Hello, I'm looking for feedback to see if I am correct in my findings with regards to a SWR power meter.
I wanted to set the output of my HF amplifier to 5 watts. I checked this with an Oscope and a power meter.
1.) I connected my amp to my Oscope (1 MOhm) via a T-BNC connector.
2.) From the T-BNC connector I connected to the SWR power meter
3.) From the SWR power meter I connected to a dummy load terminated at 50 Ohms.
I increased DC power to my amp until my Oscope read (22.4V | 24dBV) - 5 watts at 50 Ohms
My SWR meter only displayed about 3.5 watts.
This leads me to believe my SWR meter is off. Is my method correct?
Well, 3.5W instead of 5W means a voltage error of maybe 15%. You said HF so I am assuming short cables. Most SWR meters (I am guessing) would have a calibration pot so maybe that's what needs adjusting. If the frequency is much above 10 MHz I would be cautious about what to do about it. At 30 MHz the dummy load may not be 50 Ohms so it's wise to check it.
I didn't do the math but 22.4 V seems not right. Is that p-p? And what kind of amplifier can be running at 5W out? How much drive are you giving it? Is the wave sinusoidal?
There's a couple bits of trickery here and I can't read the O-scope display to be sure. Your SWR meter is supposed to measure only the power delivered to the load, and the SWR measures the power reflected from the load. It appears to be zero reflection or the meter is broken. But the Oscope measures the total voltage at the point on the line, so if you moved the Tap point along the line you would see a peak and a valley value and the ratio of those two values is the definition of SWR. But if the SWR meter reflection measure is correct, you don't have any standing wave. The upper plot on the O-Scope shows the sine wave value and normally you would need to take the peak-to-peak value divided by 2 (to get Peak voltage) and divided by square root of 2 to get RMS value of the voltage. I can't read the amplitude of the O-scope but if you did take 1/2 peak to peak as the 22.4 volts then the O-scope is reporting a higher power than the SWR meter. funny thing is you say you see 22.4 V / 24dBV but normally we use 20Log10(V) to get dBV, so 22.4 V peak would compute to 27 dBV. If the dBV reading is from the FFT, it might be already dBV rms (22.4/sqrt(2)). You might try to move your T point closer or farther from the antenna (use a shorter or longer length) and see if the voltage changes. If it does, it means you DO have a non-zero reflection and you do have some SWR, and the SWR meter has an issue. I suspect a bit the SWR meter as zero reflection is a little suspicious. You could try lowering your power to a very safe level and take off the load, to see if the SWR jumps up. If it does not, suspect the SWR meter.
22,4Vp, so 15,8Vrms equals 5W into 50
- so far everything seems to be correct.
my guess would be that the accuracy of the swr/power meter in the lower range (1-10W) is less than in the higher range (10-100W). probably that can be corrected by an internal trimmer, but I don't know the meter; maybe a manual gives more infos about that.
and don't use the FFT function of the scope to get dB-values; they are certainly way off the real world values. This would be a task for a spectrum analyzer or a power meter based on a logarithmic amp.
Okay, this response makes the most sense. I'm at 13.56 MHz. that's probably on the lower end for this meter. I'll see if I can open it up and check around for some calibration plots. Just wanted to make sure my method for approaching this problem wasn't off.
One note that you might want to pay attention to ...
The SWR of the SWR meter itself.
Nothing is an absolutely perfect impedance match. Impedance mismatch = SWR from that device. This gets more critical as frequency goes up.
I discovered this with my MFJ 864 SWR meter on the 2 meter band.
I then connected my VNA to the transmitter port and my calibration load to the antenna port. ACK! Not happiness!
This encouraged me to test every SWR meter I had in house. The VERY old Heathkit SWR meters WAY outperformed the MFJ model.
I extended this investigation to antenna switches and all that stuff that I put into my feedline without thinking. I also got looking at my various dummy loads.
This made a HUGE difference in several ways.
So, I thought this was something I needed to share with my subscribers on my Electronics for the Inquisitive Experimenter YouTube channel. You can view the video here:
https://youtu.be/1fpoViuxn3c
The classic approach to a problem like this is to perform a measurement uncertainty analysis for every item in the calibration system.
I see quite a few contributors to measurement uncertainty in your system. First of all, the FFT window function in your scope probably has 1dB of uncertainty unless it is the flat top function. I would not use an FFT to measure power accurately.
Second, trying to compute Vrms from Vpkpk on the scope sinewave will also have uncertainty because harmonics can introduce uncertainty. Harmonics at -30dBc can introduce an uncertainty window of about +/- 6% for power if you convert from Vpkpk to Vrms. Probably better to let the scope compute Vrms from its sampled data rather than do it the clunky Vpkpk way yourself. The accuracy limitations of the scope could easily add +/- 3% uncertainty in the voltage measurement and power is V^2/R. Errors in voltage measurement bite you hard with your current system.
The dummy load has to be an accurate 50R or it will introduce uncertainty. A 5% error in the load translates across to an error in the power calculation if you calculate power from voltage across the load.
My advice would be to abandon every piece of hardware in your current calibration setup and make yourself a nice low pass filter for 14MHz that has very low insertion loss. This removes uncertainty due to transmitter harmonics. It should be possible to get about 0.1dB insertion loss and very low VSWR in the filter at 14MHz. You ideally want to get all harmonics to be better than -50dBc.
Then make a precision 50R load using good quality resistors that are measured for accuracy on a DMM. You only need a load rated at a few watts. Then attach a 1N5711 schottky diode detector to the 50R load and this will measure the peak RF voltage to within 1 diode drop. Then transmit into this load with your PA and measure the DC voltage at the diode detector with a decent DVM.
The power delivered to the load will be approx ((Vdet +0.3)^2)/100 Watts. This should have an overall uncertainty of below +/- 2%. Then transfer the transmitter to the SWR meter and terminate it with the same load or a decent load. The accuracy of the load will be less critical here. Then calibrate your power meter to the known RF power level you just calculated.
I increased DC power to my amp until my Oscope read (22.4V | 24dBV) - 5 watts at 50 Ohms
My SWR meter only displayed about 3.5 watts.
This leads me to believe my SWR meter is off. Is my method correct?
yes, your method is correct. Just make sure 22.4 V is amplitude, so peak-to-peak voltage should be 44.8 V.
For SWR meter it is also ok to have some error, especially at low power. For example, if your SWR meter is rated to 100-200 W and you're trying to measure 5 W, your SWR meter can have some measurement error. This is ok.
Also note, that your amplifier can have harmonics, so classic formula for power/amplitude of sine wave can produce error. You can use RMS voltage measurement on your oscilloscope to resolve that.
Calibrating SWR meter does not make much sense. If you look at its specification sheet, most models say it's 5%. I have NEVER seen one that is actually 5% accurate. Also, response per frequency is not flat at all.
I think of SWR meter as an indicator, not a measuring instrument. Unless there is a specific need to get it exactly 5 watts (at a given frequency), I wouldn't bother. Just take it at face value and call it 5 watts....
KB4EMF/JF2DKG
Taka
Well, 3.5W instead of 5W means a voltage error of maybe 15%. You said HF so I am assuming short cables. Most SWR meters (I am guessing) would have a calibration pot so maybe that's what needs adjusting. If the frequency is much above 10 MHz I would be cautious about what to do about it. At 30 MHz the dummy load may not be 50 Ohms so it's wise to check it.
I didn't do the math but 22.4 V seems not right. Is that p-p? And what kind of amplifier can be running at 5W out? How much drive are you giving it? Is the wave sinusoidal?
Which seems to match what's reported on eHam
https://www.eham.net/reviews/view-product?id=14770they say the meter is off by about 20%, and if that's true with the meter set to 15W max input, then it's quite a lot imHo
[edit]
and here are some notes about correcting issues with a CN-101L, different from the CN-501H, but I wonder if the latter may use a similar schematic
https://pa0fri.home.xs4all.nl/Diversen/CN101L/CN101Leng.htmI wonder if the 501 is just a new "incarnation" of the 101 with a slightly different front panel
as for calibration
open the meter
connect the meter to a tx and a dummy load
set the meter to 15w
drive it with 5 or 6w
adjust the fwd "lo" range trimmer
swap (reverse) the dx and dummy load connections
repeat the above, but this time adjust the ref "lo" trimmer
It is not 100% clear from the photo but the reflected wave needle may need to be null adjusted using the screw on the front panel , as it appears to be below the null point.