January 30, 2024. The Final TFMS ~ Tune For Maximum Smoke

WTB (cheap) audio output transformer 5K plate to 3.2 or 8 ohms. Cheap is the operative word!

 

TFMS. The Fire to the Wire Circuitry where you Tune For Maximum Smoke!

I am often taken back by the claims seen in various projects, like a 2N3904 Final RF Amp that puts out 3/4 of a watt. Most of the time if you see 100 milli-watts out of a 2N3904 you have hit the lottery. I guess authors do that to entice you to build their project! 

Check those values by simulating their circuit in LT Spice. If you only see about 4 or 5 dB of gain that is a huge clue.

Then again, I frequently make contacts to the Mid-west from the left coast on 20M SSB running a real two watts. 

 

If you measure your output across a 50 Ohm resistive dummy load and you are not seeing 28 Volts Peak-To-Peak, then you are not running two real watts. 28^2=784 and taking that value times 2.5 = 1.960 watts ~ 2 watts. 

If you take the PTP value and square it and then that result X 2.5 and divide by 1000 you get watts RMS. With my approach you do not have to convert to RMS. The 2.5 factor accounts for that and the 50 Ohm Load. So for those of you who doubt. RMS starts with 1/2 Peak (14 Volts) and then you take 0.707 times that number which equals 9.898. 

 

Now you square that and you get V^2 = 97.9704 our formula says (V^2)/R = 97.9704/50 = 1.9594 (1.96) watts. So guys PTP^2*2.5 and the answer is milliwatts. 784*2.5 = 1960 or 1.96 watts. N6QW TKT.

For TFMS the first place to start is with the Driver and Final Amplifier stages and that is to set reasonable expectations. No, you can't get 750 milli-watts from a single 2N3904. The answer is in the mathematics.

The amp circuit you are using has a 13dB gain -- the device specifications told you that number. If you take the ratio of the power output (in milliwatts) to input in milliwatts and next take the log of that and finally multiply by 10 --you get the gain needed to produce that output from the drive level to the stage.

 

 

Our example: You measure 45Volts PTP across 50 Ohms and that my friend is slightly more than 5 watts or 5062.5 milliwatts. You are driving that amp with 100 milliwatts and thus 5062.5/100 = 50.625. Now 10*log(50.625) = 17dB. That is kind of hard to do with a 13dB amp. 

The real answer is a 13dB gain device will only get 2 watts with 100 Milliwatts of drive. Do the math: 2000/100 = 20. Thusly 10*log(20) = 13dB.

Now if you hit your amp with 250 milliwatts of Drive then a different story. With 5 watts out and 250 Milliwatts of Drive we get 5000/250 = 20 and we know that is 13dB. 

Low power output may not be a problem in your Final -- it likely is a Driver stage problem. Adding a second IRF510 and/or boosting the Drain voltage to 24VDC may not do anything -- it is a Drive issue!

So, to get the Fire in the Wire you have to look at both stages. 

So here are my go-to Driver and Final Amp stages. Of importance is a standard layout that focuses on keeping stray RF from finding a feedback path and then you have an oscillator. (Avoid Kb = 1)

 

Above is my universal Final RF Amp layout. While shown with an IRF510 this will also work with a RD06HHF1. I cut the Drain Pin off to avoid one of those unwanted feedback paths (Thanks KB1GMX). In the case of the RD06HHF1, the tab is the source and the right pin is the Drain. Conveniently we have a pad there for that. 

You will also need a bit more bias for this device. So, look at the 3 Terminal 5VDC regulator and simply lift the Ground Pin and install a LED between that Pin and Ground. This raises the reference level, and you get more than 5VDC for bias. Another cool feature when you apply bias from the TR circuitry the LED lights, so you know you got juice to the amp.

 

I mill out the pad where the Drain (or Source) would go so that you have access to the backing heat sink. You will also need a bigger heat sink with the RD06HHF1 -- higher idling current.

Next, we have the pad layout for the Driver stage and much thought went into this layout. The long horizontal bar is for the Base connection. On the left side is the input capacitor and on the right side is a pad so you can fit two feedback resistors from the Collector where one of those resistors has a bypass cap. The long vertical run is the Collector Pad. On the left side of the long vertical pad is the open space where the RF Choke is placed and the single pad on the top right is the output capacitor. The two small pads on the upper left house a 10 Ohm resistor and 100nF to ground which form the decupling network from the power source rail. Like I said lots of thought and compact!

With this circuit I find you can get about 10 Volts Peak to Peak which if you do the math 10^2 = 100 and that times 2.5 = 250 milliwatts. But it all backs up into the sewer -- you need sufficient drive to the Driver stage from the Pre-driver. If you clock this Driver stage at 13dB then you need about 12 milliwatts of drive from the Pre-driver. 

Doing the math: 12 Milliwatts = 2.2Volts Peak to Peak. 2.2^2 = 4.84. That value times 2.5 = 12.1 Milliwatts.

Recently I have shifted to the 2SC5706 as a Driver for two reasons with cost being one as the 2N2219A has gotten expensive and two its Tab can be soldered right to an isolated copper pad as a connection point AND heatsink.

 

 

Yes, for those with an itch in your twitch this is boring and all you want to do is heat up the iron. Boring it may be but also important when you are seeing only 2 watts to the antenna and then this is a roadmap as to where to look.

73's

Pete N6QW

January 29, 2024. Bi-directional Amplifiers

WTB (cheap) audio output transformer 5K plate to 3.2 or 8 ohms. Cheap is the operative word!

There are many forms of Bi-directional amplifiers and of course we have a segment of our ham homebrewer demographic who have a TIA Tattooed on their butt. But a bi-directional amplifier can take other forms. I will share some of my approaches with you.

TIA (Termination Insensitive Amplifier) enjoys great popularity in that the gain can be adjusted by the resistors you choose for the circuit. Then there is the all important 50 Ohms that remains 50 Ohms no matter what you hang on the front end of the amp. They are particularly useful on either end of a crystal filter as that has the overall effect of improving the signal quality of that stage, because of the constant impedance factor. Of course using a TIA makes you feel a kinship with one of its inventors, Wes Hayward, W7ZOI.

Transceiver Module with two SBL-1's, a set of TIA's and a 9MHz Filter

Quite honestly the TIA's worked great, BUT not something so spectacular that you would only use this circuit. The Plessey Module (also recognized by W7ZOI) does a great  job without any noticeable difference and is 50 Ohms in and out and a lot less parts. Then again I don't have a tattoo of a TIA on my butt.

Long ago I came up with a project idea called LBS (Let's Build Something) which along with AI6YR, Ben we co-authored a series for QRP Quarterly. We started with a Direct Conversion Receiver and using most of that hardware turned it into a 40M SSB Transceiver. 

 

I came up with an idea for a Bi-directional Amplifier that had gain pots in each leg so you could optimize that stage. Typical use might be as a Rx RF Amp in one direction and Tx Pre-Driver in the other. On Rx you want some gain BUT not max gain and on Tx you want the pedal to the metal.

 

 

Here was the neat feature of this design. The output off of the secondary of the matching transformer is always terminated in 50 Ohms with the 100 Ohm resistor in parallel with the 100 Ohm pot (a forced 50 Ohm termination). The center wiper is the tap point for the output and provides a variable output without disturbing the 50 Ohm termination. No one noticed this, but it works well! Relays are used to steer which amp is in-line depending on whether in Receive or Transmit.

 I have used bi-directional amps that were diode steered and you saw some of that in yesterday's post on Crystal Filters. Look at the 4 Diodes around the 3N211 and certain Points are either biased on with +12VDC or Grounded. Note the A and B arrows to show how the signal gets routed around the DGM. On the right most diode if the upper RH point is grounded and 12VDC applied an the Lower RH point the signal form the upper RH core gets routed to gate #1 and since the upper LH point has +12VDC on it, the Drain is supplied juice for direction B and the output appears on the upper LH matching transformer. Cool Beans guys! This circuit was found in SPRAT authored by my friend Ron Taylor, G4GXO.

Currently my approach is to use Relay steering where two relays can reverse the In's and Out's and it can be a small foot print. Here a whole IF Module is steered in two directions.

P3ST with Relay Steering!

We earlier mentioned the Plessey Circuit and many of my homebrew transceivers have that functionality (but only one has the TIA*). I like the Plessey Module since I also have a CNC layout using SMD Parts. So, it is pretty easy to cut a board -- just punch the Start Button. This makes for a compact rig. Good to 6 Meters with 17dB of Gain and 50 Ohms in/out. No toroids to wind.

 

 

I have also used MMIC devices in a bidirectional module that also has a small footprint. Good to 3.5 GHz with 20dB of gain and 50 Ohms in and out.

 

 

 

So as a designer and you as a builder there are many bi-directional forms aside from a TIA. * My take on the TIA is like the Nano VNA and Tiny SA, it certainly has excellent qualities but used primarily, I think because of the identification with Hayward and Kopski perhaps more so than termination insensitivity! Hey guys this rig uses TIA's from W7ZOI. Do cool kids have a Nano VNA Tattooed on the other side of their butt?

 

This is sort of like wearing Nike Tennis shoes, where you have a strong identity by using the brand. You know when you take your morning walk in the hood the widows and divorcee's spot you and approve because you have your Nike Tennis shoes on for the walk. But you can wear Skechers and they cost a lot less and will get you around the hood, but you might not be noticed!

 

My P3ST design specifically had as a goal No Bitx, No TIA, No EMRFD and No Facebook Groups. That radio works well and is now a kit from K7TFC.  There another piece of a SSB Transceiver project has been provided to you so get off the couch you couch potato!

 

73's

Pete N6QW