This page documents a recent homebrew signal generator project. My electronics and building experience is pretty minimal, and I could not have done this without the help of Todd's (VE7BPO) helpful site, QRP SWL Homebuilder (Amateur Radio Electronic Design) at Todd's site is very fun and friendly, with thousands of pages of useful design tutorials, discussion and photos.

I wanted a small lightweight signal generator I could use for antenna projects for HF HAM radio experiments. This means it should produce clean sinewaves in the frequency range from about 3.5 - 30MHz and be designed for low-level but gain-controlled 50ohm output.

Todd designed a circuit that seemed to fit my needs, though it is only designed for up to about 10MHz ( I eventually fitted mine with plugable inductors in order to provide the range I needed.

The project box was a found 70's vintage gizmo for telephone servicemen - I removed the front dial and circuit board only. I kept the front panel controls and power supply for possible later use.

All the parts used were already in my junk box- nothing was bought for this project other than a few J310 FET's. I'm working with what I got, which might explain some of the odd parts and configuration. Of course, I have a lot of junk.

I have limited metal working skills/tools. To mount the variable capacitors I used a pair of tin snips to cut a 3"x3" square of thin metal from a dismantled boatanchor. My 82-yr old father gave me his drill press, which has helped tremondously in drilling all those holes! Of course, I make mistakes, sometimes just 1/32 inch off will require extra work in the form of reaming or replacement. It would look much better if I had simply replaced the entire front panel, but I have no way to cut and shape metal of that heavy gauge. And of course limiting the placement of the variable caps to that small section on the left of the front panel meant many hours of time I need to rethink that! But I am thinking I may want to add an LCD panel for a counter on the right side as a future project.

I initially decided to build the front end (Hartley Oscillator) point-to-point on a terminal lug strip attached to the front panel. The rest of the circuitry I planned to build ugly style on a board (shown).

This was the first time I attempted using a rotary switch, and it had to be remote mounted with an extender. I made a shaft coupler from an aluminum spacer. It took me more time than I wanted to figure out how to repurpose that switch! But doing so saved me from using two different band-switch toggle switches or similar.

One thing you will notice is the winding on the toroid- I used some ancient magnet wire which wasn't designed for such tight corners... the enamel coating simply flaked off! You'll see in later photos that I rewound it for a plugin.

I was having trouble obtaining the range of frequencies I wanted, and I also was getting weird inconsistent results. I could not get the desired frequencies by simply switching in various amounts of capacitance- this Hartley would not oscillate for just any LC combination. I decided to try switching different toroid inductors. I started using a SPST switch, but quickly determined it would need a SPDT to switch both the tank and the tap at the same time. But this led to a mass of wires and generally a mess.

So I decided to step back and replace that point-to-point mess with a board. With some planning, I used my Dremel to create pads on a small ciruit board. I have used PCB software before to etch boards, but it takes me too long to do it that way for these simple circuits.

I spent some time thinking how I might use swappable plugin toroids, and decided to mount them onto 8 pin DIP sockets. Five wires, but two of them are grounds so I would use the four corner pins. This way I could have a stable circuit and experiment with various toroids to see which configuration would work best. Unseen in the photo below is a red soldered socket that each toroid plugin socket will plug into on the reverse side (you see this red socket from the front view in a following photo). The leads on the toroids wound with 24 gauge wire could simply be pushed into the socket, but smaller gauge wire had to be soldered to a heavier pin before being pushed into the socket.

That turned out to be a good idea, because now I could now easily and directly compare the effect of different windings by just replacing the plugin coil assemblies. Here's some data using two switched air variables:

Low range capacitance is approx 20-800pF (two ganged air variables)
High range capacitance is approx 10-40pF (air variable)

Looking at the results I think I will try a 30 turn toroid as well to see if I can get an approx 3.8-19MHz range.

Actually this was a little disappointing because I had hoped to get 3.8MHz-30MHz using only one coil, but doing these experiments proved to me that isn't possible with this particular Hartley oscillator circuit.

At some point I accidently damaged the main air variable capacitor and had to replace it; they are a lot more delicate than I had thought. Since the replacement was different in size/shape I needed to reconfigure slightly. Out went the shaft coupled rotary switch. And you can see I made a mess of the metal insert with my nibbler tool enlarging a hole.

I ran out of ceramic bypass caps, so used some plastic ones I had. In this particular project temperature compensation/control shouldn't matter too much (to me) since I don't plan to use this for anything critical, but generally temperature compensation is very important for VFO's (and Todd emphasizes this).

I built and tested the hybrid cascode amp, then followed with the feedback amp. These were fairly straight-forward compared to the oscillator circuit. This was the first time I had wound toroids to use in a project, and it took me awhile to understand how to wind the broadband transformer for the FBA (Todd has a page devoted to winding bifilar types).

These circuits were designed by Todd to be powered by approx. 12VDC, but I decided early on I wanted to use a single 9-volt battery. I won't get the optimum performance, and it draws a bit much to run on battery all the time. So I decided to also make a 9-volt regulated supply for this, again using mostly original parts and others from my junkbox.

The plugin toroids are convenient and fun but delicate- those 8-pin DIP's are so cheaply made! I didn't happen to have any with machined pins. As you can see I have temporarily stuck them to the side of the box with Velcro.

Completed Sunday 12/15/14, works well! Now onto building a dummy load, RF sampler, peak reading power/SWR meter, antenna tuner, etc.

Keith Ostertag