How to verify the amplitude of a basic 100MHz module

[enut11]

I want to verify the high frequency amplitude response of my Tek 2445 CRO to 1% or better, if possible.
I have a 100MHz module specified as 13dBm (no tolerance) which I understand is 2.8Vpp. The module is powered from a regulated 5v supply.
The 2445 shows 2.34Vpp on the 50 ohm input which looks to be a bit low for a 150MHz instrument so I want to be able to verify the output of the 100MHz module before I draw any conclusions.
I do not have an RF power meter but I do have a Nano VNA and an accurate DMM.
Question: is there a simple way to verify the 100MHz module amplitude using basic equipment and/or perhaps a DIY circuit?
enut11

[moffy]

Found this: https://techlib.com/files/detect.pdf
while looking through this: https://www.eevblog.com/forum/projects/idea-for-super-peak-detector/
For the 'detect.pdf' circuit if you used a dual schottky diode it might be possible to use one as a temperature compensated reference for the offset.

[enut11]

Hi @edavid
The module outputs a sine wave.

[radiolistener]

You can use some chinese module with logarithmic-detector. They usually have about 60-90 dB usable dynamic range. But you're needs some reference source to calibrate it.

Something like this one: https://www.aliexpress.com/item/1005007116984716.html
Or this one: https://www.aliexpress.com/item/1005007221165402.html

If you want more professional equipment, then look for RF voltmeter.

[enut11]

Thanks. What type of 'reference source' do you have in mind?

[radiolistener]

Thanks. What type of 'reference source' do you have in mind?

It can be some oscillator with stable and known amplitude. For example you can use some TCXO powered from REF voltage source and ask someone who own calibrated RF voltmeter to measure it, then you can use it as reference source.

Another but less reliable and less precise way is to measure some low frequency oscillator (about 10 MHz) with 200-300 MHz bandwidth oscilloscope, it should give you result within about 5-10% error. Then use it to calibrate logarithmic detector module from such 10 MHz reference and then use logarithmic detector module to measure 100 MHz signal.

The logarithmic detector allows you to use low frequency (about 5-10 MHz) source for calibration and then measure much higher frequency.

If you have oscilloscope with >= 400 MHz bandwidth, then you can measure 100 MHz signal amplitude directly with oscilloscope. But oscilloscope measurement is not so precise, the error will be about 5-10% or even worse if frequency is more than 1/3 oscilloscope bandwidth.


Another option is to use tinySA Ultra, it can measure signal level, show signal spectrum and can be used as reference source, but its cost about 100 USD.

If you decide to use tinySA Ultra or logarithmic-detector, then also buy 20 dB attenuator and connect it through attenuator, because 2.34Vpp (which is +11.4 dBm at 50Ω) is pretty high level. With attenuator you will need to add attenuator value to measured signal power in dBm.

So, it depends on your budget.

[David Hess]

I want to verify the high frequency amplitude response of my Tek 2445 CRO to 1% or better, if possible.
I have a 100MHz module specified as 13dBm (no tolerance) which I understand is 2.8Vpp. The module is powered from a regulated 5v supply.
The 2445 shows 2.34Vpp on the 50 ohm input which looks to be a bit low for a 150MHz instrument so I want to be able to verify the output of the 100MHz module before I draw any conclusions.

Gaussian response for a 150 MHz oscilloscope should be down to about 84% at 100 MHz, or 2.32 volts peak-to-peak, so your measurement is consistent with likely errors.

As far as accurately measuring the output of your 100 MHz source, a sampling oscilloscope or sampling voltmeter would be my first choice.  Analog Devices makes some logarithmic RF detectors which should be accurate.  A compensated peak detector using microwave diodes should also work well.

[Postal2]

vacuum temperature converter.

http://lampes-et-tubes.info/sp/sp059.php?l=d

[enut11]

Thanks everyone. Some interesting replies, especially the last one!
The logarithmic detector solution proposed by @radiolistener does not appear to meet the 1% error requirement. However, the TinySA Ultra sounds like a wonderful tool to be explored at a later date.
@David Hess points out that my 2445 may already be in spec but it might be fun to verify that in some way.

As a first approximation for a viable solution I am going to build the circuit based on an article proposed by @moffy in Reply #1, dectect.pdf.
The response is shown as linear for RF signals greater than about 200mV.
enut11

[jonpaul]

see Tektronix serv man 2445, on testing, spec and repair of input mod.

Read 1980s Tektronox Circuits Concepts book


Oscilloscope vertical amplifiers

j

[radiolistener]

The logarithmic detector solution proposed by @radiolistener does not appear to meet the 1% error requirement.

1% is about 0.08 dB. With log detectors and good multi-point calibration I'm get about 0.001 dB resolution (which is incredible result from my point of view) with very flat response within 20 kHz - 200 MHz range and with pretty stable results. But I don't have so precise standards to test it's absolute precision. I'm using a little bit more complicated frontend than Chinese modules.

I think you will get higher error due to cable and connectors, which is hard to connect in that way to get two exact measurements for two different connections. It very sensitive to cable bending and connector contact.

[jonpaul]

Oscilloscopes freq resp and flatness are not the best assessment.

Most TEK 2445..2465..2467/A/B exceed the specified freq reesp.

We use transient response as a better guage of CAL.

Leo Bodnar 30..40 pS pulser>>MiniCkts HAT-20 20 dB atten>>Scope 50 Ohm input.

Tweak V amp etc for best transient resp as per CAL proceedure in scope serv manual.

For amplitude the standard TEK SG503, SG504, PG506, PG506A SG5010 CAL gen is fine.

Enjoy,


Jon

[radiolistener]

Here is my result taken with 1 meter Chinese RG58 cable with BNC connectors with AD8310, frontend schematic also attached. By the way touching BNC connectors may add offset up to 2 dB  So it needs to carefully check connection before measurements.

The picture shows frequency sweep over 1 kHz - 90 MHz range. Amplitude is calibrated and shown in dBm value. Power calibration was made at two points: -10 dBm and -40 dBm. The input signal is -50 dBm (0.001 Vpk).

[enut11]

Hi @radiolistener. Plot is obviously flat but I am at a loss as to how this can help me cal my CRO. I am a babe-in-the-woods wrt to RF circuitry.
enut11

[enut11]

see Tektronix serv man 2445, on testing, spec and repair of input mod.

Read 1980s Tektronox Circuits Concepts book


Oscilloscope vertical amplifiers

j

Hi @jonpaul
Thanks for the tip about the Tektronix Concepts articles. This will be a very worthwhile learning exercise for me.
I have built my own pulse generator based on a 74AC14 chip but this does not help with amplitude verification.
The Tek modules you specified for amplitude are too expensive to to buy from Australia hence my DIY approach.
enut11

[radiolistener]

Hi @radiolistener. Plot is obviously flat but I am at a loss as to how this can help me cal my CRO. I am a babe-in-the-woods wrt to RF circuitry.

log detector allows to compare 100 MHz amplitude with much lower frequency signal, like 1-10 MHz which is much easier to measure with oscilloscope or even with some DMM.

But if you want very precise result, its better to find someone who has RF voltmeter with enough precision.

Since you will use coax cable to connect your module, the coax cable should be properly terminated, the terminator impedance should be equal to cable impedance. If RF voltmeter has 50 Ω you need to use 50 Ω cable and connect it directly, with no terminators.

Proper termination is very important, since you need very high precision.

And I suggest you to measure environment temperature during this measurement, because your module output amplitude may vary depends on temperature.