at the datasheet. neon1.net/nixieclock/sn74141.pdf \$\endgroup\$ – Jack Creasey Nov 28 '19 at 20:02
  • \$\begingroup\$ @JackCreasey, where do you get that values from the datasheet? The "Off-state output voltage" parameter? \$\endgroup\$ – anrieff Nov 28 '19 at 20:37
  • \$\begingroup\$ @anrieff I added it to Dwayne's answer......direct form the TI datasheet. \$\endgroup\$ – Jack Creasey Nov 28 '19 at 20:43

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.