Design a nixie multimeter

[romhunter]

I have a lot of USSR IN-12B nixie tube on hand. They look cool, and I think they are quite fit to be use in a retro looking benchmeter. So, I went out and research a bit. There seems to be a few options:

1. Use the ubiquitous ICL7107, it has almost everything I need to make it work as a voltmeter. I will just then grab the schematic of some multimeter that use it, add some glue logic (or GAL chip) to convert the 7 segment output to nixie-compatible output, then drive the digit using high voltage transistor like MPSA42.

2. Modify the schematic of some similar meter that use nixie (like the one I own - Heathkit IM-102). It has the advantage of readily-working circuit design, but I have look through the schematic once and it is quite the pain in the ass

3. Use a microcontroller

So, any thought? I really dislike the idea of a microcontroller though, just because microcontroller is not era approriate for such project. GAL seems to be a pain too, but I'm trying to use them right now

[danadak]

I would lean on the uC. You can get a 20 bit A/D in PSPC 5LP,
a digital filter handle noise and signal averaging mode,
tons of I/O and logic fabric (that you can customize) to manage
it.



This shows single ended, simple setting to handle differential. Onboard precision reference.


Cheap board for dev or final solution ($10)





http://www.cypress.com/documentation/development-kitsboards/cy8ckit-059-psoc-5lp-prototyping-kit-onboard-programmer-and


Regards, Dana.


Regards, Dana.

[romhunter]

I would lean on the uC. You can get a 20 bit A/D in PSPC 5LP,
a digital filter handle noise and signal averaging mode,
tons of I/O and logic fabric (that you can customize) to manage
it.

Sure, as much as I hate writing firmware (bugs!), I have to agree this is a better solution. But I just simply don't like it for the authentic feeling.

And to say it again, I'm not good at writing firmware.

[BrianHG]

I would go hybrid.  Take an existing multi-meter, which has an rs232 port, remove it's display and use a small microcontroller to receive the rs-232 and drive the nixies.  Now, I know you said you don't like programming, bur just making a 1$ microcontroler receive rs-232 and send that output to high-voltage line driver for the nixies is not like making an entire volt/resistance/amp, AC/DC front end which you can trust to 1kv DC/750vac which comes with existing meters.  You dont need to worry about board routing the analog properly, and if you use the optocoupled isolation built into many voltmeters, you can sell your design as an upgrade device for others to use who can trust the readout based on what will be inside, a proven voltmeter design, isolated from the nixie supply so your probe measurements will be isolated.  Though, your meter will need both batteries and AC for the nixies to achieve good floating probes.

[Kleinstein]

One may not  need much glue logic for driving the nixi tubes from an ICL7107 or similar. So this may not need an an GAL or similar - just discrete transistors may be all it takes. If multiplexed drivers are wanted one could consider the ICL7135 (4.5 digits, even if only 3 or 4 are used).
 
It only gets complicated if auto-ranging is needed. The Ohms ranges may need quite some switches. The other point to take into account is that such an old style DMM would need analog calibration / adjustment and thus lots of trimmers to do the cal.

The other option would be a small µC and a SD-ADC like the MCP3421. Thus adjustment of the ranges would be in software. This might be the easier way, especially if auto ranging is wanted. As a bonus one could add an PC interface (RS232 or isolated USB). Using µC(s) might also allow for an extra isolation between the input / ADC part and the display.

[Cyberdragon]

A three or four digit multimeter is actually incredibly simple. You could multiplex them or other complicated rubbish. Or you could just do something like this:

http://www.rfcafe.com/references/popular-electronics/build-numeric-glow-tube-dcu-2-1970-popular-electronics.htm

Just replace the three logic chips for each section with decade counter chips. There was a voltohm meter project for it which I can't seem to find. But the meter unit uses only two chips, and that was in 1970! So it's not hard to drive this thing  (the original display can update at a max of 12MHz, so not a speed issue either).

[Pawelr98]

There's 74141 chip for nixie tubes.
AKA K155ID in USSR.

A bunch of TTL chips + uC(receiving data from normal multimeter and controlling the TTL logic) and it's ready to use.

Then you have the ICL7107 which is one-chip digital voltmeter.
Bit of work would be needed to adapt such chip to Nixie display.

If you need precise Nixie voltmeter then search for schematics.
Meratronik V541 - Nixie very-precise voltmeter
Meratronik V543- Nixie multimeter

I can give you schematics for both. They are quite complicated though.

[Chris56000]

Hi!

If you really don't mind "old–skool" then download the booklet for the Advance/Gould DVM5, it's 20,000 count automatic ranging, and if you hunt around a bit, you can get the IN–19A and IN–19B Nixie's for the unit displays and +/–/Sine AC squiggles!

Well worth a look, but I warn you you'd be making it for donkey's years – there's about 40 TTL in it! If you're good at PICs or Arduino, you can replace all the manual switches in the design with press button relay control!

Regarding Nixie Anode voltage, there are a number of nice little Nixie PSUs in kit or Ready–built form for one figure (£10 or less!!), or, if you want to have an "old skool" PSU in your design, you can find several small pre–amp transformers (new!) on fleabay with a 110–150V secondary that can be rectified for the anode voltage!

If all that is too much, look for a "7–Segment to BCD Encoder" schematic and use that on the end of of any of the usual '7107 designs, for an "old skool" mains bench '7107 design you can convert to BCD for Nixie drivers, get the April 1979 and May 1979 issues of "Practical Electronics" off the American Radio History site!

If you want a self–contained battery instrument, you need three isolated O/P DC–DC converters, +9V to +5V, +9V to –5V for the '7107, then +9V to +12V for the Nixie PSU module!

Chris Williams

PS!

Considering the absolutely insane prices wassocks are asking for anything with a nixie–tube in these days, you could.certainly build your own Nixie multimeter for less!

[Cyberdragon]

9V?

Are you suggesting using a PP3? That won't last very long at all.

There are a lot of 5V nixie converters out there (for USB power probably). You can make it USB powered by using one of these modules and something like a LTC1983 inverting charge pump for the negative rail. If you wanted an internal battery you could use the guts from a power bank.

[romhunter]

I would go hybrid.  Take an existing multi-meter, which has an rs232 port, remove it's display and use a small microcontroller to receive the rs-232 and drive the nixies.  Now, I know you said you don't like programming, bur just making a 1$ microcontroler receive rs-232 and send that output to high-voltage line driver for the nixies is not like making an entire volt/resistance/amp, AC/DC front end which you can trust to 1kv DC/750vac which comes with existing meters.  You dont need to worry about board routing the analog properly, and if you use the optocoupled isolation built into many voltmeters, you can sell your design as an upgrade device for others to use who can trust the readout based on what will be inside, a proven voltmeter design, isolated from the nixie supply so your probe measurements will be isolated.  Though, your meter will need both batteries and AC for the nixies to achieve good floating probes.

that would be nice. I'll try to do this some time later this year. But I'm trying to be proficient in analog design anyway, so analog PCB (be it painful or not) is what I must be able to do.

There's 74141 chip for nixie tubes.
AKA K155ID in USSR.

I prefer the K155ID, better than the crap 74141

Then you have the ICL7107 which is one-chip digital voltmeter.
Bit of work would be needed to adapt such chip to Nixie display.

That's my original plan

If you need precise Nixie voltmeter then search for schematics.
Meratronik V541 - Nixie very-precise voltmeter
Meratronik V543- Nixie multimeter

I can give you schematics for both. They are quite complicated though.

Well, if you want to give it to me then I'll pm you my email address. Or if you don't mind publishing it then...

Hi!

If you really don't mind "old–skool" then download the booklet for the Advance/Gould DVM5, it's 20,000 count automatic ranging, and if you hunt around a bit, you can get the IN–19A and IN–19B Nixie's for the unit displays and +/–/Sine AC squiggles!

I still prefer the IN-12B, although I have some IN-19 laying around.

Regarding Nixie Anode voltage, there are a number of nice little Nixie PSUs in kit or Ready–built form for one figure (£10 or less!!), or, if you want to have an "old skool" PSU in your design, you can find several small pre–amp transformers (new!) on fleabay with a 110–150V secondary that can be rectified for the anode voltage!

I will wind my own transformer. I have the equipment and resources to do that. Wouldn't take more than an hour, and it feel a lot more "old skool".

I'm still in the idea development phase, thanks for all of these ideas. What do you all think about the ICL7107? Should I go with it then decode the output (since it doesn't multiplex), or just use a microcontroller (and stuck with BUGS!)

[Cyberdragon]

The ICL7135 has multiplexed BCD output.

[Chris56000]

Hi!

I must confess I'd not looked at the '7135!!

If you took the Practical Electronics April/May 1979 design and replaced the original DVM board  with one using a '7135, that would crack it!

If you go down that route you could put in.an AD636/737 and make a true RMS instrument!

Multiplexing a Nixie can be done by using a high–volrage (MPAA92 type) transistor connected upside–down such that it's collector is earthed to 0V with the emitter connected to the Nixie tube's anode, and an ordinary NPN interposed between the digit–drive pin of the. 7135 and the base of the nixie–switching transistor to invert the digit drive – multiplexing Nixies needs the digit drive transistors OFF when displaying the required digit!

Good luck, I'd be very interested in seeing your circuit ideas?

Chris Williams

[Cyberdragon]

Does anyone know what an MC1303L voltmeter chip is? I can only find an audio amp chip with that number but not an ADC. Unless it's using some sort of black magic or some bastard switched the chip. The other chip is just a quad logic gate.

[Kleinstein]

The MC1303 is supposed to be an old dual OP amp with external compensation. It might be used in a circuit to make a VCO and than use a counter circuit for display. That would be a kind of old style DVM.

[Cyberdragon]

The MC1303 is supposed to be an old dual OP amp with external compensation. It might be used in a circuit to make a VCO and than use a counter circuit for display. That would be a kind of old style DVM.

What I don't get is how it outputs an exact number of pulses. It would be measuring frequency then? The logic chip would have to be either part of the sampling oscillator, or control for the other outputs. Would the sample window be set manually or automatically?

[Kleinstein]

Some of the very old DVMs used the VCO + counter system. The VCO part is build from something like 1-3 OPs and a few extra parts, it could be discrete transistors too. It could be more for higher precision, but high accuracy is difficult this way.

The counter part is a more or less standard frequency counter, thus setting a fixed gate (e.g. 1 second) from a reference clock and than count.

[Cyberdragon]

Some of the very old DVMs used the VCO + counter system. The VCO part is build from something like 1-3 OPs and a few extra parts, it could be discrete transistors too. It could be more for higher precision, but high accuracy is difficult this way.

The counter part is a more or less standard frequency counter, thus setting a fixed gate (e.g. 1 second) from a reference clock and than count.

There's one dual amp (the stereo amp chip), a quad logic gate, and some transistors and diodes. There is a reference voltage going for the meter unit. There's no oscillators or anything on the display unit, just decade logic, so it must be entirely on the measurement board (there's no gating input to the display either). There is a switched knob on the front marked "hold" at the bottom. I assume this is "update rate" or something. I have no idea of accuracy since the measurement board is dead.