9:1 Transmission Line Transformers 50:5.55Ω - More Results

Ruthroff Type 9:1 on Two Ferrite Beads

As an alternative to using trifilar windings I tried two of these cores, each core with 6 bifilar turns. These are connected in such a way that a 9:1 transformer is achieved. My expectation was this was just not enough ferrite or wire to work. I was surprised then to see a useful return loss from 160m to 20m. 

Ruthroff 9:1 Transformer Two lots of 6 turns on a single core

Ruthroff Type 9:1 on Small Binocular Ferrite

Switching to a regular binocular core I wound 6 bifilar turns on each half and repeated the measurements.

Binocular core, 2 x 6t bifilar, each winding on separate half of core

Comparing these two cores with the 4 bead 4 trifilar turns blogged about last time shows the 2c beads can be just as good as this binocular core:

Upper: 6 bifilar turns on each of two beads
Middle: 6 bifilar turns on each half of binocular core
Lower: 4 trifilar turns on 4 beads

 

Surprised again I looked at the Smith Chart of S11 for the binocular based 9:1 transformer. Since parallel capacitors are often seen compensating RF transformers I captured a few tests. Unfortunately I lost track of the compensating values used but the impact is clear. As the capacitor value increases the S11 curve rotates clockwise on the Smith Chart:

2 x 6t bifilar on Binocular Core
Upper no compensation, Middle some compensation and Lower More Compensation

 The resulting plot of the magnitude of S11 shows the outcome:

2 x 6t bifilar on Binocular Core
Upper no compensation, Middle some compensation and Lower More Compensation

Compensation can improve the performance of this style of transformer. It struck that a series capacitor could be used to "slide" the uncompensated curve around the Smith Chart. A 1nF capacitor was about right at 14MHz to cancel the series inductance present:

So yes the trace slid around the Smith Chart, but the transformer became narrower in bandwidth:

So if you need compensation to achieve the desired performance perhaps a conventional style transformer with compensation can be used? I'll look into that next.

73's

Richard

9:1 Transformers 50:5.5 - Initial Results

With the success testing 4:1 transformers I examined some 9:1 transformers. A trifilar winding, being a bit thicker than a bifilar winding using the same gauge wire, proved difficult with the small cores previously used. I wondered how a single winding through 10 of these cores would perform. I passed a single trifilar winding through 10 turns, arranged them into a U or long binocular shape, and secured them with piece of heatshrink.

1 trifilar turn through 10 cores

Note the Smith chart response is curving away from the 50Ω origin upwards and to the right. This is the same with a bifilar winding.

Returning to the test results here is the sweep of S11: 

1 trifilar turn through 10 cores

 

Very useful from 160m to 30m applications and so easy to wind. Just one turn. But it looked a bit cumbersome and not as broad as I had hoped. So I tried two turns through 6 cores. This was better but still didn't improve the upper frequency. So I tried 4 turns through 4 cores. 

Comparing 1t 10cores v 2t 6cores v 4t 4 cores

Clearly 4 turns through 4 cores is a better transformer, and would handle the 1W I intend to push through it at 80m. However, not as broadband as I had hoped given the success I had with 50:12.5Ω transfomres using these core.

At present these transformers are destined to be used in the 10W Class AB amplifier for 80m I am planing to drive with the 1W Class A amp I have developed. It will be interesting to see if they handle the power.

Overall, these tiny cores are a useful part. I certainly consider them useful for 4:1 transformers but trifilar windings are more time consuming to wind and solder.

73's

Richard

4:1 Transfomers 50:12.5Ω - Success

I hadn't expected this to work out so easily. I already had a selection of bifilar wound transformers wound on these cores so it seemed like a good place to start. Terminated in 12.5Ω I looked at the 50Ω connection with the nanoVNA:

8t bifilar on a single core

Another single core with 12turns bifilar was tested and it was too many turns. Good for 160m to 40m though. 

Shifting to two cores side by side like a binocular core with 7 turns proved useful for all of the HF bands. Five turns could be a suitable starting point for 6m.

7t bifilar on two cores binocular style

To check the insertion loss I wired two of these transformers back to back. (50:12.5 + 12.5:50) 

The loss was negligible.

Loss across two transformers back to back

With 1W applied to this arrangement for 15 minutes the cores showed no sign of heating, perhaps 1degree above ambient.  

The goal was to find a suitable transformer for matching 50Ω to input of a 10W amplifier. If a 4:1 transformer is the solution then this has been achieved. But a 9:1 transformer may be required so another round of testing is needed.

73's

Richard

4:1 Transfomers 50:12.5Ω - Preliminary

Having resolved which ferrites to use for 200:50 ohm transformers it is time to see if any of the cores I have are suitable for matching from 50Ω to a lower impedance using a bifilar winding. The application in mind is for the input stage matching on a single ended mono band HF amplifier delivering 10W or more. 

While I have built push pull HF amplifiers it always struck me as ironic that semiconductor manufacturers often specify the IMD properties of RF transistors in a single ended circuit. However, that just gets ignored because push-pull configurations are claimed to deliver lower distortion and harmonic cancellation. Maybe that's possible, but I know single ended amplifiers can deliver very good results. And I struggle with the concept that a transistor in the off state can cancel harmonic energy generated by a transistor in the on state. Perhaps the overall result is less about the topology and more about the implementation?

My approach will be to determine which core can offer a good transformation from 50 to 12.5 ohms, then use two back to back to see what the loss is and how well the core copes with 1W of RF. 

73's

Richard

1 Watt RF Amplifier using General Purpose Transistors - Latest Iteration

Taking on-board the test results to date the latest iteration of the 1 watt amplifier has the following characteristics:

• 1 Watt output
• 2SC3356 > 2SD1664 > 2x 2SD1664R in Parallel
  
• First two stages use a bifilar transformer to present the collector with 200 ohms for increased gain.
• Use of resistors to provide a thermal bridge to ground for heatsinking. 

The results of my experiments have overturned my previous position. You do not need a working RF transistor, suitably rated, recovered from another radio to build 1 watt Class A HF amplifiers. 1W HF amplifiers can be reliably made using many of the general purpose transistors available that cost a few cents each.

If there is sufficient interest I'll put together a kit. My back of the envelope estimate is the cost would work out around US$5 for a PCB with all surface mount parts fitted.

Save those working RF transistors for more demanding applications!

73's

Richard

4:1 (200:50 ohm) Transformers - Yet Another Suggested Core

So this core is surprising but bang for buck is pretty amazing. It's quite small but 8 bifilar turns spaced around the core is useful from 160m (just) to 6m.

8 bifilar turns

 Perhaps you have the patience to use more turns of still finer wire?

 

12 bifilar turns

But glue, or hold while you start winding, two side by side and wind 7 bifilar turns using fine wire also results in a good transformer:

Given the extra time it takes to work with these small cores it might be false economy. But it does remind me how useful the nanoVNA is for making such measurements of unknown components. And these ferrites are really low cost.

Details of power handling and loss soon.

73's

Richard