Giant Nixie Tube Clock using MM5314N IC / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Giant Nixie Tube Clock using MM5314N IC

1 2 next

Move Thread

LAN_403

Lightening Rod

Fri Feb 05 2010, 01:11AM

Registered Member #2674 Joined: Fri Feb 05 2010, 12:35AM
Location:
Posts: 11

I found a circuit which can be modified to drive the B7971 Nixie tubes using the M5314N Clock IC but need help with mods.

Here is the link to the article and schematic:

Here is the data sheet for the MM5314N:

1. Power Supply Mod:

Modify the power supply to the chip using a 6.3 - 0 - 6.3 V AC out center tap transformer, a full wave bridge rectifier, 1500Âµf, 35V filter cap and a 7812 regulator. This will give a regulated 12 v DC supply for Vss. The center tap is left unconnected. This is better than the 9v, unregulated supply in the schematic and may fall shy of the 11v minimum required for the IC. Can a 22 ÂµF, 400v cap be used to filter the mains for the HV side of the circuit instead of the 15ÂµF, 150V as in the schematic, since this is what I have on hand? Can a smaller 1 watt bleeder resistor be used say 2.2k instead of 220K in the schematic? Why was such a high value chosen in the first place?

2. Transistor Substitution Mod:

MPSA42 (NPN) will replace of the UL624. MPSA92 (PNP) will replace the SA480. Adjustment of the multiplex frequency may be necessary. For the MPSA92s, Q8- Q13, there is a pull-up resistor from the transistor bases to â€œgroundâ€ in the schematic. A question was raised by Alan 'A.J.' Franzman: â€œI've just realized another possible mistake in the schematic, but I'm not quite sure if it's wrong or not. You may need to build a temporary breadboard version of the circuit (or at least most of it) to find out. Each of transistors Q8-Q13 is shown with a 2.2M ohm resistor going to 0 V DC. I believe these may need to go to the "+150 V DC" high voltage line instead, but because the resistors are of such high value and the transistor base inputs are capacitively coupled, the circuit may work either way.â€

Does anyone have an opinion and can connecting the transistor base to HV damage the IC? Should I put series diodes with the IC outputs to prevent stray voltages?

3. Mutiplex Timing adjustment on pin 23:

3000 ÂµF is appears to be wrong, based on Figure 3 from the MM5314N.
datasheet (it's WAY off the chart!) If the author needed to use that value to get the chip to work, it was probably because of the chip not functioning correctly due to supply voltage being too low. Either .01 ÂµF with the recommended 220K ohm resistor from 0 V to pin 23, or if I use the 100K ohm resistor as shown in the schematic, will increase the capacitor to .022 ÂµF. It seems from the article that the values were based on needing a higher frequency to allow pulses from the IC to get through series capacitors connected to the digit enable outputs. Can anyone shed more light on this issue?

Since .5 ÂµF is no longer a common capacitor value, I will substitute .47 or higher 250 v nonpolarized mylar caps.

4. Fixing â€œInter-digit Ghostingâ€

All of the above was devised to prevent flickering and display ghosting. I noticed that when the same IC was used for Numitron tubes that every segment of the Nixie tube needed a diode.

See

Perhaps the inter-digit blanking circuit in figures 9 and 10 of the MM5314 data sheet should be used? Â The example given was for the 28 pin version of the clock chip MM5309 and shows the circuit connected to pins 26 and 28. Can I assume that this same circuit will work if connected to pins 23 and 1 respectively?

Any help much appreciated!

Dan

Direct email is kuriloff at NYHNI dot org.

Conundrum

Fri Feb 05 2010, 07:40AM

Registered Member #96 Joined: Thu Feb 09 2006, 05:37PM
Location: CI, Earth
Posts: 4061

iirc it might be worth putting a 4M7 or 10M 1KV resistor across each segment as well, so that they switch cleanly.

adjust value until ghosting goes away.

also could be worth adding series resistors on each segment as it could be overcurrent causing the problem. From memory, neon devices have a negative resistance when lit and that can cause strange problems such as flickering if they are overdriven.

-A
"Bother" said Pooh, as he dumped core...

Nicko

Mon Feb 08 2010, 09:19AM

Registered Member #1334 Joined: Tue Feb 19 2008, 04:37PM
Location: Nr. London, UK
Posts: 615

This question should be active in the Yahoo! NEONIXIE-L group - the main nixie-related group on the net where answers to this sort of stuff are given all the time. I know Dan is a member there too...

Lightening Rod wrote ...

4. Fixing â€œInter-digit Ghostingâ€

All of the above was devised to prevent flickering and display ghosting. I noticed that when the same IC was used for Numitron tubes that every segment of the Nixie tube needed a diode.

See

Numitrons are incandescent displays, not cold-cathode like nixes. The diodes used are for demultiplexing, not for anything else - its a completely different issue.

Conundrum wrote ...

iirc it might be worth putting a 4M7 or 10M 1KV resistor across each segment as well, so that they switch cleanly. adjust value until ghosting goes away.

also could be worth adding series resistors on each segment as it could be overcurrent causing the problem. From memory, neon devices have a negative resistance when lit and that can cause strange problems such as flickering if they are overdriven.

Neither of these are the problem. Its not over current (which does not cause flickering, only sputtering) and I've never ever seen a nixie design where someone put resistors over the segments (and there are 100s of designs out there).

Firstly, ghosting is not caused by switching problems, it's caused by residual charge on cathodes caused by internal capacitance in the tube - generally, bigger the tube, bigger the problem. if you have a ghosting problem, the cathodes need to be pulled to about 80VDC when off rather than allowing unlit cathodes to float and possibly glow faintly. The negative resistance exhibited by cold-cathode discharge tubes is nothing to do with this. Nixies "strike" at about 180V, then maintain glow at about 120V. Say you want 2mA for a segment, that means you have to drop (180-120) volts at 2mA in the current limiting resistor = 30Kohms at 120mW dissipation, so say use a 1/4W 33K resistor. Note that you need to look at the datasheet for each tube to work this out as it depends on your HV voltage and the characteristics of the tube type - they do differ and the segment/digit current is the single most important factor in determining the life of a nixie. Slightly different rules apply when multiplexing tubes as the tube is not on 100% of the time (maybe only 25% or 17%), so you can increase the current a bit though not proportionally (i.e. not 4 or 6 fold - typically double).

The referenced design is poor on many levels. Firstly, and most importantly, it uses a "live" chassis - ground is referenced to one side of the mains, and the other side of the mains is used to directly generate the 180VDC the tubes need. Its also 110VAC specific - connect it to 230VAC in Europe and watch the destruction of your tubes.

When the article was written in the '70s that might have been acceptable behaviour - it most certainly is NOT now.

Secondly , as mentioned above, cathodes must not be allowed to "float" when switched off - a simple potential divider can be used to pull the segments to the correct mid-point - there are many examples of this on the net, e.g. see the manual for the "Geek Clock" here -

.

Ghosting is a problem that mainly manifests itself in poor multiplexing strategies or with large tubes with high internal capacitance and hence the capability of retaining significant charge when the cathodes are switched off. The MM5314 datasheet referenced by the OP even mentions this explicitly (though they call it blanking rather than ghosting) on page 5-15.

Thirdly, the B7971, like other alpha segmented-style tubes such as the ZM1350, has cathodes of differing sizes - this means that each cathode has a different recommended operation current - using the same current for each cathode is absolutely NOT recommended for such segmented tubes - the life of a nixie falls almost exponentially as you increase the operating current. Read the B7971 datasheet and you'll see the recommended cathode resistors for each segment. Note that non-segmented tubes are carefully designed to have the same surface area (if possible) for each digit. i.e. the "8" is quite skinny but the "1" a bit fatter etc. This allows component saving in that you can then get away with a single anode current limiting resistor rather than having 10 different cathode resistors - one for each digit. See

for a rather better way of driving these tubes (project page at

Basically, what you have here is a BAD design that will at best perform badly (and damage your tubes) and at worst cause you or someone else a serious, potentially fatal, injury. Don't use it.

Rant over!

Steve Conner

Mon Feb 08 2010, 10:20AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Well, that's that design assassinated, then!

In the defence of mains-operated nixie clocks, I'd say that about 50% of the home-made ones I see on the net have a live chassis. The HT current draw is very low, so it's easy to isolate using two little transformers back to back.

I've been experimenting with the Allegro A6810 VFD driver IC as a SPI input nixie tube driver, with nixed, err, mixed results. (It was working great until I overvolted it and blew it up.) It does actually pull the segments high when "off".

Nicko

Mon Feb 08 2010, 10:53AM

Registered Member #1334 Joined: Tue Feb 19 2008, 04:37PM
Location: Nr. London, UK
Posts: 615

Steve McConner wrote ...

Well, that's that design assassinated, then!

Its poor, completely unnecessarily dangerous and I hate that sort of thing!

Steve McConner wrote ...

In the defence of mains-operated nixie clocks, I'd say that about 50% of the home-made ones I see on the net have a live chassis. The HT current draw is very low, so it's easy to isolate using two little transformers back to back.

There's no defence of using a live chassis for this sort of stuff. The design you are referring to is probably Mike Harrison's at

, the use of which is pretty widespread mainly because its been around a long time. Note the use of A VERY LARGE RED WARNING at the top of his description

Mike is a superb engineer (as you know) and almost certainly wouldn't do a design like that today...

If you must build this design, Mike himself recommends using back-to-back isolating transformers. Better still, use one of the many small SMPS designs done just for this purpose (such as mine at

) If you don't want to build your own, several are available on eBay (some based on my design) and others available through web shops, typically at prices so low its uneconomical to build it yourself.

Steve Conner

Mon Feb 08 2010, 11:21AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Yes, it was Mike's design I was thinking of.

If we're going to redesign this guy's project, then there's no need to use the MM5314 clock chip, either. Any number of people have written PIC firmware to drive nixie clocks.

I was planning to modify Fred Niell's firmware to work with the A6810 instead of the Maxim VFD driver he used, but I never got round to it, and the A6810 is discontinued now.

Nicko

Mon Feb 08 2010, 01:45PM

Registered Member #1334 Joined: Tue Feb 19 2008, 04:37PM
Location: Nr. London, UK
Posts: 615

I rather like Chris Barron's Smartsockets - they make building a clock etc. absolutely trivial - that is, assuming you are happy to get a design in rather than building one from scratch. Personally, I prefer doing stuff from the ground up, but there is a Smartsocket for B7971s - Have a look at the novel-named "Smartsockets" Yahoo! group at

. Its a good stating point...

John Taylor (who also sells wonderful nixie power supplies - hint hint) has re-worked Chris' original B7971 module to be more "productised" (how I hate that word). See

Lightening Rod

Sun Feb 14 2010, 08:51PM

Registered Member #2674 Joined: Fri Feb 05 2010, 12:35AM
Location:
Posts: 11

Nick

Now that the rant is over, thanks for your comments concerning the MM5314 clock chip. I have not yet built the clock but plan to use the chip since I have several on hand and would like to see if it would work. I took your advice and ordered a few SMPS from John Taylor. One module should be able to supply all six Nixie tubes from a 9 volt AC wall adapter. The 60 Hz signal will be tapped from the AC wall wort and then rectified and filtered for the Â±11 volts needed for the clock chip.

I know from a purest stand point that this is not the best solution for driving the Tubes and yes PICs are available with bells and whistles but I would like to get this to work with a minimum of components.

As far as ghosting is concerned, I will let you know if it is a problem.

I am not sure how to deal with the different current limits of the tube segments. I saw the FLW schematic but need to redraw the MM5314 circuit. MPSA42 (NPN) will replace of the UL624. MPSA92 (PNP) will replace the SA480.

Can you help me with a schematic for this chip incorporating the FLW design for driving the individual segments at the correct current in this multiplexed design?

I know its reinventing the wheel but I just want to use these vintage clock chips...

Thanks,

Dan

Lightening Rod

Thu Feb 18 2010, 05:43AM

Registered Member #2674 Joined: Fri Feb 05 2010, 12:35AM
Location:
Posts: 11

I found another circuit using the MM5314. See attachments.

The essential differences between the two circuits involve the segment drivers and the digit drivers.

In the alternate circuit 200v is used instead of 170v to drive the anodes switched on by the high voltage transistor, the collector is tied to 100v through a 1 meg resistor and the base tied to 200v through a 1 meg resistor. In the circuit you sent me the collector is not tied to anything other than the anodes and the base goes to ground through a 2.2 meg resistor. Furthermore, this alternate circuit uses a tansistor array from the clock chip enable pins instead of a simpler capacitor signal coupling.

In the alternate circuit, the segment driver is very similar with the exception that collector is tied to the 100v line via a 1 meg resistor whereas in the circuit you sent me, the collector is not tied to a reference voltage.

One engineer mentioned that "the cathodes must not be allowed to "float" when switched off, therefore, a simple potential divider can be used to pull the segments to the correct mid-point - there are many examples of this on the net" Is this what is being done in the alternate circuit I have attached?

If I use a switchmode power supply, can I use a simple voltage divider with two resistors shown in the alternate circuit to achieve the same effect?

Which design is a better solution or are they both acceptable?

Thanks for the help.

Dan
]b7971clock.pdf[/file]
]giantnixieclock.pdf[/file]

Lightening Rod

Thu Feb 18 2010, 05:49AM

Registered Member #2674 Joined: Fri Feb 05 2010, 12:35AM
Location:
Posts: 11

Attachment Giant Nixie Clock Circuit
]giantnixieclock.pdf[/file]

1 2 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob