Greetings all,

?

Having recently purchased a Siglent SDS1202X-E, I am faced with 2 immediate questions.

?

1 – What is the safest and most accurate way to measure output power? I would think that a T-adapter (placed right before the dummy load) and a 10X probe would be the way to go, however, there is the issue of accessing the ground and signal lines. An inline unit with places to clip the ground clip to and then either hook on the probe and/or ‘the probe tip can’t slide off’ sounds good, but before I make one up, what other options are in use? Considering that I want to be able to (at least) measure up to my 50W PA at full output (141 Vpp into a 50-ohm load), connecting a T-adapter to one of my ‘scope’s inputs does not sound like a good idea…

?

2 – When probing QCX+/QDX/50W PA units with my DMM, finding a place for the ground probe is fairly simple. However, when looking for a place to clip the 'scope's probe’s ground to, it’s much harder. For those who have assembled those units and analyzed them with a ‘scope, to where have you connected your probe’s ground clip?

?

I appreciate the help!

?

Charles Johnson

KF4AYT

Doing it fairly properly :
tx to dummy load via BNC T connector at the load end, shortest leads possible?
10x scope probe into open socket port of T connector and earth clip to T connector surround. ? Let scope calculate RMS voltage if possible. ? ?V^2 /50

Check the reading is zero with no TX or you have a hum loop and it’s affecting the readings?

if you’re not measuring into 50 ohms all cakcukations are off?

Re grounding
Solder some thick wires or pins from the antenna sockets ground and a few other GND points around the board?

(Don’t get suckered by the idea that the impedance at the antenna is 50.0 ohms )

According to the datasheet your oscilloscope is rated for up to 400 volts peak, or 800 volts peak-to-peak. The limiting factor will be the oscilloscope's ability to attenuate the signal into its actual measurement range, but you're certainly not going to break it. The limiting factor will be whether the oscilloscope has a measurement range that will fit the full waveform. My own cheap oscilloscope is rated for 300 VRMS, but can only display a signal up to 80 volts peak to peak.

I think in your case what I would do is build yourself an attenuator with an L-pad in it, 10k ohms in series and 50 ohms in parallel, and hook it up to a T just in front of the dummy load. Turn on the 50 ohm terminator in your oscilloscope input (rated for up to 0.5 watt) and blammo. You can now safely measure up to 100 watts.

To expand on the previous:

Your PA will see a 10k load in parallel with the 50 ohm dummy load. It won't even notice the extra VSWR.

The attenuator will have an output impedance of 50 ohms, so it will match just fine into your 50 ohm coax and 50 ohm oscilloscope termination resistance.

Yes! ?Put a 39k resistor in parallel with that 10k and you have 50dB attenuation ?

FWIW I use that set up together with an AD 8307 style (set up for -60dBm to 0dBm) power meter. ?That 50dB attenuation then gives me 1mW to 100W. ?

That half watt rating would apply to the 10k resistor (for 100W measurement) by the way.

73

Rod
G0VKX

?

The DIY pad is unnecessary and very special resistors and construction are needed to stay linear and resistive at RF. ?It will give imprecise results at best and as described may also introduce SWR issues in the cable between it and the measuring device?

Hopefully your scope came with a good x10 probe (!). That is absolutely the best way to use a scope to sample RF voltage and on x10 a good one will not bother the main signal much or deliver too much voltage to the scope input . ?Scope probes /cables are specially constructed for exactly this purpose and to minimise SWR problems inside their own cable?

(If you want more than ‘fairly properly)
For greater precision and/or use with a spectrum analyser then using a 100W 50ohm through attenuator of 30 to 40dB is the way to go but the price of a good one is only for the very enthusiastic.)

I have one of sitting in front of me right now. Very unsophisticated construction, regular old off the shelf 1-watt resistors, the nanoVNA tells me it's within 1 ohm of the bullseye on the smith chart all the way up to 20 MHz. Seems to me that HF and below is pretty forgiving.

开云体育

SWR varies from 1.03 at 3 MHz to 1.1 at 30MHz, terrible.

It's got an attenuator - a 2k2 resistor into 2 X 100, this gives:

I've set the scale to show you how terrible it is, varying between -38.5 and -39 dB

Obviously totally useless.

Chris, G5CTH

开云体育

I bought a 10W 20dB attenuator from eBay for about ?23, wanted something to keep the spectrum analyser I've borrowed from my radio club happy.

This is much better:

This looks good enough to get reasonable measurements on a QRP transceiver.

Chris, G5CTH

FWIW? ?The trouble with reading power accurately . . .?

A 1 : 1:05 SWR from a poor match or system implies a voltage reading across 'supposedly 50 ohms' that's 5% different to nominal/expectation for that power level.?

That 5% measurement error becomes 5% squared (10.2%) as you move to reading it as power.? ?

Add your scope error and your power calculation can easily be 15 to 20% out unless you have a way to calibrate your entire measurement system individually.

This would not work at all in a professional environment but in amateur use of course you can decide what's good enough for you.? ?

Eek, I didn't realize that such small mismatches would have such a huge effect on accuracy... Consider myself corrected.

Check the accessories that came with your probes!

That scope (like most) usually comes with special little probe tips to plug them in to a BNC/TNC connector. It looks funny because it has no outside latch ring - it connects to the INSIDE of the shield.

I find it works great when the probe is set to 10x with a proper BNC TEE connector (from old RG58 networking days) and a good dummy load up to 50Mhz or so. Beyond that, the response drops off (due to the limited bandwidth of the probe's built in attenuator).

For a really good time, see my video series about using your scope to display harmonics and sidebands using the built-in FFT functions at ?. Most of the functionality of the 1204X-e for FFT is present, but some of the advanced marker features are not available in the 1202.

My 1204 will even display the FFT calculated power in a measurement field. Note that power measurements with an FFT are limited to the FFT unit resolution, but it's still really handy for tweaking! I often add another trace for the voltage to be displayed in a measurement field on screen.

This is also a great setup for examining keying rise and fall times and other behavior.

I quite like my Siglent 1204X-e for this work, but don't forget to save you setup so you can reuse it next time you need it...



M

开云体育

I see this as an example of the perfect being the enemy of the good.

We are trying to get an idea of the performance of a QRP transmitter.? We aren't measuring the temperature of the big bang. We aren't in a metrology lab.

We want an idea of how much power we have, and if a change made it better or worse.? Blanket statements that it can inly be done with great difficulty by experts using hugely expensive kit doesn't help.

Chris, G5CTH

I understand your frustration Chris, but the point of this hobby (and I believe Hans' product line) is experimentation and you cannot do that without at least SOME investment in test equipment and tools. In fact, the equipment in question is in the same price range as the products sold by QRPLabs!

Nobody in this thread is saying you MUST buy any particular test equipment.

If you need an inexpensive solution, try starting a thread with your question as the topic. If you look up at the topic of this one, it's about using an oscilloscope to measure power. There are other ways...

Sorry we hurt your feelings.

M

The scope mentioned by the OP does not have a 50 Ohm input impedance selection. He needs to get a termination adapter, but the best, most accurate measurement is by using a commercial attenuator with a terminator rather than a dummy load and tee or homebrew, uncalibrated, frequency dependent resistive attenuator. I have 20 and 100 Watt 20dB attenuators that I can follow with various other attenuators to get a safe level for harmonic and power analysis with a TinySA spectrum analyzer or power meter.
Gary
W9TD

Your suggested alternative for tapping input is also a good solution, but there's absolutely nothing wrong with the method I proposed of using the high impedance probe and a cable tee below 50Mhz. It's perfectly flat if used with the supplied probes.

I don't mean to shout but this is important: IT MUST NOT BE USED WITH 50 Ohm TERMINATED CABLES or 1x probes. The probe adapter/dummy load is designed for use with the manufacturer supplied HIGH IMPEDANCE PROBE. Trying to run it at 50 ohms will result in a 25 ohm load and 50% of the power going into your scope cable and terminator.

It is NOT flat for VHF work and can be 3db down at 2 meters.

I showed both methods in my videos. I use a 50 ohm inline termination preceded by a 100w 10db attenuator and further 10 watt attenuators as required for working above the QRP level with this device and my other test equipment. This approach is more comfortably safe for 100watt transmitters and is more accurate at 2 meters and above. The scope in question is only rated to 200 Mhz.

No offence, but please don't give or criticize advice from others that you have not actually tried or demonstrated yourself.

M

On Tue, Sep 12, 2023 at 11:22 AM, Gary W9TD wrote:

I have 20 and 100 Watt 20dB attenuators

If terminated by 50Ω, a 20dB pad makes a reasonably precise 10:1 voltage divider.? To do this I use a "50Ω feedthru termination" between the pad and 'scope input.? You can check/calibrate it with DC.? This one??should be good thru UHF at least.? This one claims 1GHz:??

73, Don N2VGU

On Tue, Sep 12, 2023 at 10:11 AM, Stephan Ahonen KE0WVA wrote:

I have one of sitting in front of me right now. Very unsophisticated construction, regular old off the shelf 1-watt resistors, the nanoVNA tells me it's within 1 ohm of the bullseye on the smith chart all the way up to 20 MHz. Seems to me that HF and below is pretty forgiving

Absolutely so.

As a further aside, the interesting if slighly simmering discussion around performance of typical amateur homebrew "bog standard" attenuators and dummy loads does leave me pondering quietly quite how much an error of, say, 1dB ever really matters.? They all "add up" of course and so we want to get the best we can out of every stage of a radio contact, but in absolute terms we're for the most part exploiting (battling?) an ionosphere that varies constantly in wonderful ways.? Signal to noise ratios, it seems to me, are the thing - and what's between our ears makes an ultimate measurement there, of intelligibility.

The numbers in power measurements feel worrying perhaps, 7dBW plus or minus or dB is one thing, where the range from 4W through 5W and up to 6.3W may to some seem a huge uncertainty.? But it's just one dB.? The losses through the "aether" from one's QTH to our contact at the other end are likely to be much more uncertain (at HF at least) and we can work to do the best we can without knowing the absolute figures.

Hans provides us with built in test equipment to get best performance out of our builds - it would be a great shame if beginners felt that they needed to invest a lot more to enjoy this wonderful hobby.? Just my two pennyworth.

73

Rod

G0VKX
one-time G8FJN