26 November 2022

The 4-weeks challenge: deliver a Nixie clock from scratch

Last weekend of October a curious teenager guest noticed the Nixie clock in our living room. The most eye-catching element was the short depoisoning routine that runs once a minute. B-5750 Nixies have their numbers 0-to-9 back-to-front, which is an even more eye-catching feature.

So I quickly explained him and his family some basics of these lovely tubes, showed them my lab/workshop/messy room and of course I told them I had designed and built that clock.

The next day I came up with a self-challenge: design and build a fully functional Nixie clock that I could give the teen at our next (and probably last for a very long time) meeting four weeks ahead.

With a small personal "library" of hardware design and AVR firmware I was quite sure I could get over all the small challenges inside this project:

  • design a working circuit and receive the PCB from the fabhouse
    • sub-task: fix the KiCAD footprint for the chosen ZM1020 Nixie
  • design a suitable case and get it 3D printed (this then turned to laser cut)
  • build it without easy access to my lab for the same period of time
  • write a user's manual

The PCB was designed with KiCAD and etched by JLCPCB. To save on board space and stay into the chosen round design I used a bare AVR ATmega uC with its internal 8 MHz clock; it also drives the voltage booster. On the board I included an ICSP header which is very useful if you're debugging or adjusting the firmware code. I used a ZM1020 round top view tube

While waiting for PCBs to arrive I learned how to use OpenSCAD to design the case, which is plain geometry applied to programming. I also risked to have no AVR microcontroller for the clock as the ATmega48 doesn't have enough program memory (or I would have needed to do heavy optimizations). I managed to unbrick a couple of ATmega168 that are enough for my firmware (wrong fuse settings).

The circuit diagram was correct. The PCB was suitable, meaning I kept HV and LV separated enough. The laser cut case (two plates) was perfect and I picked it at 3.5 weeks into the challenge. In the last week I did some firmware corrections and upgrades and today the clock has been delivered to a happy teenager.

Here it is!

Single Digit Nixie Clock with ZM1020
Single Digit Nixie Clock with ZM1020

16 November 2022

Arduino based AVR ATmega fuse doctor

Given the ongoing semiconductors shortage and needing urgently a bare ATmega168/P/328/P chip, I had to reset two 168's unresponsive to the usual avrdude programmer command.

Actually I could have spent about 10€ to buy an ATmega328P rather than spend two hours assembling the fuse doctor.

According to my own blog posts, back in 2010/2011/2012 I had already built a fuse doctor, but the link doesn't work anymore. Well, I found this Fuse Reset by Thuta Kyaw on Instructables. It uses an Arduino Nano, which I still had at home. This time I decided to do a permanent build, in case I will need it again 10 years from now.

My build of Thuta Kyaw's Fuse Reset.
My build of Thuta Kyaw's Fuse Reset.

Building it with wires requires patience and a steady hand. There is no radiofrequency involved, so there's no need to do careful planning of the layout.

Operation is easy. Fit the ATmega patient in the socket, apply 12V, wait few seconds for the LED to switch off. Enjoy.

07 November 2022

OpenSCAD for 3D printing

If you are a software programmer/developer of some sort in any high-level language and you need to design an object for 3D printing, I think that OpenSCAD is for you.

I am designing a case around a PCB I have ordered. It's a cylinder with a diagonal cut, three supports for the PCB and two holes for the buttons. In OpenSCAD the final object is described as sum/difference of basic shapes (cylinder, cube, ...) in a simple programming language.

The GUI provides both auto-completion and 3D preview.

The following code makes an empty cylinder with a diagonal.Try reading it through:

module cylcut()
            cylinder(h=80, r=39);
                cylinder(h=82, r=37);


There is a long way to go and learn before I get to the STL file for the printer, but I like it!

18 October 2022

HP 3734A Electronic Counter

I really enjoyed Mombarone HAM Fest (near Asti, Italy), October 2022. The visitors flow was relaxed and I could spend time looking above and below stands.

In a corner I spotted a device that supposedly contained Nixie tubes. The seller, who had a stand of regular wristwatches, said: "it's for the Nixies". It was wrapped in plastic, I could see it had physical damage. He asked 15€ but had no change, so I raised the swap to 20€, which is closer to the current ZM1022 quotation (2022AD). That's 4€ per tube, plus the transformer and .... a true HP device!

So I got an "HP 3734A Electronic Counter". The Nixies tested OK for gas but one side of the case has heavy damage with broken aluminium brackets! The plastic display bezel is broken as well as a corner of the protective glass (it is composed of an orange plastic sheet over a clear glass).

As looked from above it is also not square but a bit rhombic. Nothing a light hammer and a 3D printer cannot fix, right?

Since I had nothing to loose, I powered it up (with the right "163" cable). This was the pass mark to decide on fixing or scrapping the whole device (minus the Nixies).Well, the verdict is "fixing" since it did power up! Without smoke.

I haven't found a free PDF manual online, but it's not that hard to understand how it works, showing either a frequency or its period. Unfortunately it doesn't work even if the internal 100 kHz oscillator is runnng.

There is also a pulse counter function controlled with start/stop/reset buttons which obviously bypasses all signal conditioning and averaging circuitry. It counts either an EXTernal or and INTernal source. Well, selecting "INT" on the front knob combined with "INT" on the backside switch makes the counter use its own 100 kHz clock as source. Et voilĂ , Nixies were running as soon as I hit "START".

If the case can be re-aligned and rebuilt, there is enough working electronics to make a five digit clock. A big plus is that input/start/stop controls are available on the back panel too so everything can be hidden inside or behind. Nice!

HP 5512A-4A display boards

Each Nixie is mounted on a removable board. The concept of this 3734A is similar to HP 5212A, 5512A, 5232A, 5532A counters from the same lustrum (1960-1965) as they share the display architecture. Earlier 5512A had the option to install a board with 10 neon lamps or with the "inline" indicator (a Nixie).

I will look for a manual of these other devices to grab an idea of the circuit diagram.



HP 5512A-4A insideEach display board uses a mix of transistors and a rudimentary integrated circuit as visible in the last picture (click to enlarge as usual).

Next step is to remove all boards and try to straighten the case.


26 September 2022

Almost working circuit with 6LO1 CRT

A square little 6LO1 CRT and associated circuit.
A square little 6LO1 CRT and associated circuit.

The are a few reasons why I bought this homemade circuit originally meant to display the two ellipses when tuning an RTTY signal:

  • it has a small CRT, probably 6LO1
  • it has a proper transformer with filament and HV
  • it might not work right out-of-the-box 

I was three times right! The CRT lights up and responds to both X-Y inputs but

  • it is out of focus
  • the beam doesn't reach display ends
  • something overheats

At closer inspection I saw that a resistor has suffered some form of heat shock. One resistor is floating and the wiring is close to a nightmare, let alone the routing and crossing of high voltage lines.

I do have the schematic diagram which matches the PCB (oh, BTW, it is double sided but the magazine showed only the bottom side). But schematic and component placement (not shown here) differ with values especially in the CRT bias network.

The schematic diagram: it's a starting point!
The schematic diagram: it's a starting point!

Well, it is not a complex circuit and with some extra parts I should be able to get it running... one day. I should also add a Z-axis control if this is going to become a clock.



17 September 2022

A simple way to 77 GHz

The Italian government has approved a new frequency allocation document, so I checked which changes would impact the amateur radio world. I also looked at "our" bands in the GHz range and I saw we share 77 GHz with many other users.

Since I was "lucky" with 24 GHz radars, I hopped to the usual Chinese online marketplaces to see what they sell on 77 GHz.

It turns out that 77 GHz are currently used in automotive radar sensing and of course there are modules for sale.

Looking at pictures I identified a product name: Continental ARS-408-21 Radar. There is documentation on the Continental Engineering website (they also have other 77 GHz radars), so it could be adapted for a simple HAM RTX without too much "guessing". They claim to be able to detect objects at 1200 meters, much more than my 24 GHz radars.

So, probably simple but not cheap, though! These modules are currently (2022/09) selling at 80€ each, so I will just observe how the market evolves and prepare in case their price drops.


11 September 2022

Italian RAI has stopped medium wave service.

As of today, Sunday September 11th, 2022, the Italian national broadcast RAI has stopped all its medium wave transmissions. 567, 657 and 999 kHz have gone silent.

In most of the Italian territory there are no signals to be received in MW during the day with a domestic receiver.

Time to put into service those 1.000 MHz canned oscillators?

28 August 2022

Summer 2022 PCB Collection

In order to reduce the impact of returning home from holidays, this year I prepared a small present that would make me want to get back: not one, but three PCBs!

In July I designed the PCB for a secret project (it will be a birthday present). Then I added a board to simulate the spin of the WE6167 dekatron. And at the last minute I designed a board that would modulate either an HB100 or a CDM324 radar transceiver.

Each design is packed separately.
Each design is packed separately.

All the design was done in KiCAD 6 and each board had some challenges to learn something new: proper Net spacings for HV, new component footprint that includes a large "hole", adding an ICSP connector.

I chose JLCPCB as fabhouse, but I needed to slow them down so that the packet would arrive once I am already back at home, not before.

First of all, I chose three different colors for the three designs, and these colors take two extra days for production.

I submitted the order on a European Saturday afternoon and it went in production almost immediately: I didn't check that JLCPCB respects business hours, but they are open 7 days a week! Little mistake on my side, I lost one day, but otherwise I would be tempted to make more changes to the boards.

Then I paid for the slowest shipping option.

Why all this trouble? I didn't bring a computer with me on holidays, so I had to submit my order before departing.

Who came home first? PCBs, of course! They beat me for 2 days, shipping was too fast.

The secret project has passed the hardware smoke test and is now in the hands of the firmware development team (it's always me).

Three designs, three colors.
Three designs, three colors.

26 August 2022

A request to my German (speaking) readers

At Friedrichshafen fair last June 2022 I bought some stuff from the seller at table A012 (https://youtu.be/BH29tY4Sjyc?t=154 and next 30"). I would like to get in touch with him to ask if the has more (not complain about something broken! HI).

Is there a German forum where I could ask if someone knows him? If you asked me for the Italian scene, I would answer "ARI Fidenza forum".

Please let me know with a comment or via email to ik1zyw@yahoo.com .


05 August 2022

Using KiCAD on dual screen on Ubuntu

When I tried ordering a PCB from a manufacturer in 2017 my choice for generating the Gerber files was KiCAD, version 4 back then. All home computers run Ubuntu (no young boys that need Windows for videogames), so it was KiCAD on Ubuntu. With a couple of good tutorials and memories from OrCAD for DOS at high school (past century), the experiment was successful.

Fast forward past pandemic, I need again PCBs. This time I swapped the work PC on the remote working table with the lab laptop that runs KiCAD gaining an external monitor, keyboard and mouse.

Now, THAT IS a productive way to design PCBs!!

KiCAD 6 on Ubuntu on two screens. Unfinished circuit.
KiCAD 6 on Ubuntu on two screens. Unfinished circuit.

Schematic diagram on one screen, PCB on the other, with KiCAD (now version 6) you can make changes on the fly on either side and reflect it on the counterpart with a keypress. I needed to change quite a few pin assignments on the microcontroller to have a simpler routing once the general layout was done, and the dual screen was the way to reach the goal without endlessly swapping windows. And if you memorize keyboard shortcuts, it's even faster.

21 July 2022

SGS integrated circuits H100/H200 series

Shortly after posting about H 257 and H 258 never-seen-before(-by-me) ICs, I tried different search keywords and a datasheet popped up from the back of Internet.

According to that document SGS-ATES produced a series of High Level Logic I.C.'s designed specifically for applications in areas where noise is a hazard. The supply voltage is from 10.8 to 20V and they are compatible with MOS technology:


From the datasheet you learn that the suffix D1/D2/D6/B1 distinguishes between voltage+temperature range and case material. I think it was 1970-1975. Afterwards SGS should have adopted the standard naming for their IC line-up.

As already stated, contact me if you need those chips (H 257, H 258, H 202, H 203).

14 July 2022

Integrated circuits SGS "H 258 B1" and "H 257 B1"

One of the Nixie boards I bought in FN Ham Messe 2022 does not use the usual 7441 Nixie driver + 74xX Latch combo, but rather some undocumented SGS "H 258 B1" and "H 257 B1".

The pinout of H-258-B1 BCD(?)-to-Nixie driver is different from similar documented components (see http://madrona.ca/e/nixieref/index.html for a comprehensive description), so it is worth reversing it from this simple PCB.

I shot a picture of both sides of the board and created a single Xray-like view using an image manipulation software (The GIMP) as I've done in the past. This is the result and a zoomed view of one pair of these ICs:

Both sides of the board with SGS chips.
Both sides of the board with SGS chips.

SGS H257 and H258 combo.
SGS H257 and H258 combo.

The resulting pinout for H 258 B1 is:

1   input
2   input
3   input
4   digit 2
5   digit 3
6   digit 7
7   digit 6
8   GND
9   digit 4
10  digit 5
11  digit 1
12  digit 0
13  digit 8
14  digit 9
15  input
16  +V

I have no idea and no interest to discover which technology SGS used in 1974 for these parts. I will not use them in any of my projects because I would have no spares or direct replacements.

If you land here looking for a replacement, then contact me.

09 July 2022

Increase discharge current in XH-M239

The Lithium Battery True Capacity Tester Module XH-M239 discharges 18650 cells with a current of about 500 mA. It takes about 6 hours to test a single good cell: can we increase the discharge current?

The load is an 8 ohm 10W resistor while the current sensor is a 20 milliohm SMD resistor.

I quickly verified that if I parallel to the load a low-ohm power resistor the current increases and the mAh counter goes faster. So it should be possible to increase the drain on the battery under test and get a consistent result (with the error estimated in the previous post).

Would the circuit handle it? The tracks seem to be thick enough and we can easily measure their voltage drop, for example V across the battery vs V across the load. If we double the current to 1A, the 0.02 ohm sensing resistor would dissipate P = R * I * I = 0.02 * 1 * 1 = 0.02 W. 20 mW is well below the unknown rating of the resistor, which I would assume to be 250 mW.

Any resistor below 10 ohm with adequate power rating (10W or more) will do the trick if soldered in parallel to the stock load. For my quick test I went as low as 2.35 ohm load which resulted in a measured current of 1.5 A.

Note: you do it at your own risk as you understand BOTH stated AND not stated RISKS of this modification (short circuit, explosion, fire, skin burns, end of the World, end of Internet, human race extinction and so on).

Even Li-po cells from laptop battery packs should be able to withstand C to 2C discharge current so this trick is useful for testing a large amount of batteries with just one XM-M239 device.

04 July 2022

Li-Po capacity meter XH-M239: does it get it right?

In the last months I've used the FT-817 on the field and noticed that the Li-Po battery pack wasn't lasting as much as I expected. I had to choose: to build or not to build a device to measure the battery capacity? I opted for buying one and the choice was XH-M239 from aliexpress.

It tests one cell at a time with about 500 mA discharge current. It's slowish, but it mimics the consumption of a laptop I dare to say.

I set the stop voltage at 3,4 V because that's in the range for my intended use (3S "12V" pack) and all cells tested scored a very low rating! The best reached 1350 mAh, and many were below 500 mAh. None of them is brand new from a reputable source, so that was somewhat expected.

Nevertheless I started questioning XH-M239 readings. With a DVM, it's easy.

The measured battery voltage does match my DVM (that was checked against a reference voltage source).

The measured battery current does not match! See:

The real current is lower. Battery under test is connected to the screw-on terminals.
The real current is lower.
Battery under test is connected to the screw-on terminals.

The actual current flowing out of the battery is lower! The capacity is overestimated by 17%! As long as the time interval is correct. I will repeat the measurement at a higher battery voltage, in case we are hitting some sort of non-linearity. The measurement circuit uses a 0.020 ohm resistor and a "A53A" SOT-23-5 device, which is a voltage detector.

Alright. I don't need to guarantee the capacity of my battery packs, I simply want to build "good performers" with these second hand 18650 cells, so the higher mAh the better.




30 June 2022

Tune-a-Lite gas filled tuning indicator

I learned about gas filled tuning indicators in the last month, as I started reading through an Italian forum packed with information on restoration of old radio receivers ("Elettroni al tramonto"). They were used before the introduction of magic eyes, so approximately from 1930's to 1940's. It is like a shorter IN-9 or IN-13 tube. On radiomuseum there is some documentation: look for Tune-a-Lite. Of course it ended up in my "to look for" mental list.

I was very lucky as I found one at Friedrichshafen Ham Messe 2022 on the stall of a German seller that deals with vacuum stuff. I tested the component for leaks with my magic wand neon tester, which surprised two more sellers in the same day. Given the presence of neon inside, it came home with me.

On the mentioned forum there are few diagrams showing how to wire the indicator so I tried the setup while swapping the three terminals, since the pinout was unknown. Moreover, since it behaves like a Nixie, as long as the current is limited (to 1 mA) you cannot damage it.

So it did light up. The picture shows a zero, middle and full indication, plus the look from behind the slotted metal window/support.

tune-a-lite gas filled tuning indicator  

Zero, middle, full extension.

And this is the wiring layout of my specific part. You may use it as a starting point in case you come across a similar indicator mounted on its base:

tune-a-lite wiring diagram
Wiring diagram.





26 June 2022

My Ham Messe Friedrichshafen 2022

Yeah! I managed to be in Friedrichshafen ham fair 2022! We traveled on "day -1" so we were not in a hurry and we could be at the gates at the opening time. Also we could enjoy the evening in town.

Like most of these events, even the self-proclaimed #1 in Europe is shrinking. The side with commercial operators and associations was full to 2/3rds (my estimation), while the private sellers area was one hangar and 1/3rd (my estimation). Hint for the future to organisers: unless we will fit in just one hangar, keep tables further away so it's easier to move around.

I walked from 9:45 to 17:00, had lunch and an extra rest-drink, and I saw every stand.

What I bought at the Ham Messe 2022.
What I bought at the Ham Messe 2022.

 What I bought this year:

  • 3 boards with red-coated Nixies (12 pieces OK)
  • another board with 9 Nixies
  • a "Tune-a-Lite" gas filled tuning indicator
  • a probably 2AP1 CRT (got it for free actually)
  • a board with eight HP 5082-7300 displays
  • a 40m band RTX "Frog Sounds", kit, assembled
  • 2 aero DME units because of their Minitrons (13 pieces OK of out 14)
  • machined-pin IC sockets (if they're the wrong length, I will cut them)

What I haven't bought:

  • "kostenlos" (free) Telefunken E 104 Kw receiver torn apart, more than 70 kg of stuff!
  • a large device with CRT, looking heavy too!
  • some cheap '80s oscilloscopes, in the range of 20€ each, for clock conversion
  • more aero DME units, now I regret

I found out that being even slightly social helps getting discounts, as I helped the Italian seller from whom I bought the red Nixies to interact with a German person that bought a (overpriced IMHO) calculator. I also realised I considered German sellers as unavailable to haggle the asking price, while it isn't true.

Our travel schedule did not leave me time to listen to the talks and because of the queue I couldn't greet Hans G0UPL as his booth.

As usual the organiser advertised next year dates: 23-24-25 June 2023. Of course nothing will be certain until next Spring, I dare to say.

17 June 2022

The James Knights Co. JKTO component - 2

What was inside.
What was inside.
From part 1, the circuit either had no output or produced a signal on 10 MHz. When I plugged in the oscilloscope, there was no output. The cylinder did warm up, always. Curious as I am, I couldn't resist but look inside .... to find a surprise!

A previous owner has replaced the innards with a 10.000 MHz crystal oscillator and used the heater to give it a form of ovenised stabilisation.

Ingenuity in pure state. A true hack. 

I wonder if the XTAL was actually cut for ovenised operation, otherwise it would have operated in a steep area of the f/T curve. Today we can achieve better performance with a GPSDO, and it will consume less power so the circuit has been dismantled giving: a 220/9V 300mA transformer, a large 10000 kHz XTAL to play with penning (lower freq) or grinding (higher freq), an experimental octal socket and an extra cool hand warmer. 

If I will ever get to it, if the XTAL can be pulled up to 10.100+ MHz it will be part of a valve transmitter on 30m. One more project in the to-do list!

Tightly build oscillator.
Tightly build oscillator.



14 June 2022

The James Knights Co. JKTO component - 1

In Marzaglia last May I picked up an anonymous circuit (click to zoom):

I was intrigued by the large metallic cylinder on the octal base. Just an AC power cord, on/off switch and a BNC output(?). The cylinder carries the message seen here:


Built in 1957 by the James Knights Company from Sandwich, Illinois. I don't know why, but I think it is amusing. It should be a 4.0 kHz (KC) oscillator, resonator, crystal and given these labels

Average Temperature 57°C, 6.3V thermostat.
Average Temperature 57°C, 6.3V thermostat.

the letter "T" in JKTO stands for "Temperature". Under the external cylinder there is another metallic cylinder and a wire wound at its base:

Since it did not smell or look burned, I powered it up. The cylinder got warm (not 57°C outside) and there was no output. At one point I picked a 10 MHz signal at the BNC on the frequency counter, but it did not convince me.

So probably the JKTO is/was the crystal element of a temperature controlled oscillator. Something could be broken in the circuit even if voltages are OK. I drew the circuit around the octal base and it might look like a Colpitts oscillator.

Since I don't need a OCXO at 4000 Hz I will look inside the cylinder, in part 2.


11 June 2022

Low VHF OIRT band - June 2022

Living far away from those Countries that use(d) the FM OIRT band (65.8-74 MHz), it is hard to tell if it is still used nowadays. But with the help of Sporadic-E propagation you should receive few stations if you are lucky.

The video below shows my unscientific band scan on 2022-06-10 at 1705Z, while the path had been open for an hour or so. The antenna was the humble 4 element beam on the balcony facing East.

The language sounds Russian. I patiently waited for advertisements and I could grab the word "Bielorussia", White Russia. I am not sure of the frequency because my 706 is getting old and a bit off, plus I might have heard several transmitters of the same station in different locations.

If you can identify the stations please leave a comment.



05 June 2022

Marzaglia maggio 2022

Marzaglia true spirit!
Marzaglia true spirit!
YES! I managed to me in Marzaglia meeting in May 2022. In my opinion it is THE place to be if you like old radio-electronics stuff. You need to get there by 7 a.m. (if you're a buyer/visitor). Bring a hat a sun lotion!

I brought home:

  • a couple of Nixies
  • two unknown large vacuum tubes
  • an unknown circuit
  • a valve radio skeleton to build my ideas on
  • a high-impedance headphone

Nothing really big because I've run out of space in the shack, even if some prices were quite right. Now planning for Friedrichshafen 2022. See you next May, Marzaglia!

A picture with vacuum tubes
Need valves?