This is my first post and I am quite new to electronics. I have a couple of projects that I am working on, but this one could use a second (or many) set of eyes. I put the actual questions in bold, don't know if that helps.

Premise: Create a circuit for the display of numeric data ranging from 1 to 9,999,999 sourced from a REST web service.

Details. For clarity I will say that my company processes x amount of widgets per day. I would like to use nixie tubes to display a running tally of the day's work. The web-service is firmly in my wheelhouse so that part along with the appropriate data caching is not an issue.

What is an issue is the PCB layout. I have not yet decided which microcontoller I plan on using. My options are propeller, arduino, and raspberry pi. I think I can rule out the prop just based on ease of use. The arduino is easy to use, but wireless connectivity will have to be added in. The pi seems like the easy choice, but I am open to other suggestions.

Power: I have a separate power supply for the tubes providing 170V from 9V input. I will also need 5V for the drivers and shift registers which I plan to pull off the microcontroller. Since the drivers will be using the 5V and Gnd from the controller, but also need the Gnd from the 170V I will just tie the two grounds together. Do I need to do anything extra here for safety? I realize that we are talking about a ground, but I still want to make sure.

As far as the rest goes I plan on using (10) IN-12B nixie tubes (socketed) with (10) K155D1 drivers. Then use (5) 74HC595 shift registers daisy chained together for the numeric control. I am going to cheat and pretend that the decimal point indicator (DP elsewhere) is a comma and use it on the 1st,5th and 8th digit as a thousands separator. For this I plan on using an MPSA42 transistor. If my understanding of transistors is correct (which is shaky at best); I connect the collector of the transistor to ground and the emmiter to the DP pin (cathode; anode is always connected). Then if I supply +5V to the base the DP will turn on. What I absolutely guessed at is which transistor to use. I looked at a bunch of schematics for other projects and found the MPSA42 a common choice. Is the MPSA42 a suitable transistor and do I have the collector and emitters in the proper direction.

I used 22K resistors before each of the IN-12 anodes as those were the most commonly used, however, I am not so sure that is right for this case. Looking at the datasheet and spending a whole lot of time translating I find the following:


current display
digits = <2 mA
point = <0.3 mA

Current Operating with power from a DC voltage source
digits = 2 - 3.5 mA
point = 0.3-0.7 mA

Current Operating with power from the source of ripple voltage (average value)
digits 1 = 2 mA
point = 0.15-0.2 mA

Does this mean that 2-3.5 mA when the digit is turned on and then <2 mA to keep it lit?

If so wouldn't that be a 56K resistor? Or, keep a 22K for ripple (which apparently can be 50%)? [~7.7 mA]

Finally, how do I protect the DP since it cannot handle that amount of current (and it uses the same anode)?

I have attached a PDF file exported from DipTrace. I was going to post the actual DipTrace schematic but it looks like that is not allowed.

Sorry for the long read, I wanted to try and give as much information as I could. I have probably still left out details, please let me know if I did so. Any thoughts on the design would be appreciated.

Many thanks in advance.

Matthew

 

Welcome to AAC!

Do I need to do anything extra here for safety? I realize that we are talking about a ground, but I still want to make sure.

I'd current limit the 170V supply to some reasonable value; though even a few mA is sufficient to hurt you.

Is the MPSA42 a suitable transistor and do I have the collector and emitters in the proper direction.

Polarity is correct, but you should add base resistors to limit drive current and should insert a series resistor to protect the DP from overcurrent.

I used 22K resistors before each of the IN-12 anodes as those were the most commonly used, however, I am not so sure that is right for this case.

The proper resistance depends on the tube you're using and the brightness you desire.

You can post images. Relevant part of the schematic:

 

You have the MPSA42 connected OK just the emitter/collector Description is backwards.
Max.

 

Ok, I have a very simple schematic that just puts focus on the transistor. In this example the numeral "1" is always lit and a simple switch turns on and off the decimal point. I added a 340K resistor between the tube and the collector to limit current to the decimal point.

I don't think that is quite correct. I haven't figured out the specifics of calculating the currents in a transistor. It is more complex than I originally thought (most things are). Here is what I have so far.

 

In this example the numeral "1" is always lit and a simple switch turns on and off the decimal point. I added a 340K resistor between the tube and the collector to limit current to the decimal point.

Do you mean '0' is always on?

You still haven't inserted a resistor on the base of the transistor to limit base current. As drawn, the transistor will be destroyed as soon as you apply 5V to the base.

The 340kΩ resistor is probably too large. Without more information I can't offer a better value. It's still unclear that 22kΩ is adequate for the digits.

 

i dont think 22K is sufficiently large. A back of the envelope calculation shows that 170V through a 22K resistor is a current of 7.8mA. Higher than the max digit current.

 

I have re-drawn the simplified schematic with a couple of changes the 22K resistor has been changed to 56K which should limit the current to around 3 mA which is within range. I have added another resistor between the transistor and ground with a value of 10K.


I derived this value from the following:


I eliminated the R1 resistor as I didn't think I wasn't sure if I needed it. Using only R2 the current to the tube is .25 mA which is in range. Still not 100% sure, but I feel I am getting closer.

Edit: I also tried to make the "switch" more realistic. It will be an output pin from a Raspberry Pi. I known there is another ground in there, but I wasn't sure how to draw it.

 

I hate to double-post, but I was wondering if I got the calculations above correct? I am concerned about omitting R1.

Thanks!

 

I was wondering if I got the calculations above correct?

With what you've drawn, you'll get 0.3mA (260uA) through the DP.

Current is limited by the voltage on the base of the transistor and R2. When the transistor is on, it should saturate and the collector voltage will be about 2.6V.

 

Dennis,

Thank you for the reply. I see I am getting a little closer, but obviously I am still off somewhere. I was estimating 250 uA on the DP side. I will go over the formulas again to see which one isn't correct.

Matthew

 

I was estimating 250 uA on the DP side.

250uA, 260uA, 0.3mA are essentially the same value. The BE junction voltage isn't a specific value.

The transistor will start conducting at around 0.6V of forward bias and junction voltage will increase to something around 0.7-0.8V when it's saturated. I use 0.7V for "on", some use 0.6V for "starting to turn on" in calculations.

 

I agree that in this case the values of 250 uA and 260 uA are virtually identical. I am just a little on the OCD side and want to understand how I got the calculations reversed. For this use case I am all set.

Thank you again.

 

I am just a little on the OCD side and want to understand how I got the calculations reversed.

Your calculations weren't reversed. Your assumptions were incorrect.

You're operating the transistor as a switch, so it's either cutoff or saturated. When saturated, the voltage on the collector should be near the voltage on the emitter.

If the transistor was in it's linear (active) region, the voltage on the collector would depend on transistor beta, VCC, and all of the elements on the collector. In your case, you have a resistor and the nixie tube. I don't know how to determine the equivalent resistance of the nixie tube before and after it glows. From what I've seen, manufacturers just give you typical characteristics in graphs and you just use the circuit parameters they specify.

 

I promise I am not intentionally being thick, I guess I just don't quite understand. I am going to try and write back what I think I understand. I really do appreciate your time.

From the datasheet I see that the Base-Emitter Saturation Voltage is 0.9Vdc. So this is saying that anything over 0.9V will saturate the Base-Emitter region and "connect" the Emitter and Collector? This connection has nearly zero resistance from what I read on a different internet source. So in my case I would effectively have

+170 ------- (nixie)------/\/\/\/\/------Gnd (As long as I am applying more than 0.9V to the base.)

So my resistor should be calculated solely on that circuit?

As far as protecting the base from over current, I don't see a max current for the base in the datasheet.

 

From the datasheet I see that the Base-Emitter Saturation Voltage is 0.9Vdc. So this is saying that anything over 0.9V will saturate the Base-Emitter region and "connect" the Emitter and Collector?

For a transistor to be saturated, both junctions need to be forward biased. The BE junction voltage will be determined by base current, not whether the transistor is saturated.

This connection has nearly zero resistance from what I read on a different internet source.

You have to filter what you read on the WEB with what you know. If you don't have knowledge in what you're reading, you can't filter good information from bad information.

So in my case I would effectively have

+170 ------- (nixie)------/\/\/\/\/------Gnd (As long as I am applying more than 0.9V to the base.)

So my resistor should be calculated solely on that circuit?

I don't understand what you're asking. How can you calculate current without knowing the equivalent resistance of the plasma in the nixie tube?

As far as protecting the base from over current, I don't see a max current for the base in the datasheet.

Some datasheets have more information than others. Since a maximum base current isn't given, you could assume that the transistor will be okay as long as you don't exceed the maximum collector current. If you find that BE junctions fail at a lower current, you learned something and can reduce base current accordingly. BTW, good designs usually derates devices by 25-50%.

 

You have to filter what you read on the WEB with what you know. If you don't have knowledge in what you're reading, you can't filter good information from bad information.

Isn't this the purpose of forums like these? I don't know; so I am asking, if I had the ability to tell which sources were providing accurate information I wouldn't need to ask the question. The only way this works is I have to trust that the source is correct. On the subject of the resistance between the collector and emitter, I wasn't sure so I stated I read it elsewhere. I still have no idea if that is correct or not..

I don't understand what you're asking. How can you calculate current without knowing the equivalent resistance of the plasma in the nixie tube?

You said you didn't understand the question, but you still answered it. The answer was No, you have to take the resistance of the nixie tube into account. I did not realize that the resistance of the nixie tube would make a significant difference. That may have been obvious to all but me, but I didn't know that.

Now that I know I need to find the resistance of the nixie, can I just measure it directly with a dmm? If so, how best to measure? Should I measure the resistance directly across the anode and cathode, or, should I calculate the resistance based on the voltage/current?

 

Isn't this the purpose of forums like these? I don't know; so I am asking, if I had the ability to tell which sources were providing accurate information I wouldn't need to ask the question.

A forum like this can be a better source of accurate information because it's peer reviewed. That is, assuming there's actually any peer review.

I've seen crap schematics and advice on the WEB, some in forums. YouTube is a source of a lot of garbage; nothing is peer reviewed.

Now that I know I need to find the resistance of the nixie, can I just measure it directly with a dmm? If so, how best to measure? Should I measure the resistance directly across the anode and cathode, or, should I calculate the resistance based on the voltage/current?

It's not that straightforward with nixie tubes. There are several "resistances". There's one when it isn't conducting, one when it starts to flash, and one when it's conducting steady state. That's why manufacturers tell you what voltage and resistance to use. They designed the tube and they know how best to use it.

 

I have scoured the web (again) and found no documents from the manufacturer (russian gov. as far as I can tell) that give resistance values or any type of data past the current maximums.

Let's take this circuit simplified even further. The "DP" will just be "ON" and no other digit.


The total resistance of the loop containing the nixie tube would be less than .5 mA. This is based on 170 V and a resistance of 340K + an unknown resistance in the nixie. Since the resistance of the nixie couldn't be negative the overall current could only be less than .5mA.

Is that assumption correct?

If the max current for the DP is .7 mA and the resistance is zero (which I know is not likely) .5 mA is around 70% of the maximum. The resistance of the nixie tube would also be different for individual digits; I would expect the "3" to have a higher resistance than the "1". All of these factors could only make the current go down. I would get a dimmer "DP" but the current to the tube would never go higher than .5 mA.

I suppose another way to go would be to use a 270K resistor and a second trim pot of 100 K in series. This would give me some adjustment without have to re-solder anything. Even if the trim pot was turned down as low as it could the overall resistance would be less than the stated maximum, right?

 

I decided to re-draw with the trim pots.


With the 10K trim pot turned all the way down the current to the digits could not be more than about 3 mA still within the range. All the way up about 2.6 mA. For the DP; both trimmers turned all the way down the max current would be about .69 mA just inside the max. Using two trimmers I could dial in the best resistance for both the DP and the digits. I still would have to contend with each digit having a slightly different resistance, but I would suspect that the difference would not be that noticeable.

 

but I would suspect that the difference would not be that noticeable.

Black on white schematics are easier to read...


What are you trying to achieve?

The way you've connected the trimmers, they can't adjust current.

You have the transistor wired as a current sink. The current will be, at most,
\( \small I = \frac{V}{R} = \frac{3.3V - 0.7V}{290k ohms} = 9.0uA\)