29 November 2021

Dishwasher problem: solved

The fix to the dishwasher was as easy as buying a replacement cable guide.

I had to restore the broken white wire and add insulation to the other wire with visible copper. The cable guide slides underneath the body: just push it from outside, both sides already wrapped around the cables. Pay attention to position properly the nibble against the body so that the guide does not slide away when moving the lid.

The replacement part costs anywhere between 15 and 25€, depending on your source. I chose a real shop as we have quite a few in town. The human contact was invaluable: "yes, I have it, but I don't know where." According to the shop owner, it doesn't seem to be a common problem on those Whirlpool dishwashers. Or perhaps we have already used it past the design duration time, when a larger fault would force people to buy a new one rather than repair it.

Case closed.

23 November 2021

Dishwasher problem

English version is half-way down.

Domenica pomeriggio la lavastoviglie si è guastata. Si tratta di una Whirlpool WP75/1 da incasso, ma poco importa. Alla chiusura del portellone è scattato il salvavita anziché far partire la pompa di scarico.

In rete c'è il manuale di servizio, che mi ha guidato fino ai morsetti della pompa, che è OK. Lo stesso manuale indica come avviare la procedura di self-test, che ha diagnosticato: "problema allo scarico." Preziosissimo.

Quando ho avuto un momento di calma ho cercato qualche indizio sul più grande forum italiano (plcforum.it) soprattutto per capire come estrarre la scheda elettronica. Però in un post un utente diceva che il continuo movimento del portellone aveva spezzato i cavi di comando tra l'elettronica e gli attuatori.

Vuoi mica che ....

Vista del portellone da sotto.
Vista da sotto del portellone e fondo, dove passano i cavi.

ECCOLO! Devo dire che da qualche mese aprendo tutto il portellone si udiva un "clack!" proveniente dal basso, che non ho mai approfondito. Ora senza il pannello frontale è chiaro come la struttura metallica del portellone:

  1. fa scattare la plastica della (ex) canalina
  2. ha eroso la guaina protettiva dei cavi

 Se tutto va bene me la cavo sostituendo e proteggendo i cavi (sì, ma come?). Il divertimento continua.

 --------------- ENGLISH BELOW ---------------

On Sunday afternoon the dishwasher broke. It is a Whirlpool SP75/1, but that's not very relevant. When I closed the door, the circuit braker cut the electricity to the apartment, while it should have started the drain pump.

I found online the service manual that guided me to check the pump motor (OK) and start the self-test procedure. The result didn't surprise me: "draining failure."

Then I searched the largest Italian forum on these things how to get the electronics board out of its box. One user mentioned that the door movement had deteriorated and broken the control wiring between the board and motors/sensors.

Since the dishwasher is now more accessible from below I shot the picture seen above. GOTCHA!

I need to replace some wiring and protect them from further erosion. Need to find out how to do it. The fun continues.


11 November 2021

VHF elliptic filter retune with NanoVNA

Last month I bought a "surprise box" of surplus (radio)electronics. Well, the content was visually known and the box contained at least one wanted item. In between the heavy remaining stuff there were two VHF filters: one centered at 162 MHz and the second at 137 MHz.

At the time being I have no real uses for such a receive filter, but it is a very useful support to experiment with the NanoVNA, and use both of it ports. Educational!

After a little soldering and a bunch of coax adapters the setup looked like this:

Test setup and wiring.

I intended to retune the filter into the 2m HAM band. At the first glance the NanoVNA showed it is an elliptic bandpass filter. With a mix of awe and patience, adjusting the 4 screws following some fuzzy algorithm the bandbass centre moved down to 144 MHz:

Response at the desired frequency.

I am sure there must be a scientific way to tune these filters. I settled for: the two inner capacitors set the frequency, while the outer two match the impedance. With the obvious slight cross-interaction.

I stopped when I reached a 1.2:1 SWR at 144.300 MHz together with the least attenuation. The diagram matches the theory of elliptic filters. I am unsure of the best insertion loss of -4.06 dB (including the cable and coax adapters) as the label said -3 dB at 162 MHz. But I've got another one to retune and I will pay more attention.

By the way. These filters should be symmetrical, so the pattern should look the same if input and output are swapped.

View of 30 MHz span. A pretty elliptic response.

Which use for a receive VHF bandpass filter about 10 MHz wide in 2021? It should help to reduce out-of-band noise when looking for weak signals, but mind the 3-4 dB insertion loss. I will try it in front of an RTL-SDR and a wideband scanner to see/hear if it makes a difference.



06 November 2021

3R12 flashlight mania?

I have had a chance to visit the Zagreb Hrelic flea market once again in August 2021. The religious holiday, the Summer period and the forecasted high temperatures probably discouraged many sellers, with the result that many empty slots made it easier and faster to walk around.

I bought a working 3R12 metallic flashlight for 1,3€, battery included. It matches the one I bought in a previous visit since it carries the brand "CROATIA" and shares an uncommon colour tone (to be polite...).

Either I search for the wrong keywords, or these items are of no interest to collectors. Nevermind: I might be starting a collection of some sort myself. I do like the incandescent lamp glow and they can be modified to be run on Li-XX cells, let alone the obvious LED conversion.

So far I have two of them. Lacking proper model indications, I will assign my own names with reference to the place+year of purchase and sequence (in case I buy more in a single year).

Note: the white balance of the following pictures is correct, or the closest that shows their real colour in my lab room.

Left ZYWHR19.1. Right ZYWHR21.1. Front sides.

Left ZYWHR19.1. Right ZYWHR21.1. Back sides.

 ZYWHR19.1 couldn't be simpler as there is not even an on/off switch! I think it is meant to stay lit as long as the battery lasts. The Croatia brand logo is stamped on the back cover.

Inside of ZYWHR19.1 flashlight.

In contrast, ZYWHR21.1 has both a permanent on/off slider switch on the side and a N.O. button to emit flashes on top. The Croatia text brand is stamped on the front cover.

Inside of ZYWHR21.1 flashlight, with battery.

Both of them work.

02 November 2021

Surplus GPS "mouse" receiver and AVR input protection

Somewhere in the back of my mental to-do list I have a project with GPS, NMEA and precise clock (as in hh:mm:ss, not in the digital electronics sense). While I could use the ubiquitous Wi-Fi connection to query an NTP server, I prefer a solution that gets the timestamp without needing a user/password. As a bonus, it tells me the position as well.

A local HAM sells surplus GPS-only receivers (no GLONASS, no Galileo). They require 5V (later found that the range is 3.5V to 5.xV) and speak out NMEA at 4800 baud 8N1. The module inside is one of the GlobalSat series, probably BR-355S4. At least that model number provided a spot-on description of the pinout and existing wiring.

I wired it to a serial TTL/USB adapter and the dongle kept disconnecting from the computer. Both dongles I own. Weird. I needed to see what was going on, so I fired up the oscilloscope. Surprise! The NMEA data out of the surplus GPS dongle is RS-232, with -10/+10V! BTW, I couldn't get the data decoder of the Hantek scope to decode anything out of it.

In order to fix the signal levels I assembled a simple BJT adapter, which also inverts mark/space as in RS-232 logic "mark" (or logic "1") is the negative voltage and "space" (logic "0") is positive WRT ground. In TTL serial, "1" is 5V and "0" is 0V. This let me use the dongle and a GPS decoder software on the computer.

The adapter is 10k resistor into the base of any BJT (BC547, 2N2222, whatever). Emitter to ground. 4k7 from +5V to the collector. You can add a diode between ground and the base, with the cathode on the base side. The diode limits the Vbe reverse voltage to 0.7V. TTL output is at the collector.

The seller told me he could read the data with an Arduino just fine. How comes? Well, many Arduinos use an Atmel AVR ATmega328 processor, whose inputs are protected like this:

Atmel AVR input protection
From an Atmel datasheet, it's all inside the chip.


Those two diodes clip the -/+10V right into the AVR microprocessor and everything looks fine, but might break in the future.

Even if I will use the GPS mouse with an AVR micro, I will include the BC547 adapter.

31 October 2021

Topward 1302 multimeter fault?

The Topward 1302 multimeter performs as expected.

I left it plugged to AC outlet in OFF position for a few days after the last use. Then I had to move it away from the bench and I got a hot surprise: the transformer was hot enough I could not hold the finger on it.

WHAT? It was plugged but supposedly turned off, the display unlit too. Even if the ON/OFF switch would operate on the low-voltage circuits, why would an unloaded transformer get so hot?

Time to have a look at the actual circuit.

Good news. The PCB reports the product name as "TDM-104", which brings up a PDF manual from the Topward website. The schematic diagram confirms the transformer is always connected. Therefore also the diode bridge and capacitors are active when the AC cord is plugged in.

I probed voltages and they all seem fine. There is no ripple detectable with the multimeter.

Considering this product probably sat in the low-end side of the market, it was cheaper and safer to power off only the logic circuitry rather than the whole AC. Who cared about saving power in the 1980's?

06 September 2021

Laptop WiFi antenna coax

Laptop computers usually have the WiFi antenna in the upper border of the screen/cover. When you dismantle one down to the last bit, you end up with about 60 cm of coax, a super small coax plug connector and a J-pole looking antenna for 2.4 GHz (I haven't torn down a laptop with 5 GHz yet). There might/should actually be two antennas for reception diversity (improves signal quality).

Laptop wifi antenna.
Laptop computer WiFi antenna, coax and connector.

Since these antennas work at 2.4 GHz, I wanted to know if that coax is any better than our HAM radio cables. Not an easy task without markings on the cable, but starting from the connector name, probably U.FL, I found a ready-made cable with detailed specifications.

Well, the manufacturer declares 3.2 dB/metre at 2.4 GHz. As a comparison, the "tiny" RG-174 we use in HAM radio, looses "only" ~2.4 dB/metre at 2.4 GHz.

Conclusion: unless you need a WiFi antenna for some homemade module (like ESP8266/ESP32) or a very tiny coax, these things are not a keeper.

15 August 2021

Topward TFG-8104: the multimeter challenge

SPOILER ALERT. This post contains my take to the question asked in the previous post: "how to determine if it works with just a multimeter." If you want to play the same game please skip one post backwards and read my challenge.

Here we go.

The Topward TFG-8104 Function Generator powers up and generates enough RF that can be picked up by a nearby AM receiver. What else can be said with just a multimeter?

Since the output frequency can be reduced down to 0.1 Hz, all variations can be slowed down enough to be seen and appreciated even on a slow digital multimeter.

According to the manual it should produce a 20 Vpp signal. So I set:

  • lowest possible frequency, 0.1 Hz or so
  • square wave output
  • maximum amplitude
  • no DC offset
  • no modulation
  • no attenuation

The output alternates between -11.5V and +8V while it should be -10/+10V. When the amplitude is reduced readings collapse towards 0V and the asymmetry reduces, which is good. I can't think of further checks that are possible with a multimeter at the output connector. Well, there is output impedance.

So, even before checking the output at the oscilloscope, the asymmetry should mean something had happened to the output stage. A visual inspection of the circuit reveals signs of human intervention or fault on the output attenuator: some signal has tried to enter the function generator?!

While the box is open, the multimeter can check voltages. Both +15V and -15V are present. The circuit diagram mentions +/-20V but they are not marked on the board, nor there are accessible test points.

10 August 2021

Topward TFG-8104: does it work?

Does the Topward TFG-8104 Function Generator (from the previous post) work? Well, considering the source, the correct question is: "What is broken in the Topward TFG-8104 that I now own?"

A Topward Function Generator TFG-8104 in mint aesthetic conditions.

While it looks good aesthetically - if you ignore the heavy uniform yellowing of the plastic case -, some tools are needed to determine how the inner electronics are doing.

First test: power-up. It is quite likely that someone else has plugged it in before selling the device, so a power-on harm has probably already been done. There's always a bit of adrenaline shock when pushing the ON button, and the red LED lights up. Nothing else can be tested using eyes and nose.

Next check: does it generate something? This step is easily done with the recently acquired Hantek oscilloscope, that is now resting in its box out of sight and immediate use.

First of all I tried tuning the generated signal on an AM radio. It is easier to slowly sweep the generator over the tuned frequency than doing it in reverse. Got it. With and without modulation. Good sign, it's worth going further. Oscilloscope? Not yet.

I opted for a more fun approach: understand as much as possible with a simple multimeter.

How would you do that? What can you check?

You can contribute in the comments, and read how it I did in the next post.

05 August 2021

Topward multimeter and function generator - for repair

I got two Topward instruments at a fundraising event: they give away truly untested and almost certainly broken electronic devices that might have a second chance of life in a different form. Most of the hardware available is composed of old computers and parts from 10-20 years ago, with the occasional electric/electronic lab novelty.

Topward TFG-8104 and 1302
Topward TFG-8104 and 1302.

I brought home a Topward Function Generator TFG-8104 and a Topward Digital Multimeter 1302. Built in mid-1980's in Taiwan by a company that is still around in 2021, and has released scans of manuals and diagrams.

Both boxes are very light, even though they contain a real transformer. I initially thought they were empty! I would say that even back in 1980's these guys were entry level devices and according to the front panel wear, they haven't been used much.

The multimeter seems OK. Voltage readings are all a bit low. Resistance is in the right ballpark and I haven't measured a current, which is the faulty part in other multimeters around my lab.

The function generator, while quite obsolete today, it provides uncalibrated 0.1 Hz to 2 MHz sine/square/triangle and AM/FM modulation, which makes it suitable for driving optical transmitters or "transmitting" a signal to a nearby AM radio. I had a look at the schematic diagram before buying it and its use of standard components makes it very suitable for repairs, which is the fun part I am looking for.

30 July 2021

Winding QCX toroids as left-handed

The other day I resumed building the QCX transceiver kit I bought a couple of years ago in 2018. I had left it to the point of winding toroids.

All seemed fine for L4 (the red one), a bit loose but can be moved to match the PCB silkscreen.

Then comes the middle one from the output filter (L2) and ... I realised that winding toroids is another thing that left-handed people do differently! Loose ends of the winding end up on the opposite side vs. what is expected on the PCB and the inductor sits diagonally.

QCX toroids by a left-handed
QCX toroids by a left-handed builder.

When winding L1 I paid attention to the position of the wire and it matches better the silkscreen. Since I've had to remove L3 I might fix L2 as well so that these coils are perpendicular to each other as designed by Hans Summers.

Regardless the toroid is done with the right-hand or with the left-hand in the direction expected by the PCB, it really goes against the left-handed nature!

05 July 2021

From Tortona 2021, Nixie and CRT

The picture shows part of the spoils from Tortona June 2021 flea market.

On top is the original box containing a DP7-32 CRT. The letter "P" in the name means "long persistence, dual color". It has been opened to check the content presence and integrity, and that's all.

On the lower part of the picture is the display board from R&S receiver, showing a nice algorithm of hand routing PCB traces. Something had burned in this board because a resistor had been replaced but another one has a different [lower] impedance which causes lots of current to pass into a Nixie and the associated 74141 driver, which, in turn, is not functional any more. So the board gives 7 sockets, probably 6x 74141 and 5 good ZM1182 Nixies.

01 July 2021

Testing the equipment by testing other things

I found myself stuck in a testing loop while learning on-the-field how to use the Hantek oscilloscope: I was creating test cases for the oscilloscope that dubbed into test cases for accessories.

See this. The scope has a -3dB 150 MHz bandwidth, so I picked up my 14 MHz Marker Generator and tried to visualise the output terminated on a 50 ohm dummy load. The marker generator creates a short impulse in the time domain, which results in many peaks in the frequency domain. The short square impulse should be 10 ns long inside a period T of 71 ns at 14 MHz.

Grab all the probes in the lab and see which one renders better the impulse train. Don't forget to set the probe to 10x.

Probes were tested in this order:

  1. P6100, 100 MHz
  2. P6100, 100 MHz
  3. PK-8150, 150 MHz
  4. PP150B 150 MHz, Hantek
  5. 88025, 250 MHz, Greenpar

The first stitch below shows 4 periods of the test signal, while the second image is a zoom-in on one impulse.

According to the screenshots (wow!), all five probes behave almost in the same way! Even the input square wave show ringing.

Well, I do want to see an impulse train so I will either change the XTAL in the marker generator or use the arbitrary function generator embedded into the scope. Or both.

When time allows I will repeat the experiment with a 200 MHz Siglent oscilloscope (not mine).


29 June 2021

Hantek DSO2D15 first impressions

Holding the box with one hand.
The delivery of the Hantek DSO2D15 oscilloscope was really fast ("EU Direct Shipping" by AliExpress from Spain), just 3 days.

The very first positive remark is the overall weight: 2.3 kg, vs 13+ kg of what I owned until last week.

Once you power it up it feels like interacting with a computer through a "usual" oscilloscope interface with menus, settings, updates, resets, ... I had to read the manual to get a feeling of what the device is capable of.

For example, take "trigger modes": Edge, Pulse, Video, Slope, Overtime, Window, Pattern, Interval, Under Amp, UART, LIN, CAN, SPI, IIC. Some of them have options, too. Well, once you implement trigger in software, imagination is the limit. And you need to read the skimpy manual to understand how they operate. In most cases the default "Edge" and the trigger level knob do the job.

Another positive note is the absence of a fan. If you disable key-press beeps, the DSO2D15 is totally silent. But if you press the "Default Setup" button, it will beep again.

25 June 2021

Gone is the Tektronix oscilloscope, welcome Hantek

For a number of reasons I needed to relocate the Tektronix 7000 series oscilloscope: large, heavy, noisy, at times unreliable, stuck in an uncomfortable operating position. It also needed some thorough inspection, which would have meant being unable to use my lab desk until the work would have been over.

So I found a new owner, with other Tek 7000's, and it will keep showing signals.

I cannot live without an oscilloscope and the market is now offering compact and powerful devices, probably as reliable as the old unserviced Tek, at very affordable prices. I studied the market, read reviews on eevblog and finally clicked the trigger for a Hantek DSO2D15.

It is a digital storage oscilloscope, 2 channels, 150 MHz and includes an arbitrary function generator up to 25 MHz. This is quite an upgrade with almost twice the bandwidth, storage function, protocol decoding and a signal generator.

I must admit that when you look at the current offer of DSO's, even if you set a budget, there is too much choice! I opted for a product that has received firmware upgrades, has more than 4k sampling memory (as cheaper Rigol's do) and has no big issues or community-documented design flaws.

Time will tell.

23 June 2021

Tortona, 20 giugno 2021

Last Sunday I attended a ham/electronics flea market after 9 months of lockdown, with the open-air location taming the smell of dust, vintage gear (materials, components, grease, ...) and the occasional (unintended) lack of personal hygiene caused by the very hot days. The whole feeling was very positive!

A side view of the location. We were trying the optical QSO on the grass.

For the second consecutive year I have been able to attend the fair in Tortona, which is held in a parking lot under the June Sun. There were approx. 20 tables, with lots of space for chatting, moving around, digging into the boxes. I could recognise a couple of "new entries" amongst exhibitors, that are heirs of SK, which sold some stuff unseen in previous editions.

I went there to meet friends, attempt an optical QSO with Mauro IK1WVQ and deliver the heavy and bulky Tektronix R7603 oscilloscope.

I came home with a N.I.B. Philips DP7-32 CRT (dual color, dual persistence), the Nixie display board from a Rohde&Schwarz EK47 receiver (7 sockets for six ZM1182 tubes, five of which still working), a new 250 MHz probe (for the no-oscilloscope that I have :)), some coax connectors, few large VFD and a board with LED displays.

It was not possible to complete the QSO because my optical transmitter used inefficient LEDs as opposed to Mauro's blindingly TX head, and I had not built the FM modulator. Too bad, because we had a whopping 100 metres in full daylight and his setup had a lot of margin!

16 June 2021

10 GHz WBFM activity 2021

The yearly 10 GHz "old mode" FM Contest was last Saturday. That is the only chance to play with WBFM, Gunnplexer, HB100 and show your presence in some official documents.

LNB over a tree.
Three operators managed to arrange three points with mutual visibility in NW Italy, max distance above 100 km. My setup was HB100 on the TX side and LNB+RTLSDR+laptop on RX side.

Since my spot was 15 minutes hike from the car I had to minimize the weight: everything you see in the pictures was carried on my shoulders (except for the tree).

Having skipped the year 2020 contest because of the pandemic, we operators have forgotten some details of our setup, like the correct orientation to match TX and RX polarisation which caused some confusion when looking for each other when we knew we had to be heard.

In the end we all managed to work each other without using the parabolic dish boost, which is a good achievement thanks to modern sensitive LNB. The 2 MHz waterfall view does help to locate unstable transmitters, even if my HB100 constantly under the Sun moved just some 200 kHz on 2 hours.

As planned I worked the two operators at 35 km and 99 km. Once you sort out the correct polarisation, HB100, Gunn and LNB antenna patterns are wide enough to allow very comfortable beaming, to the point that my LNB was waving at the wind on the branch and there was no QSB at all.

TX head on a lightweight tripod.

It was a pleasant afternoon on the field. Too bad the season already calls for family/vacation trips and it is hard to involve more operators.

PS: my distance record on the same setup is 144 km from the same contest in 2019.

10 June 2021

Early digital scale with Nixie tubes

Italiana Macchi Mach 55
Until 6 months ago I had no idea that in early 1970's existed digital scales that displayed information with Nixie tubes. That was the beginning of the transition from "analog" to digital weighting at retail shops.

In mid-2020 I saw one scale for sale on eBay which sprung my curiosity. It was a Italiana Macchi model "Mach 55". Not much has been written on Internet about these quite rare devices, and the most comprehensive discussion so far was on tubeclockdb forum from year 2012-2016.

Those devices were probably obsoleted in a decade and their resilience to time travels was heavily affected by their inherent weight.

Considering that a Mach 55 contains about 30 Nixie tubes, I had to own one: the hunt began, for this or similar models (which consists of setting up alerts on online marketplaces). One condition turned out to be compulsory: the scale for sale has to be picked up in person because it weights 30 kg or more!

By the way, Italiana Macchi (Italian) and Bizerba (German) are still active in the weighting sector!

24 May 2021

La forza di Lorentz alla maturità (esperimento)

Il testo in italiano sull'esperimento con la forza di Lorentz è poco più sotto.

I have been contacted by a father to help his son with the high-school graduation project (you know, the school exam when you are 18-19 years old). I will explain few options to visualize the Lorentz force in my native language.


Bene ragazzi. La teoria sulla forza di Lorentz l'avete studiata a scuola. Se siete qui è perché vi servono delle indicazioni su come poterla visualizzare su un sistema casalingo, giusto?

Fino al primo decennio degli anni 2000 in tutte le case c'era un ingombrante generatore di fascio di elettroni: la televisione a tubo catodico (mi piace immaginarlo detto con la voce di Fantozzi). Anche detto TV CRT dall'acronimo inglese. Gli ultimi esemplari sono stati venduti nel 2013, dopodiché l'avvento della TV digitale terrestre (DTT), gli scomodi decoder e poi ancora il passaggio al DTT-4k ne hanno decretato l'inesorabile tramonto e scomparsa dalle nostre case.

Oggi - 2021 - queste TV si trovano nel mercato dell'usato al prezzo delle patate, se non gratis, quindi sono ottime per realizzare un mini-laboratorio casalingo in assoluta economia. Consiglio di puntare ai modelli più piccoli, che sono più maneggevoli.

Altri oggetti con tubo catodico erano i monitor per computer e gli oscilloscopi, il cui prezzo è più elevato di una TV, anche se si tratta di pezzi tecnologicamente obsoleti. Ho in serbo una variante tascabile di questo esperimento, documentata al fondo.


Problema 1: serve un segnale. Accendendo oggi una TV analogica si vedrà solo il rumore di fondo sul quale la forza di Lorentz sarà poco evidente. Ma ci sono alcune possibilità:

  1. le vecchie TV accettano un segnale a radiofrequenza analogico (introvabile);
  2. spesso hanno una presa SCART per il collegamento al videoregistratore/decoder;
  3. hanno un menù in sovra-impressione.

Scartata la 1), per la 2) potete cercare in casa un oggetto che genera un segnale video composito da iniettare nella SCART come le prime macchine fotografiche o videocamere digitali compatte (attenzione che serve il cavo!). La libreria TVout di Arduino può fare al caso vostro se ve la sentite di pacioccare con l'elettronica (o se aggiunge valore al vostro elaborato).

La soluzione 3) invece mi piace tantissimo: basta avere il telecomando della TV per attivare la visualizzazione di un quadro con delle comode informazioni statiche.

Problema 2: invocare Lorentz e spostare il fascio.

L'immagine su uno schermo CRT viene costruita colpendo lo schermo con un fascio di elettroni, riga dopo riga, pixel dopo pixel. A differenza dell'esperimento di laboratorio dove il fascio giace su una semiretta, qui si sposta velocemente per costruire l'immagine, ma poco importa.

Portando un bel magnete davanti al vetro dello schermo l'immagine dovrebbe deformarsi. Se la TV è a colori, cambieranno anche i colori dell'area interessata. Su uno schermo piccolo il fascio avrà meno energia e sarà più sensibile alla forza di Lorentz. Provate anche a mettere il magnete sul fianco dello schermo, e a girarlo.

Attenzione. Il tubo a raggi catodici funziona con tensioni letali quindi sconsiglio vivissimamente di aprire l'oggetto per posizionare il magnete in altri punti del tubo. Inoltre i condensatori restano carichi per ore dopo aver staccato la spina, quindi non aprite la TV appena spenta. Uomo avvisato, mezzo salvato.


Ammesso di trovarne ancora in giro, il monitor per computer a tubo catodico ha un ingresso VGA (connettore blu) e può essere usato anche con i computer attuali, se dotati di uscita VGA. Mal che vada si dovrà costruire un semplice circuito con Arduino che visualizzi una informazione statica.


Un oscilloscopio funzionate, se impostato correttamente, visualizza una linea orizzontale: il fascio di elettroni viene fatto scorrere da sinistra a destra in continuazione, sull'asse delle ascisse. Se si applica un segnale all'ingresso lo strumento inizia a disegnarlo variando la posizione del fascio sull'asse delle ordinate, come se usassimo una matita su un foglio. Quindi un oscilloscopio "a riposo" è sufficiente a giocare con la forza di Lorentz.

Tra l'altro la traccia si può spostare sull'asse Y per studiare gli effetti della forza nelle zone periferiche.


Ed eccoci alla fine, con la variante tascabile.

Negli anni 1970 sono stati introdotti sul mercato dei display funzionanti con lo stesso principio dei tubi catodici: un filamento riscaldato che emette elettroni all'interno di un involucro di vetro con il vuoto spinto e un target con fosfori su cui far atterrare gli elettroni in modo da illuminarlo. A differenza del tubo catodico dove il fascio di elettroni viene deviato per via elettrostatica o magnetica, nei display VFD è l'area da illuminare ad attrarre a sé le nostre particelle negative.

Bene, siccome ci sono elettroni liberi attirati verso una zona ben delimitata, è possibile deviarli con la forza di Lorentz producendo un effetto visibile.

Ma quale dispositivo ha queste caratteristiche? Una calcolatrice tascabile! Ed è anche alimentata a batterie. Come riconoscerla?

Una calcolatrice con display VFD visualizza le cifre di colore verde-azzurro su sfondo nero (buio). No rosso, no grigio. Anche in questo caso lo sperimentatore dovrà rivolgersi al mercato dell'usato o chiedere ai conoscenti se ne hanno una nel cassetto. Qui avevo fotografato una calcolatrice con display VFD.

Per visualizzare l'effetto della forza di Lorentz sulla calcolatrice io farei in modo di visualizzare il numero 8 su tutte le cifre, posizionando o spostando il magnete sopra al display.

I display VFD sono stati utilizzati anche nei videoregistratori, negli impianti stereo, nei forni a microonde, ... si riconoscono perché hanno uno o più sottilissimi fili orizzontali che coprono tutta la lunghezza del display (il filamento). Ovviamente il dispositivo deve essere acceso per "subire" la forza di Lorentz.

20 May 2021

Kit Oscilloscope Clock 8SJ31J review - 4

As these kits have re-appeared on eBay in May 2021, more questions have been posted online. I think the input current consumption is not documented, or at least confirmed. So here you go.

My one and only kit of these, with a 2BP1 CRT (spec'ed at 650 mA filament), draws less than 800 mA at 13 Vdc input.

The voltage across the neon bulb is 56 Vdc.


24 April 2021

SEPTANIX Display from JRC

On the popular auctions site I spotted a never seen before(*) calculator that caught my attention. 

(*) most calculators stay on the site forever and nobody is interested in them (or they are proposed at a stellar price), so they fill up my saved searches and I skip all of them.

Even if a bit blurry, a picture of the display showed a think dense grid on top of the usual 7 segment + decimal point layout. With a bit of searching I could confirm it was a rebranded Unitrex 1202M, which came with a multi-digit gas-filled display but not a Burroughs panaplex.

Within a week I could get the calculator on my desk and reach the display:

septanix display by JRC
SEPTANIX Display with visible anode grid.

That is a (probably uncommon) SEPTANIX display made in Japan by JRC. It is a multi-digit version of other Japanese 7-segment neon tubes, with a visible anode grid as opposed to the more discreet arrangement in Panaplex display. Also vertical segments in the SEPTANIX have the same length, while in Panaplex lower ones are a bit taller.

The calculator itself is using negative voltages for the logic, so a conversion into a clock would require some extra efforts. Nevertheless I have one more display in my collection.

12 April 2021

Breadboard socket for valve

B9A on breadboard
B9A on breadboard.
To me, playing with vacuum tubes means a lot of trial and error to get where I would like to. Everything is laid on the desk and after a few weeks of fiddling I didn't want anymore to mix the hot soldering iron with the other equipment, cables and so on.

The solution would have been to make a socket so that the tube could be mounted on the solderless breadboard. Nothing easier!

I picked a B9A socket for PCB mount (from the lot that I bought) and soldered terminals that fit the solderless board. I used leftovers from through-hole component leads: they are long and flexible enough to reach the nearest hole.

B9A on breadboard
Filament connection is left out.

Since the filament requires a good amount of current I lifted pins 4&5 and soldered a short length of thick cable. This also allows to have three pins on one side of the breadboard and four on the other side, without overlaps.

10 April 2021

Memory upgrade to HP laptop (15-g005nl)

When the lockdown started one year ago I was fortunate to have enough computers at home to allow simultaneous distance learning and remote working. All home laptops run Ubuntu Linux, that makes them breath with just 4 GB RAM and 5400 rpm spinning hard-disks (that's slow in 2021 terms), plus all updates are under my sole control.

Those computers are about 10 years old and serve 8+ hours a day without a glicth. Almost.

After one year of extensive use the HP laptop needed an upgrade (or a reinstall). The HP 15-g005nl can be upgraded to 8 GB RAM. Linux lshw command shows that there are two slots, BIOS is not so detailed. Still in doubt: do I need a second 4 GB SODIMM, or a single 8 GB module?

I looked online for a teardown video of this model or alike and it confirmed my suspicion: there is only one RAM slot! The video is a lifesaver because, amongst other things, it tells you of 2+1 hidden screws (2 become visible when the DVD reader is removed, the third is under a rubber pad).

Before buying the memory bank I worked my way down to the memory slot, just to confirm that this laptop was NOT made with servicing in mind. Even replacing the HDD requires removing the keyboard and the hand-rest frame.

The picture below shows the main board lifted from the back shell, as if looking at the laptop from the front/normal use position. You need to remove at least 6 cables and lots of screws to get here.

I was so dedicated to the RAM upgrade that I did not notice the backup battery until I reviewed the picture. Hopefully it is a rechargeable and I will not have to deep dive again in the HP anytime soon.

The SODIMM was finally replaced and thrashing (swapping fast memory to the slow hard disk back and forth) has stopped. 

27 March 2021

Help with unknown vacuum tube / valve

In the assortment of vacuum tubes I bought there are few items that look all the same and all without marking. What is worse is that they do not look like any of the almost 40 kinds of tubes that do have a marking!

So I ask my readers for help to give a name to these B9A/noval tubes. All nine pins are connected. There are two identical sections and filament runs happy at 6.3V. The anode/plate is probably connected to pin 9.

Is it a double triode? A double pentode?

Please suggest your answer in the comments or via email to ik1zyw at yahoo.com. Thanks!

"Side" view.
"Front" view.


23 March 2021

Things got out of control with the all-valve TX project

The more I read about valve (vacuum tube) oscillators and transmitters, the more I understand the inner workings of those relatively simple circuits. Once I realised that a triode-heptode or triode-pentode tube can work both as oscillator and "PA", I concentrated my efforts around a PCL-805. Until my own hands presented me the nest shown in the picture.

One tube CW transmitter (tentative)

Alright. I will take one step back and beam my first venture into the world of vacuum tubes towards two separate stages: the crystal oscillator and the power amplifier. I do have enough tubes and sockets. What I am missing is a decent assortment of high voltage capacitors and high wattage resistors! Let alone RFC!

18 March 2021

Valve sockets!

Now that I got a hundred vacuum tubes and an idea of a project, I needed some matching sockets. Without in-person flea markets, I turned to eBay and with a bit of luck I spotted a lot of assorted used sockets that nobody wanted. I think I got about 100 pieces, which makes it 0.15€ each once shipping is taken into account. Of course I will never need all of them so I probably ended up overpaying just one socket :)

Besides lots of octal, loctal and noval sockets in different shapes, there were few interesting items, proof of a time when electronics were booming and everyone tried to impose their standard (or a technological lock-in).

I've collected them for a group picture (I forgot the ubiquitous *octal).

Numbers represent the # of contact points.

Needless to say that I have no tubes that fit most of these, and some tubes that have no matching socket.

If anyone wishes to give them a name, please do so in the Comments!



08 March 2021

EF80, EF183, EF184 XTAL oscillator

(This post has been deleted by Blogger because someone flagged it as malicious. No idea why. Then they changed their mind and returned the post in Draft state.)

While passing through forgotten boxes of forgotten components looking for some interesting TTL/CMOS ICs for an artistic project, I found two large XTALs for 2.0971 MHz. I think they were meant for some valve circuit, so I looked for an oscillator circuit with one of the tubes I have most: EF80.

Using the DIY B9A base I rigged together the most promising circuit I found online of a CW TX, with resistors "close enough" to quoted values. The ugly result is shown in the picture as well as the circuit diagram.

Diagram and test circuit.

I fed 6Vdc to the filament and 150V HT and it did oscillate. Cool. Time for experiments.

I could reduce filament to 5V at 300mA and everything was fine.

I could reduce the HT to 30V and it didn't stop oscillating, alas the signal picked up by the nearby receiver was much weaker. I couldn't go with a lower HT with today test setup.

Then, while I was at it, I tested all EF80 valves I had already separated from the rest. Not happy with the result, I checked through the list and found out that EF183 and EF184 pentodes are pin-compatible with EF80, so they got tested too. I could even re-stamp a few tubes that had become anonymous.

The result of a couple of hours of fiddling with on/off switched of power supplies is that I have 28 working pentodes waiting to be used in a real transmitter.

Now: VFO or XTAL?

28 February 2021

Checking a static discharge trick with the NanoVNA at UHF

Today I received a question about a post from 8 years ago: how to tame the static build-up in a GP or collinear antenna. Back then I had done some research and concluded that a high value non-inductive resistor could be used, especially at low TX power or RX only.

Now that I own a NanoVNA I can confirm that the impedance is not affected. I quickly reused the LoRA antenna from last Summer, which turned out to be resonating at 750 MHz (if we trust the NanoVNA). This is the curve on a 300 MHz span (well, it stops at 900 MHz on the right):

Just the "open" GP antenna.

Then I added a 10 kohm 1/4W resistor across radiator and radials and the result is almost identical (the marker has moved):

After adding the 10kR across. See next picture.

The very (quick && dirty) antenna under test.

So, if we trust the NanoVNA at the upper limit of its working range, this trick does not compromise the impedance. I might also have been lucky on the single experiment I've made, so do your measurements first.

09 February 2021

Sorting a box of vacuum tubes

So, there comes the time you become the owner of a shoebox full of almost-no-value one-hundredish vacuum tubes and you need to sort them out. I do want to use (some of) them, not immediately. I also have no facilities to keep them sorted in a logical way, so I had to do it efficiently.

First of all I separated the magic eyes, since they can be easily recognised and tested.

Then ... I simply transferred the tubes one-by-one from one box to another and typed the part number in a spreadsheet, when still visible. Those tubes without a usable marking have been compared physically & visually to others and some got a supposed part number with a marker pen. A dozen were left anonymous, separated from the rest, for other projects.

Once all the names are in a spreadsheet just create a pivot table to see how many different types you've got, and how many each.

Now go back online and look for projects with those tubes :)

05 February 2021

Freeform two transistor LED blinker completed

My first freeform electronics sculpture is finished! Visually it is a replica of Mohit's work with a change in the diagram (one resistor less). The LED blinks approx once every second, and blinks become more frequent as the available voltage decreases towards LED's forward voltage.

The 1uF capacitor that sets the frequency is charged though the leakage current passing through the LED. In ultrabright clear-case LEDs you see a faint emission in this phase (dark room implied). Then you are blinded by the following pulse :)

I also added a 1F 5.5V supercapacitor that keep the blinker running for more than 12 hours with a white LED. If a red LED is used, it lasts 24 hours.

If a solar panel is added in parallel to this circuit to top up the supercap during daylight (add a diode to prevent discharge back to the panel), basically it lasts forever.

The stand-by current is fractions of microamperes [uA] and the average lies around 30 uA (calculated). Not easy to measure, huh!

The hardest parts are to get straight wires and component leads and to keep everything in place when soldering when both your hands are needed to hold the iron and the solder. Then you need to make clean solder joints.

While is may seem a useless exercise, freeform electronics forces you to think out of the box even if you are reproducing someone else's work. Give it a try.

PS: working circuits impress 10x fold the audience.

03 February 2021

Freeform two transistor LED blinker

Two transistor LED blinker, construction in progress.
Intermediate result.
As announced, I wanted to try Freeform Electronics Art. My last PCB designs for clocks were meant to be artistic so that they could be displayed to the public even without a case. Freeform Electronics uses no PCB: it's all self-standing and built "in the air".

I chose to replicate Mohit Bhoite's two-transistor LED flasher because of its simplicity and I had all parts already at home.

I deliberately chose to ignore some of his planning and construction techniques (shared in videos featuring him and his art) so that I would learn more of and during the process. Moreover this blinker is very simple, forgiving and lightweight and almost any building technique should work.

Two transistor LED blinker, construction in progress.
Another view, next to an SD/TF card for size comparison.

Pictures show an intermediate result with 2N5401, BC347C and three resistors. I am half-way to the end, component-count-wise.



31 January 2021

QYT KT-8900 12V on the mic socket and more noise

Alright. I found a spot to get 12V inside the QYT KT-8900, which is pin 5 of the audio amplifier IC. It is located near the entrance of the DC cord. I bought it to pin 1 of the MIC socket, which is unused in this radio: that's where you can push the audio output for Packet/APRS (the documented mods mentioned in the previous post).

The result is that now the buzz is present both on transmit AND receive.

The oscilloscope explained everything. The wireless dongle generates digital spikes on its power supply line. Since it is designed to operate from the internal battery, nobody cares if there are 100mV 500Hz spikes.

Solution: add 47-100uF electrolytic capacitor on the power supply line of the dongle.

Thirty seconds after finding the solution I burnt the receiving dongle. (Hint: capacitors retain the charge.) Actually the LDO inside went short circuit. I removed and bypassed it, but while 3.3V to the dongle do power it up, it does not pair with the headset and shuts down after 10 seconds.

Considering that I am not operating /M on a daily basis as in pre-pandemic times I am now thinking of possible evolutions or restoring the same setup.

30 January 2021

Better 8V on QYT KT-8900 microphone socket

I want to simplify my mobile radio operations so that I do not have to switch on the wireless microphone receiver. In the car I use a [insert_your_adjective(s)_here] QYT KT-8900 VHF/UHF mobile, that I have described in the past. It outputs 8V on the microphone socket, so I wanted to use it to power the microphone receiver dongle bypassing the embedded battery.

I built a simple adapter with a 78L05 + the required capacitors + dropping diodes, but the voltage drop was too high when the dongle was powered through it. Then I checked online and a mod is suggested in order to get real 8V on pin 2 of the mike socket, which translates in adding a 10 ohm resistor between a point in the front panel board and the socket pin, like this:

How it is supposed to be done. But adds noise!

Well, I do get stable 8V out of the radio body, but now the transmitted audio has a strong buzz. The dongle drains about 35 mA at 4V.

I think the transceiver does not like when someone messes with the 8V line, as I had discovered when tracing the buzz on the received audio (see the first link above). Next move is to bring 12V out through the unused "pin 1" on the microphone socket. Fingers crossed (that nothing melts in the process!).

25 January 2021

Getting ready for some freeform electronics

Three sizes of brass wire.

Through hack-a-day stories and events I got to know Mohit Bhoite's freeform electronic projects (link goes to his Twitter page). He builds working circuits "in the air" as a form of visual art. He is not the only one on Earth publishing freeform electronics circuits, but he has a tendency to build "squared" shapes that better suit my current artistic abilities.

In order to get started I needed the wires that act both as conductors and supports. Mohit uses brass wires in 20 and 22 AWG sizes. That translates to 0.8 mm and 1.0 mm diameter in metric units. While I was at it I also bought 1.5 mm diameter brass, just in case I will need to support something heavier that a couple of ICs and a dozen of LEDs.

My first project will be a solar-powered super-capacitor LED flasher that promises to recharge even during short Winter days (already gone-by for Winter 2020-21).



22 January 2021

DMM Input Impedance

I am working on a very low power LED blinker and the need of measuring currents in the range of microamps (uA) has opened me new challenges in the world of lab instruments.

Once you can measure milliVolts, just let the uA current flow into a large-ish resistor (100k, 1M) and measure the voltage across it and apply Ohm's law. Then you remember that your instrument has a finite impedance and it might influence your readings. So, how to estimate a DMM/DVM input impedance? Or how to confirm what is written in the accompanying leaflet? 

In this experiment I have been using an ANENG AN8002 DMM, because it is small and uses 2xAAA cells instead of a PP3 9V battery. It should have 10 Mohm input impedance in DC Volts range.

In order to confirm or calculate the desired value, build a resistive voltage divider say, with a 10 Mohm resistor like this:

+V ------- 10 Mohm -------- DVM -------- GND

Let +V be a known voltage value (measure it in advance using the same DVM) and read what the DVM display says. The inverse formula says:

Rdvm = Vdvm*10 / (V - Vdvm)  [Mohm]

Using my values I get 11 Mohm of input impedance in the >2V range. I do have measured the 10 Mohm resistor with several instruments and they all agree on its value.

Since the AN8002 is an autoranging instrument, the input impedance may change with the auto-set range. And it does indeed, but it is always around 10 Mohm.

Do I need a microAmp meter? Maybe not yet.

Oh, by the way, the blinker takes 0.4 uA (that's 400 nA) when OFF. I have repeated the math several times and taken several measurements in different ways, but they do agree! It runs 24 hours on a 1F 5V capacitor ... if you don't mess with it with a DMM :)