LED ERP vs. current

While searching for optical RX-TX systems that would carry voice, I stumbled across an article of the only Italian HAM (AFAIK) that has documented optical experiments, I4VIL. He mentions that LED efficiency at some point decreases even if the current flow increases (in Italian).

I managed to replicate his observations using my MCW TX and the simple OPT301 RX, just the detector without the active bandpass filter. I monitored the received signal strength by feeding the detector output to my laptop MIC input. Spectran software did the measurements. The TX was aimed at the ceiling, as well as the OPT301, without any lens.

I monitored the total circuit input current, not the one flowing into the LED. The MCW TX draws about 6 mA @12V in stand-by, so you can figure out how much current really flows into the LED. But at 100mA average when transmitting, 6mA are negligible.


The Y-axis shows dB reading on Spectran for the 488 Hz line, relative to the first dot (17 mA); the X-axis is the total current flowing into the circuit.

It is obvious that the emitted light reaches the maximum power somewhere between 80 and 100 mA. Going further just heats the diode.

Delta ERP is +17 dB. That would be almost 3 S-points in RF terms.

Interesting, isn't it?

HP iPAQ h5450 dead/flat battery

So, you forgot to recharge your iPAQ h5400 PPC handheld for a week and it doesn't power up anymore? Not even if you leave it plugged to the external supply for days?

Might be too late, or not.

First of all, if you read this before your battery has flatten out, go to the HP website ("software and drivers"), download and install firmware upgrades if you've never done it. There is a fix for the flat battery that doesn't get recharged.

If your h5450 has already passed away without upgrades, you need to find another h5450 (h5400 series), do the patching and then recharge your battery on it. Most probably in less than 24h you'll have your h5450 working again.

If you have a h5500 series handheld (h5550 for example), HP has not released the battery charging patch, while it seems to be affected as well (real world experience). Since h5400 and h5500 batteries are the same, let a patched h5400 recharge your battery and you'll be back on track.

Important: always do these chargings with direct connection to the external power supply. Do not use the USB charging feature.

Good luck.

MCW optical TX - first test

Finally, more than a year after the idea of an optical RX-TX system was born, I have been able to send (pseudo)information through modulated light.

Instead of a laser pointer I used a 10mm clear-case hi-bri bulk red LED driven at 10mA average. The 488Hz modulation is obtained from a 4.00 MHz XTAL divided by 8192 with a 4060 CMOS chip, keying is achieved with the reset function of the chip as of this diagram:


The RX was the prototype of the future optical RX on experimenter's board, with an OPT202 sensor, without any post-amplification. Circuit per datasheet, single supply operation. Output fed directly into laptop MIC input. At first attempts I did not modify OPT202 feedback resistor, but then a 3M3 was added in series and the signal output increased while total bandwidth was reduced. No amplification lens was used.

The test was conducted in a lighted/dark room with the receiver next to the laptop screen with retroillumination. The TX was aimed at the ceiling and Spectran software in QRSS3 mode produced the following output:


I could not broadcast the received audio in the room.
I will soon try building bounce using two adjacent windows.
I did not reduce LED drive to test RX sensitivity.
I will try to overdrive the LED to get the max power out (destructive test).
I might then add an external LED drive control.

The overall distance was about 6 metres: TX to ceiling + ceiling to RX. Not a DX record, but a good starting point!

The Joule Thief

I had it ready for months, but never finalized. This Xmas 2007 we needed a simple way to light the Nativity scene, so I took the chance to get it working.

I used the joule thief in its simplest form, as probably published November 1999 issue of EPE (Everyday Practical Electronics). It's a NPN transistor, one resistor, a bifilar transformer and a LED (any color). I found the diagram on the web and worked immediately.

Lighted using a dead saline battery.

Lighted using a dead saline battery. No flash.

The Nativity scene, with room lights ON.

The Nativity scene, with room lights OFF.

How did I wind the transformer? I saw many questions about it around the web. First of all, you need some luck since you're probably winding it on a core with unknown characteristics. I took some length of enameled copper wire (2 metres), bent in half and wound about 8 turns through the core. Windings are spread about on 80% of the core. The core is a ferrite bead recovered from WhoKnowsWhereLand. Try it first with a fresh battery. If it doesn't lit, add some turns and try again. If at 20 turns you still don't get light, then:

• check your LED is wired properly
• change the core

Don't forget to scrape off the insulation from the enameled wire at solder points!

Before someone asks, no, I have no spare cores to send around.

With the saline battery my joule thief produced noticeable light (in darkness) for 36 uninterrupted hours. I will measure how much current it draws at the next battery change.

Driving a LED from TTL/CMOS

It's the first thing you learn about TTL/CMOS chips: they cannot (should not) drive a LED directly. A transistor driver is needed, and is shown at the right in its simplest form.
When the gate output goes high the BJT saturates and current flows through your LED. A limiting resistor is needed to control the LED current.

The table below shows R values vs supply voltage vs LED current.

Vcc	R ohm	R_ohm	R_ohm
3,6	45	30	22,5
4,8	105	70	52,5
6	165	110	82,5
7,2	225	150	112,5
8,4	285	190	142,5
9,6	345	230	172,5
10,8	405	270	202,5
12	465	310	232,5
13,2	525	350	262,5
I_LED [A]	0,02	0,03	0,04
V_LED [V]	2	2	2
V_CE [V]	0,7	0,7	0,7

I will need these in my hi-brightness red LED optical MCW transmitter.

CCD as infrared sensors

This might be obvious to many of you, but it wasn't to me.

CCD, the digital camera sensor, is sensitive to infrared light, such as the one emitted by all kinds of remote controls (if you don't consider prehistoric ultrasound models).

Webcams use CCD as well, so here's a screenshot taken with an old one to prove that it actually works.

The IR generator is a handheld device running the pre-installed IR remote control software.
You can see the lighted IR-LED, a reflection on the left (double-bounce?) and another spot near the top border of the picture.

You can reproduce this experiment aiming a remote control to a digital camera lens. Then press a key while looking at the display.

Perhaps a simple way to explain some physics to kids?

Antenne Yagi per usi portatili - Q&A

Data la visibilita' che ha questo blog tramite alcuni motori di ricerca (e' una semplice constatazione), postero' qui in forma anonima le domande e risposte che ricevo sul mio articolo pubblicato su Radiokit Elettronica di novembre 2007.

La prima domanda riguarda il fissaggio degli elementi: c'e' un accorgimento particolare per evitare che ruotino attorno all'asse del boom?

No. L'incastro e' molto stretto, al punto che e' quasi impossibile inserire l'elemento a scatto senza farsi male, ma si deve farlo scorrere sul boom fino al punto previsto.

Avevo il problema della rotazione con i giunti a T forati, che dopo un po' perdevano di robustezza, ma con questi ad incastro gli elementi rimangono dove li metti. Il problema e' invece riuscire ad allinearli sullo stesso piano.

Bisogna tenere presente che comunque si tratta di antenne per installazioni temporanee, dove l'importante e' la rapidita' di s/montaggio e il peso. Il resto e' grasso che cola.

Eventualmente, se i propri incastri sono troppo laschi, si puo' montare l'antenna a testa in giu', con il boom in alto e gli elementi appesi come dei pipistrelli.

Multiband beam (70 MHz)

I have no experience on this matter, so this is a purely theoretical post.

If looking for a directional antenna that covers 50 MHz, 70 MHz and 144 MHz, why not consider a log-periodic? With a boom of 4.5m, max width of 3.6m and 11 elements, three bands are covered. A matching system is needed.

Another version, probably with a better impedance range but lower gain, counts 15 elements over a 5m boom.

So, there's plenty of room to play with. Just need to put to work some simulation software and then cut lots of pipes...