Filtro passa-banda per i 70 MHz

Usando RFsim99 ho progrettato (fatto progettare al programma...) un filtro passa-banda centrato a 70.3 MHz con larghezza di banda di 15 MHz:


Richiedendo una larghezza inferiore i valori di alcuni componenti diventano molto critici e difficilmente realizzabili.

Gli induttori sono realizzabili con:

• 24,153 nH: dia 5mm, len 7mm, 3 spire
• 1,061 uH: dia 8mm, len 18mm, 19 spire

Date le basse potenze in gioco i condensatori saranno semplici ceramici affiancati da comuni trimmer capacitivi.

Questo filtro avra' il doppio compito di abbellire il segnale trasmesso e di filtrare un po' le stazioni radio FM verso il ricevitore.

Confronto filtri di uscita

Prima di chiudere definitivamente il progetto del TX QRP per i 70 MHz ho voluto provare un altro tipo di filtro di uscita, quello a pi-greco. Entrambi sono descritti in un documento trovato in rete.

Cosi' ho sostituito il LCL con un RCLC, dove R era di 100 ohm in serie all'uscita della porta AND. La potenza risultante era nuovamente di 14 dBm (25mW) su 50 ohm e si notava un leggero peggioramento sulla risposta alle alte frequenze (colonna "LPF pi").

Ho poi rimosso la resistenza da 100 ohm pilotando direttamente il filtro con la porta logica. Non so quanto essa gradisca il trattamento, ma la potenza di uscita e' balzata a 21,5 dBm, circa 140mW. Peccato che le armoniche siano solo 40 dB sotto la portante (colonna "LPF pi NO R").

	LPF T		LPF pi		LPF pi NO R
[MHz]	[dBm]	[dBc]	[dBm]	[dBc]	[dBm]	[dBc]
70	14	0	14	0	21,5	0
140	-35	-49	-38	-52	-19	-40,5
210	-40	-54	-40	-54	-19	-40,5
280	-52	-66	-49	-63	-26	-47,5
350	-47	-61	-35	-49	-24	-45,5
470	-50	-64	-56	-70	-28	-49,5

Il grafico mette a confronto i tre tipi di filtro di uscita fino alla sesta armonica:

Altri valori misurati. Su "sonda a RF" 1,16V efficaci, ma a questi livelli di potenza e' un valore poco significativo.

L'assorbimento in TX e' passato da 44 mA a 60 mA. Considerando i 33 mA in stand-by, 5*(0,06-0,033) = 0,135W IN. Qualcosa non quadra in questi calcoli, ma forse ci vorrebbe il dettaglio dei singoli assorbimenti.

Provero' ad aggiungere un ulteriore LC per attenuare le armoniche.

Telescopic VHF/UHF antenna

This is a 30 minutes project that I did in the pre-Internet era. I wanted an antenna for my VHF handheld that would outperform the rubber duckie.

Easy: take something metallic some 50cm long that is self-supporting vertical.

I had a telescopic antenna whose length reached 50cm. Its base diameter fit loosely into a PL-259, so I screwed a thick wire to the antenna base and soldered the other side to the connector pin. I kept the antenna away from the outer ring, and vertical, with some pieces of rubber (probably RG58 black coating) pressed in with a screwdriver.

When closed the antenna is about 12cm long, including the PL259. Of course the size of any reproduction will depend on the telescopic antenna used.

(pictures not in scale)

Closed

UHF

VHF

Connector detail, FWIW

So I have 12 to 50 cm of radiator, which equates to a 1/4 lambda between 140 and 625 MHz. So it covers both VHF and UHF (and 220 MHz) HAM bands with a simple movement, and anything in between or above. It is not a multiband antenna: you need to retune it for a band change.

This antenna has now turned into a VHF/UHF facility that permanently sits in my car, together with the magmount base. The fully extended whip did not mind travelling at 130 km/h.

As a bonus you can use a similar antenna as a signal pickup on a monitor receiver when testing homebrew transmitters.

Air core coils. An update

I'm willing to try a different output filter topology on my 4m TX, so now I need a 113 nanoH coil. I would also like to fit the transmitter into a metallic box to see if RF leakage during stand-by becomes more acceptable.

113 nH is 5 turns on 8mm diameter (drill bit), about 10.5mm long.

I decided to try with a thicker wire, say 1mm diameter. Even if I wound it using some good finger pressure, it did spring back. But they are only 5 turns, so once the support was removed I tightened the coil with a circular finger/hand movement.

Now I have a self-supporting 8x11mm coil whose inductance can be adjusted by stretching/compressing it.

So the coil wire diameter vs. coil diameter plays an important role on the need for a permanent coil former. How to find out? I haven't come up with a rule, sorry! Try and see, and YMMV.

Misure sul TX per i 70 MHz

Il TX e' ultimato, o almeno diciamo utilizzabile se si dispone di un'antenna per i 70 MHz.

Ho scoperto di avere accesso ad un analizzatore di spettro (45-1200 MHz, step 50 kHz), non propriamente pensato per i segnali "CW", ma che almeno mi permette di tarare il filtro di uscita e stimare la potenza RF generata.

Il test e' stato effettuato usando 4-5 metri di RG58 tra TX e analizzatore, l'unico cavo a 50 ohm che ho intestato con BNC al momento. Questi sono i valori ottenuti:

[MHz]	[dBm]	[dBc]
70	14	0
140	-35	-49
210	-40	-54
280	-52	-66
350	-47	-61
470	-50	-64

Circa 14 dBm alla fondamentale, su 50 ohm, sono 25 mW. Sinceramente speravo qualcosa di piu', ma i dati sull'assorbimento di corrente l'avevano preannunciato:

• 33 mA in stand-by
• 44 mA in TX

Sono 11mA di differenza a 5V, ovvero 55mW IN. Efficienza del 50% sulla singola porta AND del 74AC08, non male.

Se si va a "scavare" nei bassifondi dei dBm sull'analizzatore si vede che c'e' moltissimo rumore sotto i -60dBm, portanti varie, ... l'oscillatore e annesso moltiplicatore per 5 non sono proprio silenziosi...

Ho anche notato una certa fuga di segnale verso l'analizzatore di spettro quando il circuito e' in stand-by. Si misurava -47dBm a 70 MHz e -61 dBm alla seconda armonica. Forse con un po' di schermatura il problema si risolve.

Ecco il circuito:
La bobina a sinistra fa parte del filtro passa-basso a pi-greco (v. post successivo).

Chi vuole tentare il QSO? :-)

Air core coils. Yeah, sure!

"Yeah, sure!" was the typical answer of my friend Urska when our requests were too awkward.

For my 4m TX I computed number of turns vs diameter vs length for a couple of air-cored inductors (output filter) using "mini Ring Core Calculator" software.

I picked round diameter values (10mm, 20mm) and adjusted length until I had an integer number of turns. Took adequate lengths of enameled wire and started winding on drill bit. 1 2 3 4 5 turns, done, release finger pression... CLANG! The coil springs back to a wider diameter and fewer turns.

Changed wire with a thinner one, same result plus a more instable coil.

Lesson 1: a coil former is needed. It will hold the wire in place, forever.

But I don't have any 10/20mm former!

Lesson 2: compute the coil according to available coil supports.

I have spare 16/25mm PVC pipe cut-offs. 16mm was behind the door (literally) and large enough. Re-computed coil lengths, cut the pipe to host the wire plus some extra as support.
I also cut a strain relief guide on both sides of each former that keeps the wire in tension.

Il TX per i 4m prende forma

Ecco come appariva due giorni fa, prima che aggiungessi il driver/PA:

La basetta ramata e' quella da cui ho ricavato l'adattatore SOIC, il chip e' un moltiplicatore di clock ICS512. Acceso con una batteria da 3V c'era un bel segnale a 70.308 MHz.

SOIC to deadbug extender

The "SOIC to DIP adapter" presented here actually is a "SOIC to deadbug extender", since I never spoke of drilling holes and adding pins. Did I?

In the final version I also drew diagonal cuts towards the adapter corners so that the two inner pads have a larger, triangular shape, which is much easier to handle.

Here's a SketchUp of the chip and extender:


Got the picture now? And, by the way, it works as expected! I soldered the SOIC8 with a 0.6mm RoHS Sn and 15W iron (1mm tip or so). No flux, no other helper: just a steady hand, a lot of patience and some luck.

Cheap eBay XTALs

In April 2006 I bought a batch of XTALs via eBay from a Chinese shop. The shop is still open (16 August 2007) but currently does not list XTALs.

I took 2x7030, 4x14060 and 4x14020 kHz: 10 XTALs for 18 US$ shipped. Good deal! Not so...

I recently planned to use those XTALs for my homebrew 4m TX , which started to come to life on August 14th, 2007.

On a Colpitts test oscillator, the 14060k XTALs was resonating on 14020k. OPS! The 14020k was close to 14060kHz. I tested a couple more crystals and the mis-marking was confirmed. The 7030k rock went on at 7028kHz, which is reasonably close to the marked frequency, and perhaps my oscillator was pulling it a bit.

Lesson learned? Be careful with cheap components on eBay!

SOIC to DIP homemade adapter - w/pictures

You didn't ask for it, but here is my process described with pictures.

NB: I take no responsibility for whatsoever damage you may do to any animated or inanimated object. Do it at your own risk.

Sit the IC (or an identical one) on the board and hold it in place with a crococlip, near a corner of the board.

Mark on the PCB empty areas between pins using a permanent marker with 1mm tip.

With a SOIC-8 you'll get 3 marks per side

Very carefully with a cutter and a ruler draw connections between
twin marks on each side of the IC, and extend them beyond your marks

Then etch. Note the original marks...

With the cutter also etch a line across in the middle, so that you'll have 8 pads once you cut the copper clad board with your favorite method so that the PCB is composed of 8 pads

Last but not least, check insulation between adjacent pads with an ohm-meter, "dig" more with the cutter if you measure zero ohms. I discovered that you can emboss the etch if you hold the cutter slightly diagonal.

Enjoy.

SOIC to DIP homemade adapter

Got the ICS512 chips for the 4m TX and I suddenly have to deal with a SOIC-8 package. I remember seeing some solutions around the web for adapting SOIC to DIP, to which I'll add my own:

1. Sijosae's DIY Gallery, scroll down to "Tip"
   
2. commercial adapters (search for "Browndog", or check what Futurlec or eBay offer)
3. The saltwater etch process
4. my own cutter etch process

I cannot quickly buy the commercial adapter. I also seem unable to use Sijosae's method(s), and I don't want to abuse the chip since I've gotten very few of them. So I'll go for the 4th method.

• Take a piece of unetched PCB.
• Sit the IC (or an identical one) on the board and hold it in place with a crococlip, near a corner of the board.
• Mark on the PCB empty areas between pins using a permanent marker with 1mm tip. With a SOIC-8 you'll get 3 marks per side. The result:
  

__ __
__ __
__ __

• Remove the chip
  
• Very carefully with a cutter and a ruler draw connections between twin marks on each side of the IC, and extend them beyond your marks:

_____________________
_____________________
_____________________

• With the cutter also etch a line across in the middle, so that you'll have 8 pads once...
• ...you cut the copper clad board with your favorite method so that the PCB is composed of 8 pads
  
• Check insulation between adjacent pads with an ohm-meter, "dig" more with the cutter if you measure zero ohms
• Pre-solder pads

With a little patience and firm hand you can produce several adapters in a row.

That's it. Hopefully I'll add some pictures of the process.

Finally on the air with 5 elements

In mid-July 2007 I tried the 5 element VHF Yagi out on the field.


Supporting mast was the lower section of a fiberglass fishing rod ("telescopocket 4"). Antenna performance was good. Beamwidth not much tighter than the 4 el., but I did notice better signals and reports were encouraging.

There's still one goal to reach: use a telescopic fishing rod as boom, but given its conical shape I can make up a good element-to-boom fixing harness.

Transverter 4-10 di OZ2M in kit

A questo link ci sono tutte le informazioni per comprare o riprodurre un transverter 4-10 metri in kit: http://rudius.net/oz2m/70mhz/transverter.htm

Il costo indicativo e' di 125 euro + le commissioni di pagamento (bancarie o PayPal).

In ogni caso lo schema elettrico e' pubblico e vi si puo' attingere per le proprie realizzazioni.