Excellent Transformer Tutorial

Posting this here to save for future reference. Good stuff.

A Milestone is Reached!

For the last few weeks, as time permitted, I have been working on building a 40 Meter Transceiver kit. The kit is called the 2n2/40 and it is kitted by the Norcal QRP Club (http://www.norcalqrp.org/nc2n2xx.htm). They make great kits at very reasonable prices. Since this is a labor of love for them, they do all of the development and kit production for free and sell everything at cost. They have temporarily suspended taking new orders, but I understand they should begin accepting new orders in the next week or so. The kit is available in 40 M, 30 M and 20 M versions – I ordered the 40 M version.

The kit designer is QRP legend, Jim Kortge – K8IQY. Jim perfected a circuit board building technique called Manhattan Construction. Rather than etch circuit boards and drill holes, the builder uses small islands of circuit board material and glues (Crazy Glue) them to a blank circuit board. The connections between components are made on top of these “islands” by simply bending the component leads so that it stands upright and soldering the leads to the islands. You can see an example of this construction technique at http://www.k8iqy.com/qrprigs/2n220/2n220Pictures.htm. You've got to see these pictures to appreciate the simplicity of this technique and Jim's craftsmanship. Jim has designed a whole series of rigs around the venerable 2N2222 transistor (the cockroach of the solid state world). The rigs perform as well as store bought rigs costing hundreds, if not thousands, of dollars (except they are CW only and top out at 5W).

However, K8IQY and Norcal have teamed up to produce these kits using traditional circuit boards. Now you can stuff a traditional circuit board, solder some leads and build one of Jim's rigs cheaper than you could order the parts yourself.

The assembly manual (online at the Norcal site) is every bit as good, and simple, as one of the old Heathkit manuals. However, that has not stopped me from making a few mistakes as I go along. Did you know a circuit is not likely to work if you install the diodes backwards, or put the wrong resistors in the right holes (or is it the right resistors in the wrong holes)? But the Norcal manual provides easy testing procedures at each stage of the build, and there is a very supportive Yahoo group where K8IQY lurks to help you with troubleshooting problems. I have completed the receiver portion of the 2N2/40. Here is a picture of the circuit board with the receiver strip completed (some controls are temporarily tacked in right now – awaiting their home in the cabinet. Now, on to the transmitter!



The receiver is very sensitive and the audio is good and strong and rock solid. I think I'm in love.

[Crossposted at http://ke5lep.blogspot.com/ ]

The little oscillator that could

I am continuing on my 2n2/20 build. With some wailing and nashing of teeth I have finally gotten the product detector, crystal filter and RX oscillator done. Here's a picture.

The crystal filter and product detector are in the rear (where the toroid with colored wires lives, where the camera is actually focused) and the RX Local Oscillator is in the foreground.

A little history. I built the product detector a few weeks ago. The big issue there was getting the toroid wound and all the wires color coded. That took ol' shakey (me) several hours. I put everything away for the weekend and came back to work on it a week later. There was no toroid to be found.

What happened? Where did it go? Did I misplace it? Did I...oh yeah, this is probably the work of "clepto cat" (aka "Missy" - full name: "Misadventure"). Missy likes shiny things. She takes them and, apparently, saves them for her retirement or something. Anyway, I was sure I was a victim of clepto cat. I looked all over the shack, but no luck. However, that's not surprising because she usually takes things as far away from the scene of the crime as possible. She does this to confuse the people from CSI.

I finally gave up looking for the toroid and spent another few hours winding toroid version 2.0. That's the toroid with the colored wires that you see in the picture. Got it all wound up and installed all the parts for the product detector and crystal filter. Later that night I found toroid version 1.0 stuck behind the network switch that lives under my desk. Dang clepto cat.

I was getting pretty tired of digging cat hair out of my soldering iron. So my lovely female spousal unit (Sue) got me a special Valentine's day present: she made me a "clepto cat guard." Here's a picture of my workbench without the clepto guard:

Here it is with the clepto guard in place:

Pretty sexy Valentine's Day present, huh? It's made from a science fair presentation board that she fancied up. It seems to work, so far.

So I continued with my build. I built the RX local oscillator next. It was a pretty easy build, even though it was a little cramped. All went well until I fired it up. I was able to see the oscillator working on my scope, but the signal was not strong enough for my freqency meter to pick it up. What's wrong? My scope said that I was only getting about 100mv p-p. Using my advanced liberal arts major math skills, I think that translates to about a -30dbm. The schematic said I should be getting a+7dbm signal out of the oscillator. What did I do now?

Mike, WA5PSE, said I might be using the 10:1 scope lead. Nope. I did some quick and dirty scope calibration checks and it was reading correctly [i.e.: close enough]. So I started following voltages around the circiut. Everything seemed healthy until I worked my way up to the "output" at the L8-R55 junction. That's where everything seemed to shut down.

I looked to see if I had the proper value components at that spot. The resistor was OK. And the inductor was the proper 27uH value. Wait a minute, that seems a little high. So I squinted at the schematic under bright light. Sure enough...what I thought was supposed to be a 27uH inductor was actually supposed to be a 2.7uH inductor. Oh man.

I replaced the 27uH inductor with a 2.7 uH. Yeah, that makes a difference! The scope now says I'm getting out about 1.6v p-p; .8 v peak; .56 v rms = +7 dbm (about). Am I a techno stud, or what? Here's a picture of the oscillator running with my freq. meter attached. I knew you'd be interested:

TC7 seems to adjust out that extra ".12" - but I'll deal with that later, if need be.

See. You thought I couldn't do this, didn't you?

73

72

WA5PSA

2N2/20 Build

Mike and I have been building K8IQY's 2N2/20 transceiver for awhile. I got off to an early lead. But Mike shot on ahead. OK...I'm slow.

The 2n2/20 is a 20M transceiver. CW only. 5 watts. Built using the manhattan style of construction.

Here's where I am right now.

Pretty impressive, huh? Yeah, I don't think so either.

Obviously, most of my progress is along the west side of the circuit board. This is all of the receiver audio circuitry, including the receive mute switching circuit. In the upper right hand corner lives the transmit/receive switching. The empty pads are for the roofing filter and product detector.

Here's another shot of the product detector and audio pre-amp. This shot helps you understand why this is called "Manhattan" style construction. Because it looks like you could get mugged down there. No, actually, seen from the side like this, the components look like a city skyline. It also looks like I need to figure out how to take macro pictures with good lighting.

Here's an aerial view of the audio preamp. See the four diodes lined up over in the product detector section? I've wound a toroid that will snuggle in right next to those diodes. Let's see if I can take a better picture of the board.

Not too much better. Oh well.

I just got through winding the toroid that will live just to the right of the four diodes that are standing up near the center of the photo. That installs next.

73

WA5PSA

Continuing with the fix-em-up.

No one on QRP-L had a manual or schematic for this particular model MFJ antenna analyzer. I emailed MFJ to see if they'd send me a pdf of the schematic, but they wanted $5 for the manual, by snail mail. Too much hassle. So, I decided to dig into it and just flail around.


I spent a bit of time trying to trace out the schematic from the printed circuit board, and that was a loser. Or maybe I was. Man, that's hard. I knew I really just needed to isolate the components for Band D, but I was having trouble doing even that. Then I decided to see if maybe the silly thing was oscillating, but at such a different freq., that I just couldn't hear it on the radio.


So, I hooked it up to my O-Scope. Might as well learn how to trace RF with the O-Scope while I'm doing this. I hooked the scope probe to the antenna out terminal, and the scope ground to ground foil on the MFJ. I also hooked another scope probe up the same way and attached it to the scope trigger input. I'm trying to learn about scope triggers. When I do that, and switch the scope to external trigger, it seems to produce a steadier trace.








I turned the MFJ to band A and looked at the scope. Nuthin'. Nuts. I tweaked and messed around with the scope, and then the MFJ. Nuthin. What am I doing wrong? Then it hit me. I hadn't turned the MFJ on! OK. That could cause low output...


I switched on the MFJ (acting noncholant, since the cats were in the room and I didn't want to look stupid). The scope voomped to life and, after a little arm (knob) twisting, I could see the sine wave from the MFJ's oscillator on the scope screen. Here's Band A:

Now that's what I'm talking about. My guess is that the flatish bottom, and the little anomoly around the zero point is a scope problem and not the oscillator.

Just for grins, I thought I'd see what Band B looked like.

Sure-nuff - the wave peaks are much closer together (higher freq.). Moving the frequency dial on the MFJ caused the wave peaks to move closer or separate (higher or lower frequency). Kind of like an accordian being played. I'm sure real techs have seen this a bajillion times; but it's fun to see in practice what you've only read about.

But when I switch to Band D some strange things happen. On the lower part of the band, there are no oscillations at all. Nada. From 12 to 14 MHz it is dead. Then, around 15MHz on the dial, it pops to life and oscillates its little heart out from 15MHz to 30MHz. Something's sick about the MFJ. However, I think I'm not looking for open coils or bad solder joints like I thought I was. I was thinking that all of Band D was dead (open coil or bad solder joint). But the scope showed me that only part of Band D didn't oscillate. There's something else going on.

Oh...another cute thing I noticed. When I swished the frequency control on the MFJ past 20MHz, the amplitude of the signal on the scope went waaaayyy down. Coincidently, the rated bandwidth on my scope is 20MHz. Go higher than 20MHz, and the scope gets the vapors.

I also wondered if loading could have something to do with it. My base tests were done with no antenna hooked up to the MFJ. I then hooked up my outside dipole to the MFJ and went through the bands again. Noticed some neat things. The amplitude of the wave forms on the different bands varied widely depending on how I set my antenna tuner. I could "see" the antenna tuner hit resonance at a certain frequency because the needle on the MFJ would dip, but the signal on the scope would increase in magnitude. It all roughly corresponded to settings that I had already determined were resonant for the antenna tuner at various frequencies. Kuel.

Also, when I attach the antenna to the MFJ, the MFJ, Band D reacts differently. With the antenna attached, depending on where the antenna tuner is set, Band D will begin oscillating at various points. In other words, the "dead" portion of the band will expand or shrink.

So, I tried an experiment. I put the MFJ into a "dead" part of Band D, where I simply had not seen it oscillate before (around 14.5 MHz). I then played with the antenna tuner at different settings. Oddly, at a few critical settings on the antenna tuner, the MFJ oscillator springs to life. However, it does not act like the rest of the bands act...the oscillator only seems to start up when the antenna tuner is at a certain spot (maybe resonance?). On the other bands, there is always some sort of signal from the MFJ, even if the antenna tuner is out of resonance. Hhhmmm.

Ya know...I wonder what frequency the osciallator is really running at on Band D (when it runs)? Maybe it's just way off. I think I'll hook up my freq counter. Problem: the MFJ and the freq counter both run on 9V, and we've only got one 9V battery in the whole house. But wait...I've got a variable bench supply. Duh.

When I hook up the freq counter, wierd things happen. The frequency starts wandering all over creation and the oscillator kicks in and out. I wonder, with both a scope input and a freq counter, hooked to the MFJ output, if I'm presenting the oscillator with too many strange loads. I unhook the scope. Sure enough, the freq counter settles down and the freq stabalizes. OK. Maybe I just should have started with the freq counter. Oh well...saw some cool things on the scope.

With just the freq counter hooked up, it looks like the top of Band D is oscillating just fine (the counter reads up to about 35MHz with the MFJ dial past the 30MHz point). Oscillations stop at various points, but they usually stop anywhere between 19 and 25 MHz and are dead to the bottom of Band D.

It's gotta be some Band D specific component that's off kilter. I just find it hard to trace out the Band D components with this circuit board. I thought I could just use my Inductance Meter to identify the Band D inductor, but there are several more inductors than there are bands, so I can't isolate it. Besides...I just realized, if this thing is wired anything like the MFJ204B, then all the inductors are in series and the band switch switches different parts of them in and out of the circuit. But the inductor for the high band (Band D on my unit) is always in the circuit. Since the other bands work OK, my guess is that the inductor is OK. Rats. Since the inductors are the only band specific components in the circuit, that means there's something else going on. Not gonna be so easy.

So, what could cause an oscillator to work OK from 1.5 MHz to 12 MHz, and then stop oscillating (depending on the load put on it) from 12MHz to 19MHz, only to start up again and osciallate OK from 19MHz to 35 MHz? Makes my head Hertz.

WA5PSA

How I fixed my MFJ Antenna Bridge

At this point, the title is a bit optomistic. A) I haven't fixed it yet. B) I'm not sure what's really wrong yet. But I thought I'd create this log of my floundering.

Operational test.

1. Is the resistance calibration of the MFJ accurate?
A. Connect antenna to MFJ.
B. Set band switch to C range.
C. Turn on ICOM 706 receiver and set for 10.110 MHz.
D. Turn on MFJ and sweep frequency until MFJ oscillator can be heard in 706. (706 reading: 10.100; MFJ reading: 10.5 MHz.)
E. Calibration chart on back of MFJ says that 50 ohm resistance is "C 1/2" on the resistance dial. Set resistance dial for C 1/2.
F. Tune antenna tuner (MFJ Versa Tuner) for maximum dip on the MFJ Antenna Bridge meter.
G. MFJ Versa Tuner readings: Antenna 3; Inductance D; Transmitter 4.
H. Disconnect antenna from MFJ Antenna Bridge and connect to ICOM 706.
I. Read SWR on ICOM 706: 1.1:1.
J. Conclusion: the "C 1/2" 50 ohm setting on the MFJ Antenna Bridge is pretty accurate. Using the antenna tuner to dip the Antenna Bridge's meter seems to produce an Antenna Tuner setting that is pretty close to 50 ohms.

2. Is the frequency dial on the MFJ Antenna Bridge accurate?
A. Test by listening to the oscillator of the Antenna Bridge.
B. Readings:
15M: Can not hear oscillator.
17M: Can not hear oscillator.
20M: Can not hear oscillator.
30M: Actual: 10.1; Reading: 10.5 MHz
40M: Actual: 7.0; Reading: 7.7 MHz
80M: Actual: 3.6; Reading: 3.65 MHz
C. Conclusion: B range is fairly close. C range is reading about 500KHz high. D range doesn't seem to be oscillating.

I've got a schematic for the MFJ 204B; but not one for the 204. There seems to be substantial differences. The 204 only has two transistors; the 204B has 4. The 204 has 4 bands; the 204B has 5. I need to find a schematic for the 204. I think I'll drop a line onto the QRP-L website.

WA5PSA

This Old Signal Generator

My radio buddy, Mike, WA5PSE, got me a "new" signal generator at a hamfest. Well, not new, exactly. I don't know how old the actual unit is, but the manual I found at BAMA is copyright 1954. A fella's got to be a few years older than his signal generator.

This is an EICO Model 324 (you can see more information here). Mike said it was in good shape...and it is. I opened the cabinet up and it looks like it was built yesterday. Mike said it was in pretty good calibration, and it is.

The 324 has 6 bands, from 150KC to 455MC (don't ask it to do anything in KHz or MHz; this thing was built when we were still using "kilocycles" instead of "kilohertz.") Anyway, to calibrate the 324, you've got to adjust some slug tuned coils, one for each band, on the back of the band switch.

I intially thought I would check calibration by simply listening to the signal generator's oscillator on my ICOM 706.

Yeah, that wasn't a good idea. I could find the 324's output on the 706, but it was quickly apparent that this was a pretty inefficient way to calibrate the 324. When I adjusted the coils on the 324, it, of course, would cause the 324's signal to disappear from "view." I then had to chase the signal back down and readjust. Each band seemed to be about 20KHz off; so setting the 706 up to receive a calibration frequency (say, 7.0 MHz), and trying to move the 324 onto that frequency, was really shooting in the dark. I wasn't sure I was headed in the right direction (up or down). I wasn't sure how far I was from the calibration frequency at any point in time. I felt like I was doing it backwards.

What I needed was something that simply gave me a read out of the 324's real current frequency. Oh...you mean like a frequency counter? Like the Norcal FCC-1 that I just built a couple months ago? Yeah, like that - dipwad. I forgot I even had the thing.

So, I got out the FCC-1; hooked it up to the 324 and lo and behold - nothing. The FCC-1 just sat there, in all of it's digitalness, looking extremely bored. But it weren't countin no frequency! Why not? I had it properly hooked up to the 324's output. What gives? I started fiddling with the 324 and, after a little coaxing, the FCC-1 finally wakes up and displays a frequency (sorry 324; the FCC-1 reads out in KHz, not KC). Anyway, it appears that the output level of the 324 was set too low for the FCC-1 to even respond to the input. When I cranked the gain controls on the 324 up to about mid range, the FCC-1 started reading frequency like a little digital champ.

Then it was a simple matter of figuring out which slug tuned inductor to adjust for each of the 324's bands. I fiddled around looking at the schematic for awhile, but quickly decided to ignore that and just look at the coils themselves. Obviously the coils with the greatest coilness (that's a technical term for the "most wire") had the most inductance and would correspond to the lowest frequency bands (the bulkiest coil would be band "A", etc.). No duh.

I started switching from band to band and calibrating the 324 like crazy. It wasn't too far off (except on Band E, which was about 2MHz off). I got it dialed in pretty close on each band. Close enough for use in Smith Labs anyway. Besides, the old analog dial, even with its vernier knob, would still only allow for a certain amount of precision. Trying to dial it in too precise was just fooling myself. If I need that much precision when I use it, I'll just sic the FCC-1 on it.

It works like a champ.

WA5PSA

Hello World

This is just an initial post to get this blog off the ground. I'm going to use this blog to document some of my more nerdy interests, such as ham radio and electronics.