IV-6 VFD Clock w/ 4000 Series Logic

This is my take of a classic project dating back to the 1970's. The amount of more modern projects I've done has worn me down, having to fiddle around with tiny SMD components and waiting on PCBs to be delivered before I can do anything. I've jumped back in time, back to the kinds of things I was doing at college, using through hole components. Being able to pull out a breadboard and start pushing components into it to make a circuit, is refreshing. Testing ideas there and then makes things much more interesting. The only trouble is, ideas pile up!

For the longest time I've had the parts for building a clock from discrete logic. I just never knew how to design it to be usable. By this I mean, do I have a bare PCB, an enclosure, multiple PCBs, or something to hang from the wall. So the project got abandoned.

It wasn't until I saw a kit for such a clock on AliExpress, that my spark for this came back. I grabbed the kit and as soon as it arrived I began reversing the circuit board. I've made an exact copy of that kit in Kicad. Maybe I'll put that up here also.

So this project is a mixture of different ideas and existing projects. The main difference being the display technology used here is vacuum fluorescent displays. These displays being right angled, made it a simpler design as a single PCB laid flat would work. Stick it on a base, line up the displays and they point sideways. (My next project isn't so simple).

Project Breakdown

Features

There are not many features in this project. It is a bare bones style clock. It's mostly thing that would be taken for granted, such as, voltage regulation that compensates for input supply voltage changes, hardware button debouncing, compatibility with modern USB C chargers, a pause switch and the main thing, a display on off control. This can be used for turning off the displays while the clock continues to keep time, or just to save the life of the tubes while not being used.

Displays

The reason you are probably reading this, the display technology. I'll not write the history of vacuum tubes here, there are plenty of places to read about that. The main thing to know about them is that they are beautiful.

I chose the IV-6 as I've previously used IV-3's for another project, and I had 6x IV-6 in stock. Mostly as I already had them. They are also quite common in the scope of these old tubes.

One thing to note when buying them is that, due to age or poor build quality, a "batch" purchased can vary in brightness. This batch has a really dim tube and a really bright one. The others are mostly similar in brightness. Good job I have more of these already, so mixing and matching should sort that out.

I personally think I have spaced the displays too far apart from each other. I've no intention of redesigning the PCB, I'll just live with it. It still looks fine. Plus, it's my first clock project, so I'll get better from here.

CD4060

• This is the heart of the clock, quite literally. It is what generates the pulses which the whole clock is based upon. It's a 14 stage ripple counter. Which isn't as complex as it sounds.

• The ripple counter divides the frequency of an oscillator by up to 14 times. This is not to say that the frequency is divided by 14 but to say divided by 2, 14 times. With a 32.768KHz crystal, that divides as follows.

• 32.768KHz, 16.384KHz, 8.192KHz, 4.096KHz, 2.048KHz, 1.024KHz, 512Hz, 256Hz, 128Hz, 64Hz, 32Hz, 16Hz, 8Hz, 4Hz, 2Hz.

• On the 14th divide, magically there is a steady 2Hz frequency. Not what you would expect for a clock. 1Hz would make sense, 1 pulse per second. This is sorted later.

CD4511

• The 4511 is responsible for taking the output of the counters, then transforming it into a set of signals to light the correct parts of the displays. This is to display human readable digits.

• The outputs are driven high, to 5v, when active. This lines up with common cathode displays. The same type as all vacuum tube displays. The displays require a positive voltage on the segments to attract electrons, these electrons hitting the segments cause the phosphor on the segments to glow.

CD4013

• This is responsible for getting the steady 1Hz signal for counting seconds. By connecting the inverted output to the data input, the chip provides its own data to latch. The latch is triggered by the 2Hz signal.

• Each rising edge of the 2Hz signal, triggers the latching of data in the flip flop. As the data input is always the opposite of the main output, two pulses are second are needed to change the output state at a 1Hz frequency. One rising edge to change the output level high and one rising edge to lower the output level.

• The inverted output is also used for the colon separator. The regular output could also be used but the colons normally illuminate when the seconds digit changes. Otherwise the colons switch off when the digit changes.

CD4518

• These chips are backbone of the clock. They take the pulses generated by the flip flop to count them. Each count is pushed to the output pins in binary format. They automatically know to reset to 0 after counting up from 9. The output is 4-bit wide, which can count up to 16 but in this case, not all combinations are used.

• 0000 = 0, 0001 = 1, 0010 = 2, …, 1000 = 8, 1001 = 9

• The 1Hz clock input to these is connected to the enable pin. Looking at the datasheet, there is a dedicated clock pin. The name is a little misleading but, it really is an enable pin. The clock input is gated with enable in AND logic. The clock input is inverted before this. The enable must be held high to allow clock pulses through. Depending on when the counter needs to count, these pins can be swapped. When used as described, the counter changes to the next number when the clock input goes high. In this setup, to carry to the next counter when the first counter loops back to 0, it would need extra logic. To get around this, the enable pin is used as a clock input. This way the counter increments when the clock signal goes low. This means when the binary 8 output goes low, if connected to the enable of the next counter, it will count.

• The problem comes when resetting the tens digit. Time works on 60 seconds to a minute and 60 minutes to an hour. The tens digit needs to be reset when changing to 60. This does require external logic and the use of the reset pin of the counter. To increment the hour digit, the same principle can be used, only this time using the extra logic to trigger the count for the next digit. Simply, when binary 60 is reached, the extra logic outputs a high level signal, resetting the counter to 0 from 6. When this happens, the logic output goes low and triggers the count for the next counter.

CD4081

• This is simply the quad two channel AND gate chip. Three of the four gates are used to reset the tens digit of each hour, minute and second when they reach 60 or 24. They also are used as the counter for carrying to the next digit. Eg, 60 seconds triggers a count of 1 min, 60 mins triggers a count of 1 hour.

2N2981A

• These are for level shifting. Used in this way called high side switching. This is needed due to the display tubes needing 28v for operation. The logic circuits used here all run at 5v, not 28v. You can't simply control 28v with 5v. It is quite simple to control, but not in the sense that you can use a single transistor to switch on and off 28v. This chip shrinks down all the components that would be needed otherwise. A simple discrete logic alternative would require 2 transistors per digit segment, 14 transistors per tube and a total of 84 transistors. Without counting the resistors required that would be in the hundreds.

• These are a useful little guy. It's a shame they have been out of production for years now, as there isn't really a good replacement. There are some SMD alternatives, but in the DIP world there just isn't. The 2N2904 is something I considered, but it requires an extra transistor for each of the outputs, making the project even larger.

MC34063

• In the effort to keep the project looking mostly like it came from 1985, the 34063 is an easy to use buck/boost supply controller that was around back then.

• This project pushes the chip to its limit. It has an output switching current of 1.5A while this project requires 1.37A for the 28v rail. The actual current needed is more like 100mA. The switching current is just the peak current through the inductor, which if the input voltage is being boosted a lot, will be a high current.

• The display tubes require a low voltage for the directly heated cathode. This is also created from the use of the 34063 in a buck converter topology. The heater voltage can easily be created with a series resistor from the 5v supply. This has the problem of the voltage changing directly in proportion with the supply. The USB voltage can vary a large amount, this would cause the display brightness to vary a lot. It could even damage the cathode. Should the voltage drift too high. Using the buck converter for powering the cathodes, allows the 34063 to adjust the output voltage regardless of the input voltage.

Project Timeline

This isn't a comprehensive documentation of the design and build. I didn't intend to make one. I had took some photos along the way for my own use. I thought they would be worht sharing.

• Here is the design for both power supplies needed for the tubes to run. I had used the MC34063 before, so it seemed the easy choice for me. It matches well with that old sytle visuals too being it is very old itself.

• The first prototype went worse than I'd hoped. It was not a failure though, as it was useful enough to iron out all mistakes and oversights. The main issue being I got the pin out for the tubes wrong. At the time I was messing with the IV-11 tubes and mistakenly used that pinout for the IV-6 footprint. I thought I'd just got it rotated somehow, but no, it was the pinout for the IV-11. On the bright side, both power supplies were working and at least one of the high side drivers. The display would also change the garbage it displayed about every second. So it seemed to be counting.

• I wondered if to just correct the pin out and order a new PCB, hoping that there would be no other problems. After sitting on things, I figured I could use LEDs to see if the BCD signals were changing correctly. Not realising until connected up I'd made a binary clock. I'd somehow got the hours digits backwards and the hours reset function was broken. I'd spent hours debugging why the hours increment stepped in twos at a time. A 22K resistor slipped into my 2K2 resistor box...

• I got the changes made and ordered a new PCB. Everything on here working just fine, quite a relief. That was until during testing, the hours passed 23. There are in fact not 30 hours in a day. The reset circuit was still broken. Fortunately the hours reset at 42, so I knew I'd got the digits reset signals flipped. I'm not building another one, the sockets aren't cheap to keep wasting; some wires underneath and some trace cutting fixed it up. The released files conatain a fixed version of this.

• Here I was trying my tube bases to improve the looks by hiding the spindley legs. They should be mounted lower down to the PCB, but I didn't want to cut the legs short. I also figure if it looks silly I can lower them further. I can't raise them back up after the legs are cut. I can release the files for the PCB base and tube bases if wanted.

• It's new home in my cosy, cute, corner. It's a work in progress <3

• A bonus photo. I was intending on make this reuseable for IV-11 tubes by simpley changing the cathode resistors and tube footprints. It could work, but I've abandoned that idea. I think I'll use them in an AVR based clock instead.

Troubleshooting

There are some digits brighter than others

• That's soviet build quality for you. They are not exactly build with care and precision. They couldn't even get the datasheets right. To minimise this, swap the cathode resistor for that display with a higher one. The value in the schematic is the minimum value to not cause damage, with a little head room. My suggestion is that if a digit is only a little brighter, leave it alone. If it is noticeibly brighter, try an 8R2 resistor in place of the 6R8. If it is a lot brighter, try a 10R or 12R in place of the 6R8 resistor. Going much higher could cause long term damage due to the cathode not being hot enough, causing sputter.

Some digits are much dimmer than others

• That's soviet build quality for you. They are not exactly build with care and precision. They couldn't even get the datasheets right. There isn't a lot to be done about this. It could well be that the tubes have begun letting air in or they were always like that. You can try changing the cathode resistor from 6R8 to 4R7. This didn't make much difference for me, the glow from the cathode gets brighter and any phosphour brightness increase is distracted from the red glow. Going to a 2R2 resistor just turns it into a bulb.

Digits skipping when setting time

• The quality of the buttons used make a difference here. Some will bounce more than others, that is what is causing the digits to skip. Try either a larger capacitor near it or swap the switch in hopes it will not bounce so much. You can sometimes see each digit as it bounces through them.

Digits incrementing 2 or more when setting time

• Check the component values in the button debounce circuit. A too high value resistor will cause the digits to change so fast, you can't see digit in between. Similar to the preious issue.

Digits are dim when a certain number is displayed and segment missing

• Turn it off. Your switching IC and its output components are probably burning hot, along with one or more UDN2981A chips. If a chip it hot, it will point you to the digit causing the problem. You will have a short between the digit pins somewhere after the output pins of the UDN2891A. Likely to ground or on of the cathode pins.

28v supply is outputting 5v

• The boost circuit isn't working. The way it is wired the input supply is directly connected to the output supply through a diode, resistor and inductor. The chip isn't switching. Pull it out, check the connections. Look at the components around it, make sure they are soldered properly, make sure they are the right values and not mixed around. Make sure the display switch is on.

1.5v supply is outputting nothing

• The buck circuit isn't working. The way it is wired the output supply comes from the emitter of the internal transistor of the chip. The chip isn't switching. Pull it out, check the connections. Look at the components around it, make sure they are soldered properly, make sure they are the right values and not mixed around. Make sure the display switch is on.

• Make sure the USB connector is soldered properly. Make sure there are no bridged pins.

There's a bright red glow in the tube from top to bottom

• Turn it off. The tube will be damaged permanently. You have a short somewhere allowing a higher voltage to be across the cathode. If that is not the case, check the output voltage of the buck converter chip. It should be close to 1.5v. If not, check the component values of the buck converter circuit are correct and none are mixed around.

• The tube needs 1.25v, make sure the 6R8 resistors are in properly and there are no shorts.

There's a slight red glow in the top from top to bottom

• This is fine and normal. The cathode should be glowing slightly under operation. Should the display be a bit dim, the glow will be visible. With a bright display, the glow will disappear from being drowned out by the florescence.

• If you are worried, swap the 6R8 resistor for a higher value. Try 8R2 first.

Some/one tube(s) don't light up.

• Before thinking it is a bad tube. Make sure you have the correct voltages on its pins. There should be 28v on the grid pin. 28v on any of the display segments that should be lit. 1.25v on the cathode input pin and ground on the other cathode pin. If these are missing the problem is elsewhere. See the section on power supplies.

• If the voltages are fine, it’s a bad tube.

The clock drifts out slowly

• Yeah it's an old fashioned clock, what do you want?

• It shouldn't be more than a few seconds per day though. Check the values of the oscillator circuit. Higher accuracy components would be better than normal ones. Normal ones could be 10% out, so even swapping them with another of the same might be of help.