|
He who measures the temperature with more than one thermometer doesn't know the temperature.
He who reads more than one time keeping device doesn't know what time it is.
He who measures the humidity with more than one instrument doesn't know what the humidity is.?
On and on and on....
Where do the error bars overlap??? The venerable Bird watt meter is only good to ¡À5% of full scale.?
I'd presume you used an attenuator in front of the SA to prevent burnout of the SA.? What is the error bar on the attenuator?
Dave - W?LEV
Brian,
Your measurement technique and math are both correct.?
(EE and extra class license here so I know enough to be dangerous!)
I looked up your SWR meter and it is spec'd as "Accuracy: mean +/- 5%".
I'm not exactly sure of what is meant by this, % of reading or % of full scale.
Instruments rated at % of full scale may not be very accurate at the low end of course.
So I looked for reviews for your power meter, some guys say it's accurate, some not.
Probably a lot of operator misunderstanding or error though.
Anyway, if I understand correctly you measure 4.35 watts with your power meter and 2.7 watts with the Tiny.
So 36.4 dbm and 34.3 dbm respectively.
If the power meter is reading 5% high then you would actually have only 4.13 Watts or 36.2 dbm.
If the Tiny is 10% low 2.95 Watts or 34.7 dbm would be your output.
36.4 and 34.3 is a difference of 2.1 dbm.? ?(Check my math - I didn't take the time to).
Assuming one device low and the other high brings the measurements closer but you still have enough difference to question a bit while realizing we can't expect HP/Agilent type accuracy out of these inexpensive but wonderful instruments.? I am very happy with my TinySA. Considering the low cost I think it's accuracy is fine for purpose. I do have an ancient HP432A? calibrated probably decades ago to compare it to and it's pretty good.
Hopefully you can come up with a 3rd option to narrow this down.? Maybe take your HT to a club meeting or hamfest and find someone with another instrument to try.? ?Good luck.
|
Brian, After you double checked with the cal output I can only say that your tinySA should be within the +/- 2dB error and it is not messed up. I can not explain the difference with your wattmeter. Maybe one more test you can do is to use the 40dB attenuator as the dummy load for your wattmeter so you can measure with the tinySA and the wattmeter at the same time using exactly the same load to your HT --
For more info on the tinySA go to https://tinysa.org/wiki/
|
On Sat, Dec 10, 2022 at 08:16 AM, Erik Kaashoek wrote:
You can check your attenuator using the tinySA.
I never would have thought of using the CAL function in this manner - works like a charm: ? -- John AE5X
|
On Mon, Dec 12, 2022 at 07:57 PM, W0LEV wrote:
He who measures the temperature with more than one thermometer doesn't know the temperature.
?
He who reads more than one time keeping device doesn't know what time it is.
?
Oh how true, I was very happy with my HP3400A. until I got a second one. ?There were some frequencies with a 16% discrepancy. ??????????????????????????????????????????????????? Mikek
|
On Tue, Dec 13, 2022 at 06:41 AM, John AE5X wrote:
On Sat, Dec 10, 2022 at 08:16 AM, Erik Kaashoek wrote:
You can check your attenuator using the tinySA.
I never would have thought of using the CAL function in this manner - works like a charm:
?
John, ? You can also measure the attenuation without the math by first normalizing the CAL output signal as below: CALOUT: [email protected]?  TRACE1->NORMALIZE ?  20dB Attenuator between CAL OUT and RF IN. ?  Herb
|
The main problem with this power measurements is that also if the
power meter and the attenuator are in good condition the
measurements can be vary a lot.
The reason is that a HT has in most cases not a true 50R output
impedance.
So if you connect the HT output to a load you have reflections
that can bring very big errors to the measurement.
This is the reason why I always use a feed through power sensor,
like the BIld or the R&S with a very very short connection to
the HT and after that a directional coupler with termination and
the SA.
So you can measure the real power much better.
Also it isn't in most cases of real help to know the output power
of a HT, because a HT is always a more or less matched pair with
the antenna. So what of real interest is, is the transmitted
energy, that is influenced by the surrounding environment a lot. A
good HT has a tolerant output stage that can handle a lot of
impedance fluctuations without producing distortion and noise.
Nothing that can be measured with a power meter or an SA by
themself.
A nice test setup is using a length variable stub to make a?
variable complex load and look what the power, the spectral purity
and the power consumption is doing. And then repeat this with a
whip antenna instead of the dummy load.
I have done some measurements with a 2m HT standing free, holding
with the hand, holding near the mouth, inside the pocket. After
this tests it was clear to me why a good old Motorola GP900 had a
better real live performance then a cheap HT. It is some years ago
that I have done that, but the basics don't change. Also I prefer
to use singe band low bandwidth HTs for real work. The receivers
are much better regarding out of band sensitivity.
When measuring whip antennas it is necessary to build a
counterbalance to the antenna. A can with a socket soldered inside
and a common mode filter on the feed cable is a good start.
One more point is that because of the unknown source impedance it
is possible that the meters show different values because they are
often not power meters, they are voltage meters with a 50R
resistor.
To conclude this. You can't use a HT as a measurement source if
it isn't carefully measured and specified for its behavior. At
minimum add a 3 dB attenuator directly without adapters to the HT
output to make the source impedance better, 10dB if you can
compensate that power loss in your measurement chain.
All the Best,
Isidro
Am 13.12.2022 um 17:25 schrieb hwalker:
toggle quoted message
Show quoted text
On Tue, Dec 13, 2022 at 06:41 AM, John AE5X wrote:
On Sat, Dec 10, 2022 at 08:16 AM, Erik Kaashoek wrote:
You can check your attenuator using the tinySA.
I never would have thought of using the CAL function in this
manner - works like a charm:
?
John,
? You can also measure the attenuation without the math by first
normalizing the CAL output signal as below:
CALOUT: [email protected]
?
TRACE1->NORMALIZE
?
20dB Attenuator between CAL OUT and RF IN.
?
Herb
|
Excellent explanation.? Brings back many lab experiences.
?
Kind Regards,
Dave
Dave Reiser, WA3TRS
?
toggle quoted message
Show quoted text
From: [email protected] < [email protected]>
On Behalf Of Isidro Berniol
Sent: Tuesday, December 13, 2022 5:30 PM
To: [email protected]
Subject: Re: [tinysa] #tinysa Power Reading
?
The main problem with this power measurements is that also if the power meter and the attenuator are in good condition the measurements can be vary a lot.
The reason is that a HT has in most cases not a true 50R output impedance.
So if you connect the HT output to a load you have reflections that can bring very big errors to the measurement.
This is the reason why I always use a feed through power sensor, like the BIld or the R&S with a very very short connection to the HT and after that a directional coupler with termination and the SA.
So you can measure the real power much better.
Also it isn't in most cases of real help to know the output power of a HT, because a HT is always a more or less matched pair with the antenna. So what of real interest is, is the transmitted energy, that is influenced by the surrounding
environment a lot. A good HT has a tolerant output stage that can handle a lot of impedance fluctuations without producing distortion and noise. Nothing that can be measured with a power meter or an SA by themself.
A nice test setup is using a length variable stub to make a? variable complex load and look what the power, the spectral purity and the power consumption is doing. And then repeat this with a whip antenna instead of the dummy load.
?
I have done some measurements with a 2m HT standing free, holding with the hand, holding near the mouth, inside the pocket. After this tests it was clear to me why a good old Motorola GP900 had a better real live performance then
a cheap HT. It is some years ago that I have done that, but the basics don't change. Also I prefer to use singe band low bandwidth HTs for real work. The receivers are much better regarding out of band sensitivity.
When measuring whip antennas it is necessary to build a counterbalance to the antenna. A can with a socket soldered inside and a common mode filter on the feed cable is a good start.
?
One more point is that because of the unknown source impedance it is possible that the meters show different values because they are often not power meters, they are voltage meters with a 50R resistor.
?
To conclude this. You can't use a HT as a measurement source if it isn't carefully measured and specified for its behavior. At minimum add a 3 dB attenuator directly without adapters to the HT output to make the source impedance
better, 10dB if you can compensate that power loss in your measurement chain.
?
All the Best,
Isidro
?
?
Am 13.12.2022 um 17:25 schrieb hwalker:
On Tue, Dec 13, 2022 at 06:41 AM, John AE5X wrote:
On Sat, Dec 10, 2022 at 08:16 AM, Erik Kaashoek wrote:
You can check your attenuator using the tinySA.
I never would have thought of using the CAL function in this manner - works like a charm:
?
John,
? You can also measure the attenuation without the math by first normalizing the CAL output signal as below:
CALOUT: [email protected]
?
TRACE1->NORMALIZE
?
20dB Attenuator between CAL OUT and RF IN.
?
Herb
|
I've also found the output power of an HT is dependent on the charge level of the included batteries.? This can change from one transmission to another transmission and even within a long-winded transmission.?
Before over analyzing the two measurements, consider all the variables that go into the two measurements.?
1)? Variable HT output power with battery condition. 2)? Variable HT output power with heating of the output and driver stages. 3)? Accuracy of the attenuator(s) used in the measurement(s).* 4)? Accuracy of each of the two measurement devices. 5)? Consistency of the coaxial cables used in the two measurements.* 6)? What ever I've forgotten or not lested.
*? Certainly the attenuator was used for the TinySA measurement.? Possibly it was not used for the additional measurement.? Also, different or additional coaxial cables were used for the two differing measurements.?
Dave - W?LEV?
The main problem with this power measurements is that also if the
power meter and the attenuator are in good condition the
measurements can be vary a lot.
The reason is that a HT has in most cases not a true 50R output
impedance.
So if you connect the HT output to a load you have reflections
that can bring very big errors to the measurement.
This is the reason why I always use a feed through power sensor,
like the BIld or the R&S with a very very short connection to
the HT and after that a directional coupler with termination and
the SA.
So you can measure the real power much better.
Also it isn't in most cases of real help to know the output power
of a HT, because a HT is always a more or less matched pair with
the antenna. So what of real interest is, is the transmitted
energy, that is influenced by the surrounding environment a lot. A
good HT has a tolerant output stage that can handle a lot of
impedance fluctuations without producing distortion and noise.
Nothing that can be measured with a power meter or an SA by
themself.
A nice test setup is using a length variable stub to make a?
variable complex load and look what the power, the spectral purity
and the power consumption is doing. And then repeat this with a
whip antenna instead of the dummy load.
I have done some measurements with a 2m HT standing free, holding
with the hand, holding near the mouth, inside the pocket. After
this tests it was clear to me why a good old Motorola GP900 had a
better real live performance then a cheap HT. It is some years ago
that I have done that, but the basics don't change. Also I prefer
to use singe band low bandwidth HTs for real work. The receivers
are much better regarding out of band sensitivity.
When measuring whip antennas it is necessary to build a
counterbalance to the antenna. A can with a socket soldered inside
and a common mode filter on the feed cable is a good start.
One more point is that because of the unknown source impedance it
is possible that the meters show different values because they are
often not power meters, they are voltage meters with a 50R
resistor.
To conclude this. You can't use a HT as a measurement source if
it isn't carefully measured and specified for its behavior. At
minimum add a 3 dB attenuator directly without adapters to the HT
output to make the source impedance better, 10dB if you can
compensate that power loss in your measurement chain.
All the Best,
Isidro
Am 13.12.2022 um 17:25 schrieb hwalker:
On Tue, Dec 13, 2022 at 06:41 AM, John AE5X wrote:
On Sat, Dec 10, 2022 at 08:16 AM, Erik Kaashoek wrote:
You can check your attenuator using the tinySA.
I never would have thought of using the CAL function in this
manner - works like a charm:
?
John,
? You can also measure the attenuation without the math by first
normalizing the CAL output signal as below:
CALOUT: [email protected]
?
TRACE1->NORMALIZE
?
20dB Attenuator between CAL OUT and RF IN.
?
Herb
|
The reason is that a HT has in most cases not a true 50R output impedance.
So if you connect the HT output to a load you have reflections that can bring very big errors to the measurement.
That is not the case. Most Pas are not 50ohm, they are designed to give a specified power 'into* a 50 ohm load. The o/p impedance of the PA has no effect on any reflections from the load, that is entirely dependent on how well the load matches the transmission line. The only time the PA impedances comes into play is what happened to any power that is reflected from the load that arrives back at the PA. 73 Jeff G8HUL
|
On Wed, Dec 14, 2022 at 10:12 AM, G8HUL wrote:
Most Pas are not 50 ohm, they are designed to give a specified power 'into* a 50 ohm load.
Jeff,
?
That's right, impedance matching is necessary for sources that can only deliver a small amount of electrical power and whose power is to be passed on as comprehensively as possible. Such weak sources are, for example, antennas. With linear sources and sinks, in the best case only half of the available power is actually radiated; the efficiency is then 50 %. The other half is lost as power dissipation in the internal resistance of the source.
Electrical energy sources, e.g. a power station, but also a large PA - whether audio or high frequency - are therefore not designed for power matching, but for voltage matching (i.e. low source resistance); their nominal power would therefore be exceeded long before any (theoretical) power matching.
?
Martin
?
|
My understanding is that any time there is a change in impedance, there is a reflected wave.? This is the definition of Gamma, the Reflection Coefficient.? The RF power amplifier has a certain impedance and by whatever method is used to remove that RF power, it also has an associated impedance.??
At that interface, there is reflected power unless both impedances are equal.? If you choose to calculate the efficiency of the power amplifier using RF power out divided by DC power input the power reflected by the interface impedance is already in the mix.??
That power amplifier efficiency can always be increased or decreased by changing either the output impedance of the power amplifier or the input impedance to whatever is on the other side of the interface.? Once those impedances are equal, there will be no reflected power and the calculated power efficiency will be at it's highest.
Every time there is a step in impedance, there is a reflected wave.? It doesn't matter where that impedance step is located, it will have it's effect on the final power delivered to the load.
Regards,
Jim
Logic: Method used to arrive at the wrong conclusion, with confidence.? Murphy
-----Original Message-----
From: Ho-Ro <linuxaudio@...>
To: [email protected]
Sent: Wed, Dec 14, 2022 4:02 am
Subject: Re: [tinysa] #tinysa Power Reading
On Wed, Dec 14, 2022 at 10:12 AM, G8HUL wrote:
Most Pas are not 50 ohm, they are designed to give a specified power 'into* a 50 ohm load.
Jeff,
?
That's right, impedance matching is necessary for sources that can only deliver a small amount of electrical power and whose power is to be passed on as comprehensively as possible. Such weak sources are, for example, antennas. With linear sources and sinks, in the best case only half of the available power is actually radiated; the efficiency is then 50 %. The other half is lost as power dissipation in the internal resistance of the source.
Electrical energy sources, e.g. a power station, but also a large PA - whether audio or high frequency - are therefore not designed for power matching, but for voltage matching (i.e. low source resistance); their nominal power would therefore be exceeded long before any (theoretical) power matching.
?
Martin
?
|
DAVE REISER WA3TRS
Ho-Ro