开云体育

Hello.

I am working on an ATtiny power / SWR meter, basically a smarter
replacement of the famous Tayloe SWR bridge. It will be a 1W to 5W (maybe a
bit higher, up to 10W) QRP resistive bridge with 6 LEDs analog-like
baragraph, with the intensity being interpolated between two successive
LEDs enhancing the baragraph resolution to at least 4x6 levels (I can
easily recognize 1 LED illuminated from 2 LEDs equally illuminated and 1
LED illuminated significantly stronger than the other). I have the
baragraph electronics and code prototyped, it works quite nicely.

The electronics (ATtiny13A driving the 6 LEDs) will be powered from a 3V
low drop low low quiescent current 16V max input linear regulator, and the
linear regulator will be powered with a 1:2 or 1:3 transformer from the
radio RF power, rectified by a single 1n5711 shottky diode (the diodes will
be used in the bridge anyway, so to reduce the BOM complexity). Look ma, no
batteries.

There will be a larger capacitor at the input side of the regulator, so the
microprocessor will have couple of seconds to perform additional tasks
after key up, that is to show the input power with quite a high accuracy.
So there will be no button: SWR will be shown during key down, and power
will be shown after a second or so after key up.

I feel quite confident with digital design and programming, but I am not
quite sure about the RF
side. Namely, what transformer should I use to feed the linear regulator?

These are my back of the envelope calculations:

On the transceiver output, when loaded with an open resistive bridge (all
three resistors 50 Ohm), the SWR is 2:1 and the voltage at the transceiver
output is 1.3333 * of Vpeak (assuming 50 Ohm output impedance of the
transceiver). Likely the maximum voltage will be lower as the output
impedance of the transceiver PA is not resistive. These are the maxima:
10W -> 42.16V Vpeak at SWR 2:1
7W -> 35.28V Vpeak at SWR 2:1
6W -> 32.65V Vpeak at SWR 2:1
5W -> 29.81V Vpeak at SWR 2:1

The minimum Vpeak at 1W transceiver output with the resistive bridge
shorted (again producing SWR 2:1 at the transceiver) will be minimum 8V
assuming the transceiver PA output is resistive. If not, the output voltage
will be closer to 10V peak.

Now the digital circuit requires a minimum 3.1V at 1.5mA DC before the
linear stabilizer, and the power will be rectified with a single diode, but
the micro will likely work happily down to 2V even if the linear regulator
just passes the voltage unregulated, as the micro may use its internal 1.1V
voltage reference for measurements. Operating the micro unregulated sounds
unusual, but it may be a viable option if tested, especially if the input
voltage will be stable enough due to the large capacitor before the
regulator.

Now my question is:
1) Should I use a 1:2 or 1:3 transformer? Using the 1:2 transformer may
require additional zener diode and likely a resistor before the linear
stabilizer if the input voltage could exceed 16V peak.
2) What will be the tiniest / cheapest ferrite and how many widings will be
optimal? I suppose this decision will be strongly dependent on the bands
supported. My personal goal is a 40m-20m EFHW tuner.

And also:

Will there be an interest for a kit?
Through hole or SMD? It may be a nice "introduction to SMD" kit as the
number of parts will be pretty low.
What bands?
What power range?
Should it contain an EFHW tuner? My personal preference would be a "NJQRP
Rainbow Tuner" like setup covering 40m-20m with a T50-6 transformer, mica
compression capacitor trimmer, jumpers to select the primary winding with
another jumper to add a capacitance for 40m coverage.

Thanks and 73,
Vojtech OK1IAK (former AB2ZA)

I pondered the idea an using a QRP radio or even a 20W radio where
some larger chunk of power out goes to SWR/power meter tp power it
seems wrong. Even if the suggested electronics uses a watt on a 5W
radio that's a lot of power not going to the antenna.

Also Tayloe bridge SWR meters are a source of some 6-10db of
insertion loss. This is good while tuning the antenna as it
protects the transmitter from a bad swr but unacceptable
for inline use. So the SWR detector (directional coupler) for inline
use should be of the Tandem match type or similar and even
that sampler eats a little transmitter power. The only point
for it is now reflected power so a simple display works fine.
For power a meter will be far higher resolution.

Why not just use power from the radios 9-15V battery or an internal one?
Saves a lot of extra electronics and doesn't hurt the power out of the TX.
For intermittent use the battery would be small.

As to the EFHW antenna matching unit make it outboard in case
you use a different antenna. For 40-202m a T50-2 or T68-2 should
do the job there are lots of schemes on the net for doing this.

Allison
--
Please reply on the group, no private emails, we a group get to share info that way.

Even if the suggested electronics uses a watt on a 5W radio that's a lot of power not going to the antenna.

The suggested SWR meter shall draw roughly 1.5mA using a 1:2 or 1:3 transformer from the transceiver. Using a 1:2 transformer, with the assumption that only the positive polarity peak will be rectified, to source a 1.5mA voltage regulator, the circuit will draw 3mA from the source at the positive voltage peaks. Transformed by the 1:2 transformer, the current requirement will be halved, that is 1.5mA.

At 5W the maximum peak voltage at the transmitter terminals when loaded with an open resistive bridge will be 29.81V. That corresponds to a 45mW power loss mostly in the linear regulator, that is roughly 1% of the input power.
At 10W it will be 63mW.
At 1W the power drawn to source the SWR meter electronics will be 10V * 1.5mA = 15mW, that is still 1.5% of the input power, that corresponds to -18dB power lost.

The SWR bridge current requirements may likely be even a bit lower, the 1.5mA is quite a conservative estimate using lower CPU speed and highest efficiency LED diodes.

I would be thankful if you correct my thoughs if I am wrong. I am more a software engineer than an electronics engineer.

Vojtech



Vojtech

If your running a micro 1.5ma seems very low and LEDs you can see in
daylight want a few ma to themselves.

Also don't forget the transformer and rectifier is a load ont he transmitter and
that is in parallel with the antenna or dummy load which contributes to
SWR . That whole DC recovery has an impedance and what is that?
How will the transmitter behave with that addded (loss in power
and stability with a mismatched load).

There is no something for nothing in RF if your using power from the
transmitter you also contributing load and possible reactance to the
transmitters total load. Also the transformer, rectifier and power
converter do not deliver efficiency as components have loss.

That and the spec for the MPU used may hit low power but
soon as the IO and A/D does something that goes up.

If the power needed is so low a few coin cells would power
it for a very long time. A CR3020 cell is about 225mah.
That would save messing with RF and dealing with
inefficiencies of power conversion. You already have a
battery to run the transmitter why suffer the cost in
efficiency and losses (transfomer loss, rectifier loss,
regulator loss) to convert DC to RF and then back to
DC. Keep in mind that a CW rig running class E might
be 90% efficient but if it s class C its only 70% and
it can easily be lower. If SSB those linear amps are maybe
55% at full power and at less than half power 25% would
be good. So converting precious battery to RF then back
to DC is at best inefficient.

You are correct save for your also incomplete in your loss projections.
Transformers are not 100% efficient, same for diodes (.2V at 1.5ma is
0.3 mw of loss per diode). Since the diodes are switches they generate
harmonic waveforms that will be contributed to the output of the
transmitter. That's harder to predict as it varies with power in and
load power needs. In any case it needs to sty below the requirent
but the telecommunications rehulatory (FCC, PTT, other)
agencies as well as good engineering practice.

In the end you get pecked to death a little at a time by ducks.

I do hardware, software, as well as RF for the last 50 years.

Allison
--
Please reply on the group, no private emails we as a group get to share info that way.

There are two reasons to try it: First, it is challenge therefore fun. Second, it shall be an external box and it is great to not to worry about flat batteries.

1 mA flowing through a clear high intensity red LED makes my eyes hurt indoors.

Since the diodes are switches they generate

harmonic waveforms that will be contributed to the output of the
transmitter.

I will draw the current before the resistive bridge, therefore the harmonics will be attenuated. And the harmonics will be generated during the tune procedure only. I suppose that at 1.5% power consumed by the rectifier these harmonics will be quite weak.

I think it is time to test it. If I fail, at least I will learn something. Going shopping for the linear regulators.

Vojtech

Unless you expect your antenna impedance to change radically while you are transmitting, there is no need to continue monitoring your SWR once you have properly adjusted your antenna tuner.? There is a circuit now referred to as a Tayloe SWR bridge (Google it) which uses a resistive bridge during measurement of the SWR. While measurement is in progress it also supplies a protective resistance in series with the transmitter output, so the transmitter cannot see a dangerously mismatched impedance before adjustments are completed.
Once the SWR is measured you switch the bridge out of the circuit entirely, so every milliwatt of output goes to your tuner and antenna.
I? will modestly point out that I described this circuit in an April 1977? QST Hints and Kinks article. But I'm sure I'm not the first person to have thought of it.?

Al Woodhull?? N1AW ex-W1GSJ
Leyden, MA

I have a prototype working, see the attached photos showing the infinite SWR with open antenna terminals by lighting up the last LED, and showing a SWR roughly 1:1.1 with a 50 Ohm load. The bridge is powered with around 1.5W input power at 40m.

The six orange LEDs are interpolated by PWM, showing smooth transitions between levels. The SWR scale is following:
1st LED - 1:1
2nd LED - 1:1.5
3rd LED - 1:2
4th LED - 1:3
5th LED - 1:5
6th LED - 1:8 and higher

therefore the 1st LED being illuminated significantly more than the 2nd LED indicates SWR somwhere between 1:1 and 1:125.

On key up, the SWR meter is powered from a large 1000uF capacitor for around 1 second to blink the power detected on the 6 LED scale. I think that is all the functionality I can manage to squeeze into the 1kB of FLASH of this tiny $1 micro.

The device works somehow, but it is far from optimal. The micro is powered by a full wave rectifier with an artificial ground in the middle of the resistive bridge. As long as the current drawn is low, such arrangement should not cause significant read out errors or harmonics generated. I plan to do simulations of harmonic content in ltspice. The SWR bridge electronics is grounded to the virtual ground with a low impedance wire, therefore it shall be physically as tiny as possible to produce as little as possible capacitance against the other poles of the bridge, and this capacitance shall be balanced. The other three connections from the electronics to the SWR bridge are high impedance (1kOhm for power, 39kOhm for the forward and reflected power read outs).
I do not consider myself an expert in RF electronics, though I have a feeling that such a setup will be less than optimal above 20m due to the low impedance path between the electronics ground and the virtual ground at the bridge, but with today's tiny SMD circuitry the capacity may be lower than expected.

There is a lot of work to be done on fine tuning the firmware and possibly the circuit. The firmware should utilize power saving features of the micro to a full extent, it should run the A/D conversion with all the processing disabled to lower the noise, it should use averaging to improve read out SNR, the diode detector read out shall be linearised. Currently the electronics draws around 2.5mA from the transceiver with 1mA flowing through the LEDs. To lower the micro current below the planned 1.5mA, as low as possible Vcc shall be used, likely around 2.2 to 2.5V. The micro would work from 1.8V, but a little bit higher voltag is required to have enough of current regulation on the resistors powering the low voltage red LEDs. The CPU shall be operated at as low as possible CPU frequency.

It is an interesting engineering optimization task, it is a lot of fun, and it is manageable due to the simplicity of the circuit and the ATtiny13A micro. As always, I will be thankful for constructive criticism of my endeavors.

Vojtech OK1IAK

Hi Vojtek,

Would love to see the full document write up for the project. My guess is
you are using a resistive absortion bridge for direct/reverse wave sampling
and then a small microcontroller but I would love to see more info. TIA!

73, 72 de Juanjo, EC5ACA. EA-QRP #104, G-QRP #9742, QRP-L #1662,
FP #899.

Juanjo Pastor
C/San Roque, 4-1?
46460 Silla
SPAIN

e-mail: ec5aca@...
web:
web del club:
Tel.: +034 96 120 17 67
Móvil: 651 35 35 11


-----Mensaje original-----
De: [email protected] [mailto:[email protected]] En nombre de Vojtech
Bubnik
Enviado el: lunes, 16 de septiembre de 2019 9:26
Para: [email protected]
Asunto: Re: [qrp-tech] powering an SWR meter from transceiver output

I have a prototype working, see the attached photos showing the infinite SWR
with open antenna terminals by lighting up the last LED, and showing a SWR
roughly 1:1.1 with a 50 Ohm load. The bridge is powered with around 1.5W
input power at 40m.

The six orange LEDs are interpolated by PWM, showing smooth transitions
between levels. The SWR scale is following:
1st LED - 1:1
2nd LED - 1:1.5
3rd LED - 1:2
4th LED - 1:3
5th LED - 1:5
6th LED - 1:8 and higher

therefore the 1st LED being illuminated significantly more than the 2nd LED
indicates SWR somwhere between 1:1 and 1:125.

On key up, the SWR meter is powered from a large 1000uF capacitor for around
1 second to blink the power detected on the 6 LED scale. I think that is all
the functionality I can manage to squeeze into the 1kB of FLASH of this tiny
$1 micro.

The device works somehow, but it is far from optimal. The micro is powered
by a full wave rectifier with an artificial ground in the middle of the
resistive bridge. As long as the current drawn is low, such arrangement
should not cause significant read out errors or harmonics generated. I plan
to do simulations of harmonic content in ltspice. The SWR bridge electronics
is grounded to the virtual ground with a low impedance wire, therefore it
shall be physically as tiny as possible to produce as little as possible
capacitance against the other poles of the bridge, and this capacitance
shall be balanced. The other three connections from the electronics to the
SWR bridge are high impedance (1kOhm for power, 39kOhm for the forward and
reflected power read outs).
I do not consider myself an expert in RF electronics, though I have a
feeling that such a setup will be less than optimal above 20m due to the low
impedance path between the electronics ground and the virtual ground at the
bridge, but with today's tiny SMD circuitry the capacity may be lower than
expected.

There is a lot of work to be done on fine tuning the firmware and possibly
the circuit. The firmware should utilize power saving features of the micro
to a full extent, it should run the A/D conversion with all the processing
disabled to lower the noise, it should use averaging to improve read out
SNR, the diode detector read out shall be linearised. Currently the
electronics draws around 2.5mA from the transceiver with 1mA flowing through
the LEDs. To lower the micro current below the planned 1.5mA, as low as
possible Vcc shall be used, likely around 2.2 to 2.5V. The micro would work
from 1.8V, but a little bit higher voltag is required to have enough of
current regulation on the resistors powering the low voltage red LEDs. The
CPU shall be operated at as low as possible CPU frequency.

It is an interesting engineering optimization task, it is a lot of fun, and
it is manageable due to the simplicity of the circuit and the ATtiny13A
micro. As always, I will be thankful for constructive criticism of my
endeavors.

Vojtech OK1IAK

Hi Juanjo,

Thanks for your interest.

The project will be published completely on github. The firmware is just a working sketch as of now, and there is no printed board. I am in the process of studying details the ATtiny13a datasheet to fully make use of the power saving and A/D noise reduction features, while Dan OK1HRA (who runs ) is working on the PCB.

The circuit is so simple that you can basically read it out from the photos, hi.

Vojtech OK1IAK

Here is a preliminary documentation of the EFHW tuner with a real SWR meter:






The schematic drawings and printed board design are by OK1HRA.
The firmware works, but the diode curve calibration is not yet integrated, and power saving methods are not yet implemented into the firmware.

We will be thankful for constructive feedback on the circuit design and printed board layout.

73, Vojtech OK1IAK

Hi Vojtech,

Bunch of diode curves plotted here:


Which might be useful for checking spice computed curves.

Also of interest is section "1N5711 Response with 10 kΩ Load
with and without Compensation Amplifier" which demonstrates using an
operational amplifier with matched diode in feedback. Feedback diode
greatly compensated out diode voltage knee. Looking at red trace in the
plot, notice RMS V = detected voltage times a constant. therefore no
curve necessary.

Following Error plot suggests less than 0.3dB error down to 0.1 Vrms
input. Flat from 0.4 Vrms and above.

John KN5L

Hi John,

Thanks for the data. I will look into them.

I was reading a lot during the work commute (2x 45 minutes in public transport here in Prague every working day). I think I can increase the resistor values loading the detector diodes significantly, up to 100kOhm - 200kOhm. This will change the compensation curves again.

I am sure the compensation circuit would work, but this design of mine is an exercise in minimalism, and the op-amp would add to the size, price and current draw of the circuit.
My goal is a tiny back packing end fed tuner which will just work, no worries about the battery or cables to battery. I am not excited about the temperature dependence which would likely be compensated for by the circuit you suggested, but I can live with it.

Vojtech

I ran some simulation of the power recovery circuit in ltspice. With the assumption, that the controller driving the LED baragraph draws 1.2mA at 2.4V, the following peak currents of the power recovery diodes were simulated, and the following harmonic distortion values were simulated as well. Only 3rd harmonic levels below the carrier are shown as they are most pronounced. Long story short, the worst harmonic simulation was 3rd harmonic at -34dB below the carrier. I would say this is a pretty harmless value for a circuit, that is inserted between the transceiver and antenna only during antenna adjustment.

1.2mA at 2.4V is a realistic value if all the AtTiny13A power saving features are used and LED current is limited to 1mA.

At 1W input power:
At SWR 1:1, the diode peak current is 7.4mA, which is 3.7% of 200mA input peak current, and 3rd harmonic introduced by the power recovery circuit is -36dB below the carrier.
When antenna open, the diode peak current is 8,3mA, which is 6.1% of 135mA input peak current, and 3rd harmonic introduced by the power recovery circuit is -34dB below the carrier. This is the worst distortion value.
When antenna shorted, the diode peak current is 7mA, which is 2.9% of 240mA input peak current, and 3rd harmonic introduced by the power recovery circuit is -39dB below the carrier.

At 10W input power:
At SWR 1:1, the diode peak current is 11.5mA, which is 1.8% of 636mA input peak current, and 3rd harmonic introduced by the power recovery circuit is -36dB below the carrier.
When antenna open, the diode peak current is 13mA, which is 3% of 427mA input peak current, and 3rd harmonic introduced by the power recovery circuit is -35dB below the carrier.
When antenna shorted, the diode peak current is 11mA, which is 1.4% of 764mA input peak current, and 3rd harmonic introduced by the power recovery circuit is -36dB below the carrier.