开云体育

The $10 qrpguys dummy load looks good, about what Allison was suggesting except that?
they include a cap across the meter leads, Allison's scheme assumes the meter itself
provides the needed capacitance.
Their manual is worth looking over:??https://qrpguys.com/12w-dummy-load-power-meter
The qrpguys suggest calculating watts using? ?VoltsPk * VoltsPk / 100
where VoltsPk is half of the peak-to-peak voltage across the entire waveform.
That comes from? ?VoltsRMS * AmpsRMS = VoltsPk*0.707 * (VoltsPk*0.707 / Ohms)

On Mouser, you can get the needed parts for around a dollar if $10 is over your budget.
? ? 1x 511-1N5711? ?and 4x 283-200-RC
That Xicon 3W 200 ohm metal film resistor has many siblings at other resistance values.
If you get 20 of them at 1k in parallel, that's 50 ohms able to dissipate 60 watts in air.
Or a couple hundred in Jack's french fry oil.?
Though the 1n5711 diode will blow at reverse voltages in excess of 70v,
so max peak-to-peak voltage of the RF into 50 ohms is 70 volts, Watts is 70/2 * 70/2 / 100 = 12.25 Watts
when using the 1n5711.

The 1n5711 will drop around 0.2v at the currents involved.
Let's assume the meter reads 1 volt DC from the 1n5711, how much RF power?
? 1 * 1 /100 = 10mW
If we correct for the 0.2v drop across the 1n5711, we find that the power is actually
? 1.2 * 1.2 / 100 = 14.4 mW
which is close enough.? At the 1 watt level we are off by only 4%.
You could correct for most of this by simply adding 0.2v to the meter reading.
I doubt you need to calibrate anything.

No DC is present in the transmitter output, so no blocking cap is required.


On a related note, the RF diode probes that Arv points to in post 52474 are rather odd
? ??
though as he notes, "There are many variations on this design."
First one has the diode reversed from most, so the meter will see a negative voltage,
and good luck getting a GEX66 diode.? Otherwise, that seems a good design.
Second one has a a voltage doubler, though it also has two diode drops so no real gain in sensitivity or accuracy.
These diode probes are mentioned in most every radio handbook of the last 100 years.
And seldom explained.

Here's a diode RF probe that's more typical:?
The 1n34a diode is hard to get, many vendors just ship a Schottky.? I'd use a 1n5711.?
For those that don't want to follow the link, it's a 0.01 uF cap in series with the probe tip,
a 1n34a diode from ground to the other end of that cap, and a series 4.7 meg resistor from
that cap + diode junction out to the meter.? Simple.? But confusing as hell.

The cap blocks any DC coming in on the probe tip, but is zero ohms to AC.
So the cap + diode junction has an AC waveform equivalent to what is being probed.
The diode conducts whenever that junction goes below ground, so the bottom of that AC waveform
will have a bottom tip a diode drop below ground.? The average voltage of that AC waveform
will be about equal to the peak AC voltage at the probe tip.? The 4.7 meg resistor plus the stray?
capacitance inside the meter form a low pass filter, so the meter sees a DC voltage equal
to the peak AC voltage at the probe tip.? Many meters present an 11 ohm load to our RF probe,
so voltage is attenuated down to? 11/(4.7+11) = 0.7 of the peak voltage, and the meter reads
something approximating the RMS voltage of the AC assuming the AC is a sine wave.

I have these issues with the N5ESE probe:
1)? ?Not all meters are 11 meg
2)? ?Not all waveforms we look at are sine waves
3)? ?Output capacitance must be included when driving something like a Nano analog pin
4)? ?The dependence on meter input resistance and stray capacitance hides some complexity that few will ever figure out.

I'd go with the N5ESE circuit, except:
a)? Use the cheap and easily available 1n5711.? (Or BAT54S)
b)? Reduce the 4.7 meg to less than one tenth of the meter impedance so this reads honest peak voltages.
c)? Add a 0.01uF cap from output to ground to make that RC filter explicit.

We should standardize on something like that, feed it into A7? (Atmel claims a 100 meg input impedance)
Whenever anybody has a question about where between mike and antenna their uBitx is failing,
we will have a common instrument to debug it.? And hopefully get common results.

Oh, and also standardize on that Harbor Freight DVM.? ?;-)

Jerry, KE7ER



On Wed, Jun 20, 2018 at 08:54 am, Doug W wrote:

On Wed, Jun 20, 2018 at 08:36 am, ajparent1/KB1GMX wrote:

You need a load (10W) such as six 300 ohm 2 watt metal film resistors in parallel
and a 1n5711 diode to rectify the RF.

If anyone is looking for a kit here what I have ? sure you could piece together the parts for a little less but I didn't have everything on hand.
?

The $10 qrpguys dummy load looks good, about what Allison was suggesting except that?
they include a cap across the meter leads, Allison's scheme assumes the meter itself
provides the needed capacitance.
Their manual is worth looking over:??https://qrpguys.com/12w-dummy-load-power-meter
The qrpguys suggest calculating watts using? ?VoltsPk * VoltsPk / 100
where VoltsPk is half of the peak-to-peak voltage across the entire waveform.
That comes from? ?VoltsRMS * AmpsRMS = VoltsPk*0.707 * (VoltsPk*0.707 / Ohms)

On Mouser, you can get the needed parts for around a dollar if $10 is over your budget.
? ? 1x 511-1N5711? ?and 4x 283-200-RC
That Xicon 3W 200 ohm metal film resistor has many siblings at other resistance values.
If you get 20 of them at 1k in parallel, that's 50 ohms able to dissipate 60 watts in air.
Or a couple hundred in Jack's french fry oil.?
Though the 1n5711 diode will blow at reverse voltages in excess of 70v,
so max peak-to-peak voltage of the RF into 50 ohms is 70 volts, Watts is 70/2 * 70/2 / 100 = 12.25 Watts
when using the 1n5711.

The 1n5711 will drop around 0.2v at the currents involved.
Let's assume the meter reads 1 volt DC from the 1n5711, how much RF power?
? 1 * 1 /100 = 10mW
If we correct for the 0.2v drop across the 1n5711, we find that the power is actually
? 1.2 * 1.2 / 100 = 14.4 mW
which is close enough.? At the 1 watt level we are off by only 4%.
You could correct for most of this by simply adding 0.2v to the meter reading.
I doubt you need to calibrate anything.

No DC is present in the transmitter output, so no blocking cap is required.


On a related note, the RF diode probes that Arv points to in post 52474 are rather odd
? ??
though as he notes, "There are many variations on this design."
First one has the diode reversed from most, so the meter will see a negative voltage,
and good luck getting a GEX66 diode.? Otherwise, that seems a good design.
Second one has a a voltage doubler, though it also has two diode drops so no real gain in sensitivity or accuracy.
These diode probes are mentioned in most every radio handbook of the last 100 years.
And seldom explained.

Here's a diode RF probe that's more typical:?
The 1n34a diode is hard to get, many vendors just ship a Schottky.? I'd use a 1n5711.?
For those that don't want to follow the link, it's a 0.01 uF cap in series with the probe tip,
a 1n34a diode from ground to the other end of that cap, and a series 4.7 meg resistor from
that cap + diode junction out to the meter.? Simple.? But confusing as hell.

The cap blocks any DC coming in on the probe tip, but is zero ohms to AC.
So the cap + diode junction has an AC waveform equivalent to what is being probed.
The diode conducts whenever that junction goes below ground, so the bottom of that AC waveform
will have a bottom tip a diode drop below ground.? The average voltage of that AC waveform
will be about equal to the peak AC voltage at the probe tip.? The 4.7 meg resistor plus the stray?
capacitance inside the meter form a low pass filter, so the meter sees a DC voltage equal
to the peak AC voltage at the probe tip.? Many meters present an 11 ohm load to our RF probe,
so voltage is attenuated down to? 11/(4.7+11) = 0.7 of the peak voltage, and the meter reads
something approximating the RMS voltage of the AC assuming the AC is a sine wave.

I have these issues with the N5ESE probe:
1)? ?Not all meters are 11 meg
2)? ?Not all waveforms we look at are sine waves
3)? ?Output capacitance must be included when driving something like a Nano analog pin
4)? ?The dependence on meter input resistance and stray capacitance hides some complexity that few will ever figure out.

I'd go with the N5ESE circuit, except:
a)? Use the cheap and easily available 1n5711.? (Or BAT54S)
b)? Reduce the 4.7 meg to less than one tenth of the meter impedance so this reads honest peak voltages.
c)? Add a 0.01uF cap from output to ground to make that RC filter explicit.

We should standardize on something like that, feed it into A7? (Atmel claims a 100 meg input impedance)
Whenever anybody has a question about where between mike and antenna their uBitx is failing,
we will have a common instrument to debug it.? And hopefully get common results.

Oh, and also standardize on that Harbor Freight DVM.? ?;-)

Jerry, KE7ER



On Wed, Jun 20, 2018 at 08:54 am, Doug W wrote:

On Wed, Jun 20, 2018 at 08:36 am, ajparent1/KB1GMX wrote:

You need a load (10W) such as six 300 ohm 2 watt metal film resistors in parallel
and a 1n5711 diode to rectify the RF.

If anyone is looking for a kit here what I have ? sure you could piece together the parts for a little less but I didn't have everything on hand.
?

Though the 1n5711 diode will blow at reverse voltages in excess of 70v,
so max peak-to-peak voltage of the RF into 50 ohms is 70 volts, Watts is 70/2 * 70/2 / 100 = 12.25 Watts
when using the 1n5711.