1. This web page imbues the spirit of experimental electronics — sharing ideas
and potentially future site content.
I edit my essays continuously. The schematics on this page may contain errors.
2. I rotate some original photos near the page bottom. Mouse over images to read a brief description.
3. New, permanent content is announced on the QRP - Log page.
With deep respect and condolences to his family: VE4PU expires — a big loss for amateur radio in Canada. ARRL write up: http://www.arrl.org/news/view/j-kenneth-pulfer-ve3pu-sk
April 30, 2013 — Anti-blog for all Strata and Sects of Builders
Against the grain — I’ve never tried Facebook, Twitter, Google’s
email/blogs/G+, Linkedin or any other social media. I sometimes feel guilty
writing this blog since reading it zaps time from your [real] analog life.
Does slavery to online activities impair or enhance our lives? Sometimes I wonder. Isn’t it better to talk face-to-face with friends and family, spend more time petting our cats, hugging our wife and children and to rise above the mediocrity, trivia and/or time wastage perpetuated by @YourAFoolToFollowMe and the like?
Perhaps modern, wireless people prefer to live vicariously through others blogs and tweets? If someone blogs daily/weekly/ often — what hidden tragedy lurks in their life? Perhaps that person is the smart one — selling you stuff, or making money off your mouse clicks, or maybe just boosting their ego.
Ignorance and poverty threaten the world’s safety. Read the newspaper — they really do. I’m so thankful my parents had means and they and others helped me learn to read and acquire the skills and motivation to rise above a paycheck-to-paycheck existence. Sadly, I know folks who can but don’t save or invest to boost their retirement income; perhaps sucker-victim consumers who lust for the next Smartphone that, according to plan, will go obsolete 6 months before they pay it off.
I find it ironic that the best time to start retirement investing passes long before we embrace the reality of our mortality and the need for retirement income. Sadly, many, trapped in poverty, can’t even consider investment. What can we do to help poor people improve their lives? Lots!
With personal success comes an obligation to pay it forward and I think
about/act on this a lot. For example, consider literacy and its role in
poverty. We all can help
others learn to read and write; through volunteering or donation. Consider
John Wood from Room to Read. He’s
the real deal— a guy who created an outcome-driven, transparent,
business-model run charity that involves books, libraries, community
investment and respect for both genders. Sadly many charities lack
transparency [you know who I’m talkin’ about — those that hire big ad
agencies and spend ½ of their donations on overhead and their fat cat CEO]
Room to Read, along with the Canadian SPCA are just 2 of the many
exceptions amongst the bloat charities. Discern wisely in all things.
So I won’t blog in May and ask you to join me in putting some of your online time into taking up a cause to better your [real] community.
If you need some free parts email me. I’ve acquired some benefactors:
builders who want to pay it forward and share our hobby with those who can’t
find or get parts for whatever reason. Big thanks to those who help me with gear,
parts, + offer advice and encouragement. I’m off the bench until October.
I wish you a great Spring/Summer + fun on and off the bench.
April 22, 2013 — Bad Bench Basics in a Popcorn Prescaler
Above — Popcorn prescaler breadboard with leaded parts for all but the
IC. I wanted to make a popcorn prescaler to allow my 225 MHz HP frequency
counter to work up to 1.1 GHz. I chose the popular MC12080 SOIC part since it
offers divide-by 10, 20, 40 and 80 per the arrangement of 3 external switches.
Focusing on learning the output voltage + what integer I should choose, I
casually built it on a single-sided copper board and threw breadboarding caution
to the wind.
Punishing oscillations from about 10-70 MHz destroyed input sensitivity and signal viewing + annoyed me. Click for output signal without the 220 pF filter capacitor sans the source follower. Click for the source follower output with the 220 pF cap in place as a low-pass filter. Messy stuff huh?
Above — My schematic with a * 220 pF cap added to help attenuate (but not remove the wideband oscillations). With some direct B+ filtering and better bypassing, I measured the output frequency and the MC12080 worked as expected. Normally we would rebuild the circuit properly and not add a cluster-kludge like the 220 pF filter cap.
The source follower looks odd and normally is unrequired, however, I wanted a small module with 50 Ω input and output so I could use a standard 50 Ω cable to connect the output to my counter to preserve maximum sensitivity.
I placed a 1K8 Ω output R since I wanted to try all possible divide-by integers. In my real build, I'll hardwire it as divide-by-10, plus solder in a 820 Ω output resistor. Tight SMT chip husbandry, UHF precautions, double-sided FR4 board and so forth should stabilize the unwanted oscillations and boost input sensitivity to spec. I like to show both my successes and bench mistakes to cultivate humility and since readers email me to say they like seeing my failures too. Rest assured; I make many errors, or get lazy/sloppy/stupid often enough.
Above — 50 MHz into the MC12080 set to divide-by-20 yielded 2.50 MHz as expected. Click for divide-by-80.
While cooking on Sunday, I badly cut my thumb with a kitchen knife and will be off the bench for at least 1 week. Since I stop bench experiments each May 1 — that might be it for bench experiments until October. I generated some new content this Winter and will slowly post it on the site as I get time over the Summer. Cheers!
April 9, 2013 — Hey
Above — Cats supervise all aspects of this web site and I'm glad.
I've built a string of failures lately — bad audio filters. Trying to push it too much? Perhaps. We can measure our experiment output in parts consumed — in all, 15 op-amps, ~100 resistors, hoards of polyester film caps + copper board chunks and of course, burnt fingers. The cat doesn't mind 1 bit for he's glad to just be there; occasionally chortling that he wants a stroke or 2. Since I must wash my hands to remove lead before I pet him, he slows down my breadboarding a little.
Above — Essentially a DC receiver AF stage from product detector to speaker jack. My breadboard features 2 volume control pots to allow different experiments. I've crammed a lot into this chassis and thus increased my experiment difficulty by about 6 dB — that was stupid. Small copper boards = big experimental breadboarding hassles.
Above — A breadboard of 1 experimental AF filter: it's actually 2 filters that I hoped to mix to set the perfect amount of low-pass filtration to suit any particular QRM situation. I found lots of problems: DC offset issues, volume pot interplay (that's fixable), difficulty choosing the proper fixed gain for each filter relative to the other, and finally, what cutoff frequency to choose for each side. Calculations help, but only listening tests find the lemons.
I spent a few hours checking radio blogs out — there must be a million now. A few blew me away; sharing amazing and original technical content. Others blog about other people's blogs, some share trivia, still others narrate great radio adventures, while a few read like "Dear Diary". I prefer real content sites where the author makes stuff and tells why and how he did it, and of course, cites measurements. Something for everyone I guess.
April 1, 2013 — @ Audio
I'm finishing off my Winter Session experiments at audio frequency (with just a dash @ UHF too). To set the context, the following blog-rant covers from a receiver product detector to the speaker. Recall that I'm making a superhet receiver to update my Popcorn Superhet from 1998.
Above — My post-product detector amplifer: a common emitter/ emitter follower pair arranged for a 50 Ω input impedance with shunt and series feedback. Getting the biggest, clean sine wave in and out posed my only goal — the 3.54 mA current source on the follower helps that cause. The input Z tested quite insensitive to output load variations and the (S11) return loss of 24.1 dB seems okay.
Above — The "Full Monty". A MCL SBL-1 diode ring mixer drives the 50 Ω input Z feedback pair and finally a 5532 op-amp with adjustable gain. The 25K gain trimmer pot allows the builder to accommodate factors such as whether the stage is a DC receiver, or an RF-AF block for a superheterodyne receiver. I'll set the gain as low as possible and hope to build up my AC voltage with active, gain producing low-pass filter op-amp stages.
The output is directly coupled to an op-amp AF filter. With software we can easily design a myriad of audio filters: band-pass, low-pass, high-pass, all-pass etc. I only run low-pass filters in this slot and even here we must decide whether to make our reponse Gaussian to 6 dB (or 12 dB), Bessel, Butterworth, or a Chebyshev with some degree of ripple. I'm testing out a few designs with an eye towards simplicity and 10% parts.
Spurred on by curiosity, with my ears as the main evaluation tool; I currently have a 5532 paper design to deliver 20 dB gain with a 500 Hz pass-band, Gaussian to 12 dB low-pass response. Click for schematic, I chose a 500 Hz cut off frequency since the skirt shape by definition is subtle and the Q low. This trick should better roll off the AF above 1 KHz a little faster without the ugly group delay associated with steeper filters employing more poles. We shall see!
Above — The compact breadboard of Figure 2.
Above — A 965 mW AF power amp with a 10 dB voltage gain stage incorporating a Chebyshev 1 KHz low-pass filter to roll off some of the higher frequency din and any hash noise. In Sundry Experiments 2013, Section 8. Popcorn AF Amplifier for Receivers — Reprise, I presented a power follower stage that may be AC coupled to any voltage amplifier stage you want. The beauty of this is we can drive the PA stage with a low-pass, band-pass amp etc. with whatever gain or cutoff frequency(s) we choose. The possibilities are endless! The amplifer above is the lowest gain power amp stage I've ever designed or built. Click for the maximum power clean signal test at 1 KHz — 965 mW. Do yourself a favor and stop just using the LM386.
Audio Samples as a 7 MHz DC Receiver
After connecting a 7 MHz RF band-pass module to its RF port, I spliced Figure 2 to Figure 3 for evaluation. By design Figure 2, will not require a coupling capacitor between it and the planned Sallen-Key AF filter that follows; however for this test (sans filter), I needed to add 1. I tested a cap value between 1 μF and 0.22 μF. At 1 μF, the audio sounded mushy and SSB QSOs lacked crispness. I preferred between 0.22 and 0.47 μF (0.22 μF a touch more) since this allowed me to exert a high-pass response simply by choosing a specific coupling capacitor. Subjectively, we might prefer a different response and our choice may differ whether we are tuning SSB, or CW; or even by the quantity of QRM present.
In this example, I tuned someone calling CQ and then raised the beat
frequency by re-tuning. To my ears, the retuned, slightly higher beat
frequency sounds louder, despite the AF amplitude remaining equal when
measured by software. The audio samples are heavily compressed into mp3
files to reduce file size.
Click for example 1.
Click for example 2. I'm milling around the 40 M band and tuned with my homebrew, bench HF signal generator. Band noise with my vertical antenna loomed high — my SWL listening antenna came crashing down during a February storm. The only antenna work I plan to do this Summer will target the VHF and UHF bands.
For my ears, I want an audio filter that I can vary the low-pass cutoff frequency -- this may be tough too achieve with an AF filter that provides 20 dB gain, however, I'll keep plugging along with my design work.
I'm pleased with the preamp and power amp fidelity. For testing, I used a speaker lying on the floor to avoid any speaker cabinet response, or microphonics since the cabinet doesn't yet sit on rubber feet. Fidelity seems great, my circuit noise ultra low, and hum absent even with a longish test lead joining the 2 audio stages, no cover on the box and jumper wires strewn about. Click for tonight's experiment photo. Audio experiments transform us in many ways.
We log every byte of web site activity and a 1-day log file runs between 9 and 12 MB in file size — they're kept forever. I rarely read log files except when server or security concerns arise. Once and awhile we study daily server activity and since its inception in 1998, despite the ever-growing number of blogs and digital evolution of radio electronics, site readership grows steadily. My mean unique visit total lies at 2K per day now.
I hope to continue presenting insightful + actionable info for the DIY analog builder. Readers email - how can I help the site for I don't ask for, nor accept money. 1 way is to send parts. I send free parts all over the world and welcome your support in kind. Got a few extra 5532s kicking around? I can always use them. BF998s, new BJTs, or even trimmer caps; send me an email. Thank you.
I added a Pin Outs web page. Mostly because many search engine returns of (for example) "SBL-1 datasheet" or "SBL-1 pin out" return web pages like "Datasheets are us DOT com" that either give spam, no information, or spawn a wild goose chase....and all I wanted was a pin out. I'll slowly photograph and add some staple items for the QRP - SWL homebuilder on this page. I don't want to offer hype; it's just something new. I'll link it on the Design Center web page.
Above — A graphic showing unique visitor site activity by hour for March 18. Our annual thoughput is in the 10's of terabyes now. We run a rack with 4 computers in parallel and yes, home-built our computers and network.
March 18, 2013 — Analog Andy
DVB-T dongles employing the Realtek RTL2832U controlled by code
first written by Antti Palosaari brought cheap SDR to the masses. And how!
We’ve witnessed an explosion of exciting activities including inspired
homebrew, exploration of novel radio bands such as aircraft ADS-B signals + the FUNcube satellite project,
and some new kit offerings.
As an analog, component-level, hardware experimenter, I’ve been asked "Does this signal the end of analog receivers? " My answer remains an enthusiastic no — it’s really just the beginning of a new chapter in our hobby. Analog circuit knowledge remains vital to compliment the digital stuff and visa-versa.
What analog topics can the dongle experimenter explore?
• Antenna design and matching
• Low noise amplifiers
• High-pass, low-pass, band-pass, band-blocking filters etc.
• HF converters
• Shielding, decoupling, RF noise immunity /reduction
• Test equipment! My sole purpose in this hobby — Measure-holics anonymous here I come!
As an AM and FM DXer, the current cheapo Realtek-based dongles don’t give me the sensitivity and noise performance I need, although I could build a better antenna, design some front-end filters and such. So far, I’ve only helped others and not built anything dongle for my QTH. Dongles can be prone to things like RFI from a myriad of stuff in and outside the house (including your own computer and monitor) plus base-band aliasing and such — likely this might all be fixed 1 day.
Alas, my heart lies on learning the analog side for now — that and photography. I'll be experimenting mostly at AF for the next few weeks. Only this weekend, I added some popcorn AF experiments to the Sundry 2012-13 web page as Section 8. Here's a photo of my main bread board from this weekend. I'll connect my best preamp, audio filter and power amp to a diode ring mixer and test it as a DC receiver. Then that RF-to-AF block will become the AF portion of my upcoming popcorn superhet receiver.
Already, we're pruning trees and tilling the garden. It won't be long before I pack it up until Fall. Cheers!
March 3, 2013 — Diplexer Perspectives
Above — I tested the W1JR (Joe Reisert) bridged-tee diplexer against another design used by Ten-Tek. I'll show the outcomes on the Almost Popcorn receiver page some day, but the W1JR better fullfills its purpose; termination from DC to daylight at the mixer IF port. I calculated the L and C values for a 11.0592 MHz IF with my Universal Diplexer application from many moons ago.
Above — SPICE provides a great way to assess a diplexer since you can plot both S21 and S11 against frequency.
Above — A fantastic match! Thanks Joe for the design and to Wes, W7ZOI for his help with diplexers and the SPICE plots through many years of related experiments. Since web publishing the diplexer page, at least 500 diplexer-related emails have come in. Human translated versions of the web page exist in at least 2 languages.
Above — The evaluation breadboard of my W1JR diplexer. Those are T30-6 toroids -- they look huge in this photo.
Above — Tracking generator plus spectrum analysis sweep. SPICE does a better of examining the high-pass side.
Above — A look at the diode ring mixer, diplexer and post-mixer amp from the current version of my Almost Popcorn Receiver. I tested the diplexer and RF amp as one unit (see the S-values). After terminating the 200 Ω output with a 4:1 Z transmission transformer, I tested it like any other 50 Ω circuit. Even with the 6 dB attenuator pad and losses from the diplexer, the gain is still 15.5 dB.
It did not make sense to use a transmission line transformer to get 50 Ω on the amp collector and then use another transformer to build it back up to the needed 200 Ω impedance for the crystal filter, so I just employed a single choke driving a 200 Ω Z attenuator pad. S22 = the 200 Ω transformed into 50 Ω — it measured -30.6 dB, indicating my xtal filter input Z is well established. Click for a big picture of the breadboard.
Feb 26, 2013 — We Build Stuff
We had no power on and off for 2 nights, so nothing got done. Email piled up badly, but even that's okay now. I've been mostly working at VHF, but performed some more experiments for the Almost Popcorn superhet update. Some ideas follow:
Above — Ugly breadboard shot with a 105 mm lens for a change. It's a Beaverton Special feedback amp (FBA) biased for ~40 mA. This will go as my post-mixer amp. All that's left in the IF box: design the diplexer and solder down a Level-10 diode ring mixer.
Above — The Beaverton Special feedback amp biased for ~40 mA. The output return loss at 35.2 dB is at the directivity of my return loss bridge -- even without the pad, S22 measured blissfully strong. The FBA with no 5 dB pad: S11 -18.5 dB; S12 -34.6 dB; S21 22.1 dB and S22 -31.8 dB. I love this FBA for its simplicity and well defined port impedance. My IF = 11.0592 MHz. Click for a discarded (but cool) photo.
Above — A high-resolution 'scope capture that's just breaking into clipping
as I boosted the drive from the maximum
4.74 Vpp up to 4.76 Vpp, the negative peak is just starting to flatten out. Other shots show some harder distortion when pushed vigorously. All the base drive comes from a fast current feedback op-amp that can drive an output current of 100 mA peak by itself. I enjoyed reading the datasheet for the LM6181 -- a great part. This AF power amp with a maximum sine wave power of 351 mW won't go in the receiver, however, it spawned some ideas for future experiments
My biggest accomplishment was finding a SMT replacement transistor for the 2N5179/PN5179. I'll blog about it later, but I'm really stoked: fT = 3 GHz, NF = 1.5 dB. I can't wait to bias up and sweep this part once it arrives safely. More later...
Feb 10, 2013 — Crystal Filter Experiments
I spent a pleasant Saturday afternoon designing, building and testing Gaussian-to-6 dB crystal filters. I inputted some xtal parameters + constants from Zverev into Xlad08.exe from EMRFD. I love this program from Wes.
Above — A screen capture from Xlad. This app allows skilled builders to depart from the typical Cohn (Min-Loss) filter and tune filter poles to derive beautiful skirt shapes and low insertion loss if wanted. I'm not a great crystal filter designer, but I've seen work from friends who have a knack for it: imagine making a N= 8 to 10 crystal CW filter with just ~1 dB insertion loss like these builders! Click for the GPLA simulation.
Above — Gaussian-to-6 dB xtal filer schematic where N = 3.
Above — Circuit to allow evaluation of the filter in my 50 Ω sweep system. Transmission transformers provide the needed 200 Ω filter termination. I inserted 4 trimmer caps to permit tweaking of the filter under test to establish the very best filter shape. With trimmer capacitor adjustment, I was able to trash the filter or peak it well. Click for 3 different swept bandwidths: Sweep-1 Sweep-2 Sweep-3.
After peaking, I removed all the caps and measured them as shown in red font. I'll replicate these values with nearest standard-value fixed capacitors. The trimmer caps nearest to the ports proved more sensistive compared to the 2 middle trimmers capacitors during tuning.
This might be the coolest thing I've ever done on my test bench.
Above — My N = 3 filter breadboard. I left the capacitor leads long for re-use. I consume huge numbers of parts and recycle as possible to keep costs down.
Above — I built 2 N=4 filters. Click for a sweep of this 1.
Feb 2, 2013 — Almost Popcorn 40M Band Receiver IF Strip Ideas
Above — Version 2 of a popcorn IF amplifer idea with BF998s. IF =
I have no use for AGC in my personal receivers since I listen through a speaker and always ride the RF gain control with my finger. Since the BF998 has gain into UHF, wideband decoupling + bypass SMT R-C networks filter each MOSFET.
A network matches the 200 Ω xtal filter output impedance to the 3K3 input Z of the IF amp cascade; I'll soon breadboard and tweak this network. The L-C-C network has become my favorite matching device — 2 tweaks is often better than the 1 offered by a garden variety L-network. The product detector that follows this stage is a Level 7 diode ring mixer soldered onto the audio preamp board. This receiver will feature some new, exciting circuits aimed at low noise + high fidelity but still stay "almost popcorn". Anyhow, that's my hope.
Above — A breadboard of the Popcorn IF amp sans the input network. I tested this stage using an oscillator with a 3K3 output impedance and liked what I saw. Click for a reverse angle view. After this build, I ordered some SMT versions of the 1N4148 diode. The BNC connector was only soldered on the board for testing purposes.
I rotate photos on the bottom of my blog. From May to October I stop radio experiments and study photography, watch football and we grow vegetables and flowers. I'm adding captions to most future photos since portable devices might not show the 'tool tips' captions normally seen by hovering over a photo.
Above — My tube electronics mentor Tom, VE7TW. Back in the day with an early TV broadcast tape machine.
Above — Live from our garden. Papaver nudicaule — known as Champagne Bubbles, Icleand Poppy, or Arctic Poppy. Great Spring/early Summer flower.
Above — A photo taken by my photography teacher Peter. We've never met face to face; but email constantly. He sends me his photos and links to great photos in order to condition my brain about what makes great composition, lighting, exposure, lens choices and story telling. While millions of photos are taken daily, great photos prove quite rare.
Our discussion includes a lot of lighting topics. Making pictures is about 'sculpting with light'. I've received much instruction about flash photography and like small Nikon Speed Lights for portability. Here's a page from the Peter's notebook — lots to learn. My first flash discovery (an epiphany): I dislike on-camera flash.
Photography mirrors my interest in learning electronics: creativity, experiments, success and failure + fun.
Above — Candidly photographing photographers lies amongst my favorite street-photography subjects. Peter took this photo near the Golden Gate bridge and the expression on the guys face/body language looks priceless.
Above — My photo of a friend with his son. Alternate take while I was determining the best angle, light and exposure. This became my favorite picture of 2012. After examining my entire photo collection in 2013, I realized that only 4 pictures taken during my whole lifetime look 'really good' — this is 1 of them. I enjoy taking portraits and plan to work on portraiture this Summer.
Above — Cousin photograph from long ago. Film. Some people can light a room with their genuine smile. She did.
Above — A Ukrainian woman that my friend hung out with. Slavik woman often photograph beautifully. Cheap point and shoot camera.
Above — 7 digital cameras watch inside and outside the server site. 1 infrared camera covers the cat door. Last year, camera 4, also IR, caught the night time break-in of a nearby building. Clear video of the 3 criminals, their 2 vehicles and licence plates was gladly accepted by the police. Mostly we just view the cat's adventures. All the code to control the multiple cameras, compress data, and rapidly analyze tons of footage was written by family members — it's sold commercially.
Above — Our kitty with a 105 mm lens. Click for my favorite photo of 2011: an AVRO Lancaster taken in Alberta with my street camera in natural light (35 mm lens ----- APS-C sensor, so 52.5 mm focal length). The exposure looks perfect due to the fabulous diffuse light surrounding the plane. When you get great light take lots of pictures and bracket your exposure -- you're bound to get 1-2 keepers.
Above — A relative getting married during WW2. Uniformed ceremony with canopy of rifles.
A pro I do repair work for wanted another dual triode added to his 40+ year old "no-name" 12 Watt tube guitar amp laden with two 6V6 finals in Class A + a solid state rectifier. The chassis already contained a a big hole from a former tube socket, so I bolted in some copper clad board and added a 12AX7a. In the board space of the former preamp stage, in total fear of hum and noise, I wired a classic Fender preamp and a 1-pot tone circuit with some scooped mids. After power up — we enjoyed low hum, a kill-you-dead blues mojo; and the hardest part was watching him leave with his new blues machine. The mostly unseen tone circuit is connected right on the pot and it's 1 of my own designs.
All web site photos taken by me or my family unless stated otherwise. No Photoshop. Thanks.
Apart from a few specifically copyrighted schematics on my web site, this is an open-access site, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication. Best regards!