I can get to the "Reference Level" menu item but I can't seem to change the actual position of the triangular REF LVL marker on the left side of the display. I'm trying to measure the gain of a preamp that should have 10-20dB gain.

What's the valid range for this field and what do the values mean?
What happens if you enter a value that's illegal or too big or too small?

All of my attempts seem to be ignored. I checked the on-line documentation (as of a few days ago) and while it tells me how to get to it in the menu system and what indicator shows its position, that's all it says about it.

I'm used to a spectrum analyzer where you set the REF LVL as the dBm value that you want the top line to represent.

Bob,
On the NanoVNA, the term used is "Reference Position", and it is the location (in screen divisions) of the 0dB position on the screen. To change it, make sure the trace you want is selected (its name should be in reverse video at the top of the screen). (Select "DISPLAY/ TRACE/TRACE 1" to select or toggle Trace 1 display, if needed.) "DISPLAY/SCALE/REFERENCE_POSITION" lets you change the display position of the 0dB (or whatever) of the trace.
If you are measuring devices with gain, beware of limited dynamic range. Too large an input will overload the mixers. The NanoVNA is set up to most easily handle passive devices. To measure a preamp, put an attenuator of about the expected gain on the input of the preamp and mentally adjust the measured gain to account for the attenuation. This will also avoid overdriving your preamp. If you want to measure the input characteristics of your preamp, you might need to put an attenuator of 10dB or so on Port 0, do your calibration at that point and then connect the preamp. This will avoid overloading the preamp input and possibly obscuring its true input impedance.
--John Gord



On Tue, Sep 1, 2020 at 01:16 PM, Bob M. wrote:

>
> I can get to the "Reference Level" menu item but I can't seem to change the
> actual position of the triangular REF LVL marker on the left side of the
> display. I'm trying to measure the gain of a preamp that should have 10-20dB
> gain.
>
> What's the valid range for this field and what do the values mean?
> What happens if you enter a value that's illegal or too big or too small?
>
> All of my attempts seem to be ignored. I checked the on-line documentation (as
> of a few days ago) and while it tells me how to get to it in the menu system
> and what indicator shows its position, that's all it says about it.
>
> I'm used to a spectrum analyzer where you set the REF LVL as the dBm value
> that you want the top line to represent.
>

Thanks. Sorry for mis-labeling it. Too many years using an SA.

Yes, the preamp has about 13dB of gain and It would be helpful to add some attenuation. Unfortunately I have no such attenuators with SMA connectors. I did want to measure the gain and eventually the return loss and impedances.

I also didn't find anything that controls the output level from Port2 that would be feeding the input of the preamp. Again, with my SA and TG, I have full control of both of those levels, and that's what I'm looking for, and it just may not exist in the VNA world.

Bob,
Fancy VNAs ($$$) usually have variable attenuators built in. The NanoVNA has some very limited control of source level through the output current setting of the synthesizer chip. I am not sure the NanoVNA-V2 has this. You are probably going to need some external attenuators.
--John Gord

On Tue, Sep 1, 2020 at 04:15 PM, Bob M. wrote:

>
> Thanks. Sorry for mis-labeling it. Too many years using an SA.
>
> Yes, the preamp has about 13dB of gain and It would be helpful to add some
> attenuation. Unfortunately I have no such attenuators with SMA connectors. I
> did want to measure the gain and eventually the return loss and impedances.
>
> I also didn't find anything that controls the output level from Port2 that
> would be feeding the input of the preamp. Again, with my SA and TG, I have
> full control of both of those levels, and that's what I'm looking for, and it
> just may not exist in the VNA world.
>

Bob,
I should mention that you can calibrate out the uncertainties of the attenuator you add to your setup: Put the attenuator on Port 0 and then run the calibration including "Thru" from that point. Now insert your preamp, leaving the attenuator on Port 0 of the NanoVNA. The gain will be indicated directly, and so will the preamp input match.
--John Gord

On Tue, Sep 1, 2020 at 04:25 PM, John Gord wrote:

>
> Bob,
> Fancy VNAs ($$$) usually have variable attenuators built in. The NanoVNA has
> some very limited control of source level through the output current setting
> of the synthesizer chip. I am not sure the NanoVNA-V2 has this. You are
> probably going to need some external attenuators.
> --John Gord
>
> On Tue, Sep 1, 2020 at 04:15 PM, Bob M. wrote:
>
> >
> > Thanks. Sorry for mis-labeling it. Too many years using an SA.
> >
> > Yes, the preamp has about 13dB of gain and It would be helpful to add some
> > attenuation. Unfortunately I have no such attenuators with SMA connectors. I
> > did want to measure the gain and eventually the return loss and impedances.
> >
> > I also didn't find anything that controls the output level from Port2 that
> > would be feeding the input of the preamp. Again, with my SA and TG, I have
> > full control of both of those levels, and that's what I'm looking for, and
> it
> > just may not exist in the VNA world.
> >
>

Bob,

Your mention of Port 2 and controlling its output level causes me to believe that you may be confused on how to connect your VNA to test your preamp. CH0 is where the RF test signal is emitted from and CH1 is an input only. So your hook up will be from CH0 through an attenuator, to the input of your preamp, then to CH1. If I recall correctly, the output from CH0 is around -10 dBm, pretty high for a preamp, and will likely overload it since a 13 dB gain will be asking for a 3 dBm output and the CH1 input won't be happy either.

Fixed value SMA attenuators are fairly inexpensive and a small selection of various values can be stacked to achieve the needed value.

Thanks. I'll check my connections to make sure that the preamp's output goes to CH1 and its input goes to CH0.

I wish the Port # and Channel # were both based on zero or one, i.e. CH0 = Port0, CH1 = Port1. Too much computer programming background.

The preamp only has about 13dB gain. Something must be working properly because the NanoVNA is reporting 3.something dB along the top of the screen, so if the output feeding it is indeed -10dBm, that's about the right value for 13dB gain.

I'll be looking for attenuators today. I wasn't sure if I could calibrate SOLT with an attenuator at Port1 for each step. I know I could throw it in after calibration and get the loss through it then add the preamp and do the math myself.

I have another opinion. I wish the Port # and Channel # were both based on zero or one, i.e. CH1 = Port1, CH2 = Port2.

The reason is that S11 is measured on the Port1, on port 2 the S21 is measured. In scattering matrix logic there is no Port0.

"In scattering matrix logic there is no Port0."
Yes indeed! I am actually surprised that the good VNA RF design engineers (presumably with some analog background), actually allowed this "digital" naming anomaly.
- Hector

I had mis-stated my desire to know the output level coming out of Port2 above. My preamp IS connected properly: its input is going to Port1 and its output is going to Port2.

FYI, the output level on Port1 is very close to -10dBm. It seems to drop off slightly at the higher frequencies. This was measured over the 100-900 MHz (default) range using Peak Hold on my spectrum analyzer. Nothing at all comes out of Port2, as expected.

As to the zero-based or one-based numbering scheme, if the ports are Port1/Port2, then the channels should be CH1/CH2, in my opinion, i.e. both one-based.

It was stated: "I should mention that you can calibrate out the uncertainties of the attenuator you add to your setup: Put the attenuator on Port 0 and then run the calibration including "Thru" from that point. Now insert your preamp, leaving the attenuator on Port 0 of the NanoVNA. The gain will be indicated directly, and so will the preamp input match."

Won't this have a bad effect on the "short" and "open" calibrations? I'm thinking I should put the 10dB pad on the input of Port2 so it will only have an effect on the "thru" calibration.

With my unit, the output level of Port1 is very close to -10dBm and it's quite constant. I would expect that if I connected Port1 to Port2 directly in the "thru" configuration, the LOGMAG would show -10dBm and it should be a fairly straight line. With a 10dB pad inserted in the "thru" calibration step, will this change the apparent input sensitivity such that is now reports -20dBm, or does this 10dB loss get calibrated out such that the unit would still report -10dBm? I can't test this myself because my 10dB pads haven't arrived yet.

I did determine that the preamp I'm testing has about 12-13dB gain and compression starts with an input signal of around -16dBm, so the -10dBm coming out of Port1 is overdriving the preamp and messing things up. I had previously tested the preamp with my SA/TG set to -30dBm output. Using the NanoVNA for other tests at the moment.

It also took me a while to realize that the gain and reference (and most likely a lot of other) parameters are done on a per-trace basis. If you have multiple traces being displayed, the various settings only seem to affect the most recently selected trace. I found it most helpful to turn off all but one trace, set it the way I want, turn on a second trace, turn off the first, and adjust the settings for that, and continue until they're all set the way I want.

The attenuator essentially becomes part of the test port during calibration.
If you want to lower the input power to your device under test, you need to connect it to port 1, then connect the open/short/load to the attenuator during the calibration.
If you are fine with the input power and just need to attenuate the output power to avoid overdriving the receiver, you need to connect the attenuator to port 2, attach open/short/load directly to port 1 as usual. The attenuator is only involved with thru calibration.

In your case, you will have to attach the attenuator to port 1 to stay in the linear region of the LNA. In either case, the effect of the attenuator will be completely calibrated out. If you had the attenuator inserted during thru calibration, then measured the attenuator itself, transmission will read as 0dB (_not_ dBm, the readings are always relative to the source power).

The same is true for the reflection. If you calibrated port 1 with the attenuator attached, and then measure a short, you will still measure 0dB reflection. Note however that the noise floor will be 20dB worse than without the attenuator. That is because in a reflection measurement the signal has to pass the attenuator twice (forward and backward). The calibration boosts the measured signal by 20dB so the correct value is shown, but it also boosts the noise. For this reason it may be preferable to put the attenuator on port 2 when you can.

In a proper VNA, the signal is usually attenuated before the reflection bridge, so only the source is attenuated but not the reflection. The increases the noies floor by only 10dB. Unfortunately, this is not possible with the NanoVNA.

I've got two 10dB SMA attenuators on the way; they should be here Tuesday.

In the meantime I connected my return loss bridge (RLB) to my spectrum analyzer (SA) and discovered that a high signal level feeding the preamp causes undesirable things to happen. With 0dBm feeding the RLB, the Return Loss (RL) is horrible (7-8dB) however when I reduced the drive to -30dBm, the RL increased to almost 20dB and the tuning changed drastically as well. Now the maximum gain point is very close to the best RL point. The same tuning change happens when I use the SA for measuring the passband gain of the preamp. Even an input level of -10dBm alters the gain tuning.

My issue with having attenuators at Port1 is that there will no longer be a true "open" nor a true "short" condition if those calibration plugs are placed after the attenuator. They will reduce the signal level going to the preamp and will improve the indicated RL greatly, which will not give me the true RL of the preamp itself, which apparently can not be done with the NanoVNA I have. While the NanoVNA is great for dealing with bandpass filters and other passive devices, it just doesn't seem to be the right tool for the job when testing some preamps that can't deal with the -10dBm output level of the NanoVNA I have.

hmmm what about using attenuator plus an external bridge ...

adding attenuators keeps the amp happy (and keeps his input power in an
area where its real input rl is fine) ... after the attenuator the
bridge (so you can accurate measure on amp end of the attenuator) ...
and the return port of the bridge in second port of vna ... you then
measure your s11 value as s21 value in the device ... problem solved

greetz sigi dg9bfc

Am 06.09.2020 um 11:12 schrieb Bob M.:

> I've got two 10dB SMA attenuators on the way; they should be here Tuesday.
>
> In the meantime I connected my return loss bridge (RLB) to my spectrum analyzer (SA) and discovered that a high signal level feeding the preamp causes undesirable things to happen. With 0dBm feeding the RLB, the Return Loss (RL) is horrible (7-8dB) however when I reduced the drive to -30dBm, the RL increased to almost 20dB and the tuning changed drastically as well. Now the maximum gain point is very close to the best RL point. The same tuning change happens when I use the SA for measuring the passband gain of the preamp. Even an input level of -10dBm alters the gain tuning.
>
> My issue with having attenuators at Port1 is that there will no longer be a true "open" nor a true "short" condition if those calibration plugs are placed after the attenuator. They will reduce the signal level going to the preamp and will improve the indicated RL greatly, which will not give me the true RL of the preamp itself, which apparently can not be done with the NanoVNA I have. While the NanoVNA is great for dealing with bandpass filters and other passive devices, it just doesn't seem to be the right tool for the job when testing some preamps that can't deal with the -10dBm output level of the NanoVNA I have.
>
>
>

On Sun, Sep 6, 2020 at 01:12 PM, Bob M. wrote:
> My issue with having attenuators at Port1 is that there will no longer be a
> true "open" nor a true "short" condition if those calibration plugs are placed
> after the attenuator. They will reduce the signal level going to the preamp
> and will improve the indicated RL greatly, which will not give me the true RL
> of the preamp itself, which apparently can not be done with the NanoVNA I
> have.

This is not the case. If the calibration is performed correctly (with the open/short/load standards attached after the attenuator), it will indicate the correct return loss. The VNA does of course see an improved _raw_ return loss, but calibration compensates for this.

> While the NanoVNA is great for dealing with bandpass filters and other
> passive devices, it just doesn't seem to be the right tool for the job when
> testing some preamps that can't deal with the -10dBm output level of the
> NanoVNA I have.
>

I agree that the NanoVNA is quite limited for measuring active devices. The lack of an internal attenuator (and thus no power sweep) is a big deal here. When you get into the non-linear region, the high harmonic content of the output may also become a serious issue.
Padding the attenuation down to -20dBm externally is likely fine though. At -30dBm is another matter, gain measurement is fine, but return loss will likely be too noisy to be useful for most purposes (the calibration would nominally still be correct though!)

It is worth noting that for the >140MHz frequency range, there is the possibility to decrease the output power by up to 9dB internally. It is just not implemented in the software.

If you use a attenuator in front of a RLB you should also use an amp at
the coupler port to have enough level for a good dynamic.

This is the reason why professional VNAs have often 2 SMA jumpers on the
input  port and the coupler port of the RLB.

This makes it possible to work with higher or lower levels.

You can try also to modify your MiniVNA to use a externel RLB. Then you
are free to add attenuators and amplifiers like you wish.

A sidenote. Thr NanoVNA isn't suitable for amplifier measurement. First
because of this level issue, second because of the missing ability to do
power sweeps. So no automated way to record the 1dB compression point
over the frequency.

To do measurements an amps, mixers, multipliers, downconverters, etc.
you also need the ability to have frequency translated calibration,
nothing to be done that easy, because then you need an absolute power
calibration of port 1 via level and frequency.

I also don't have such nice tools, this is why I do that with my
sweeper, a RLB connected to my Wiltron 560 display unit and a
directional coupler at the output of the DUT going to a second detector
for S21 plus the coupling port going to my spectrum analyzer. So I can
do some simple 1dB compression point measurements and also most of the
times the 2nd and 3rd harmonic increases very fast when a wideband amp
in going into compression. Because of a nice linearity over level and
frequency of the detector heads and a nice power stability over
frequency also frequency translation devices can be measured. It is much
more timeconsuming then with a modern VNA, but it is hobby, so no need
to do it faster :-)

The major difference in using a SNA or a VNA is that the SNA has
wideband detectors, not tuned receivers like a VNA. Because of very high
quality made detectors they have a level linearity of +-1dB over a wide
frequency range. If you add then a sweeper accuracy of +-1dB and some
hopefully short cables you can meaesure frequency conversion devices to
+-3dB over a wide range and non frequency conversion devices to +- 0.5dB
in a +- 10dB range with an old but well calibrated SNA setup. Also the
SNA sweeps much faster. I have a sweep rate of 100ms for a full band
10MHz to 18GHz sweep. perfect to find faulty cables and other
intermittent faults.

The NanoVNA V2 is using different signal sources for the source and the
receiver tuning. so it should be possible to have a different tuning at
Port 1 and Port 2. So theoretically also frequency translation devices
could be measured, but this needs an absolute power calibration of Port
1 and a high power stability. Nothing that can be done without a
temperature compensation or a compensated leveling loop.

To add such features to the NanoVNA would mean for my opinion to
redesign the whole source to a stable leveled and calibrated design.
Nothing that can be done with this design.

I could imagine that a extension with a pin diode leveler and a temp.
compensated wideband detector could help to reach that goal.

Also the behavior with the pulsed output isn't suitable for amplifier
mesurement, because you like to measure also the power consumptin over
power and frequency, also the thermal behavior. For this a constant
input level is required.

Regards,

Isidro

DB1SBI



Am 06.09.2020 um 13:51 schrieb Siegfried Jackstien:

> hmmm what about using attenuator plus an external bridge ...
>
> adding attenuators keeps the amp happy (and keeps his input power in
> an area where its real input rl is fine) ... after the attenuator the
> bridge (so you can accurate measure on amp end of the attenuator) ...
> and the return port of the bridge in second port of vna ... you then
> measure your s11 value as s21 value in the device ... problem solved
>
> greetz sigi dg9bfc
>
> Am 06.09.2020 um 11:12 schrieb Bob M.:
>> I've got two 10dB SMA attenuators on the way; they should be here
>> Tuesday.
>>
>> In the meantime I connected my return loss bridge (RLB) to my
>> spectrum analyzer (SA) and discovered that a high signal level
>> feeding the preamp causes undesirable things to happen. With 0dBm
>> feeding the RLB, the Return Loss (RL) is horrible (7-8dB) however
>> when I reduced the drive to -30dBm, the RL increased to almost 20dB
>> and the tuning changed drastically as well. Now the maximum gain
>> point is very close to the best RL point. The same tuning change
>> happens when I use the SA for measuring the passband gain of the
>> preamp. Even an input level of -10dBm alters the gain tuning.
>>
>> My issue with having attenuators at Port1 is that there will no
>> longer be a true "open" nor a true "short" condition if those
>> calibration plugs are placed after the attenuator. They will reduce
>> the signal level going to the preamp and will improve the indicated
>> RL greatly, which will not give me the true RL of the preamp itself,
>> which apparently can not be done with the NanoVNA I have. While the
>> NanoVNA is great for dealing with bandpass filters and other passive
>> devices, it just doesn't seem to be the right tool for the job when
>> testing some preamps that can't deal with the -10dBm output level of
>> the NanoVNA I have.
>>
>>
>>
>
>
>

for amp testing you also can use satsagen and a pluto (see picture)

a 60db attenuator added after poweramp not to cook the rx port with 20w
rf power :-)

dg9bfc sigi


Am 06.09.2020 um 12:28 schrieb Isidro Berniol:

> If you use a attenuator in front of a RLB you should also use an amp
> at the coupler port to have enough level for a good dynamic.
>
> This is the reason why professional VNAs have often 2 SMA jumpers on
> the input  port and the coupler port of the RLB.
>
> This makes it possible to work with higher or lower levels.
>
> You can try also to modify your MiniVNA to use a externel RLB. Then
> you are free to add attenuators and amplifiers like you wish.
>
> A sidenote. Thr NanoVNA isn't suitable for amplifier measurement.
> First because of this level issue, second because of the missing
> ability to do power sweeps. So no automated way to record the 1dB
> compression point over the frequency.
>
> To do measurements an amps, mixers, multipliers, downconverters, etc.
> you also need the ability to have frequency translated calibration,
> nothing to be done that easy, because then you need an absolute power
> calibration of port 1 via level and frequency.
>
> I also don't have such nice tools, this is why I do that with my
> sweeper, a RLB connected to my Wiltron 560 display unit and a
> directional coupler at the output of the DUT going to a second
> detector for S21 plus the coupling port going to my spectrum analyzer.
> So I can do some simple 1dB compression point measurements and also
> most of the times the 2nd and 3rd harmonic increases very fast when a
> wideband amp in going into compression. Because of a nice linearity
> over level and frequency of the detector heads and a nice power
> stability over frequency also frequency translation devices can be
> measured. It is much more timeconsuming then with a modern VNA, but it
> is hobby, so no need to do it faster :-)
>
> The major difference in using a SNA or a VNA is that the SNA has
> wideband detectors, not tuned receivers like a VNA. Because of very
> high quality made detectors they have a level linearity of +-1dB over
> a wide frequency range. If you add then a sweeper accuracy of +-1dB
> and some hopefully short cables you can meaesure frequency conversion
> devices to +-3dB over a wide range and non frequency conversion
> devices to +- 0.5dB in a +- 10dB range with an old but well calibrated
> SNA setup. Also the SNA sweeps much faster. I have a sweep rate of
> 100ms for a full band 10MHz to 18GHz sweep. perfect to find faulty
> cables and other intermittent faults.
>
> The NanoVNA V2 is using different signal sources for the source and
> the receiver tuning. so it should be possible to have a different
> tuning at Port 1 and Port 2. So theoretically also frequency
> translation devices could be measured, but this needs an absolute
> power calibration of Port 1 and a high power stability. Nothing that
> can be done without a temperature compensation or a compensated
> leveling loop.
>
> To add such features to the NanoVNA would mean for my opinion to
> redesign the whole source to a stable leveled and calibrated design.
> Nothing that can be done with this design.
>
> I could imagine that a extension with a pin diode leveler and a temp.
> compensated wideband detector could help to reach that goal.
>
> Also the behavior with the pulsed output isn't suitable for amplifier
> mesurement, because you like to measure also the power consumptin over
> power and frequency, also the thermal behavior. For this a constant
> input level is required.
>
> Regards,
>
> Isidro
>
> DB1SBI
>
>
>
> Am 06.09.2020 um 13:51 schrieb Siegfried Jackstien:
>> hmmm what about using attenuator plus an external bridge ...
>>
>> adding attenuators keeps the amp happy (and keeps his input power in
>> an area where its real input rl is fine) ... after the attenuator the
>> bridge (so you can accurate measure on amp end of the attenuator) ...
>> and the return port of the bridge in second port of vna ... you then
>> measure your s11 value as s21 value in the device ... problem solved
>>
>> greetz sigi dg9bfc
>>
>> Am 06.09.2020 um 11:12 schrieb Bob M.:
>>> I've got two 10dB SMA attenuators on the way; they should be here
>>> Tuesday.
>>>
>>> In the meantime I connected my return loss bridge (RLB) to my
>>> spectrum analyzer (SA) and discovered that a high signal level
>>> feeding the preamp causes undesirable things to happen. With 0dBm
>>> feeding the RLB, the Return Loss (RL) is horrible (7-8dB) however
>>> when I reduced the drive to -30dBm, the RL increased to almost 20dB
>>> and the tuning changed drastically as well. Now the maximum gain
>>> point is very close to the best RL point. The same tuning change
>>> happens when I use the SA for measuring the passband gain of the
>>> preamp. Even an input level of -10dBm alters the gain tuning.
>>>
>>> My issue with having attenuators at Port1 is that there will no
>>> longer be a true "open" nor a true "short" condition if those
>>> calibration plugs are placed after the attenuator. They will reduce
>>> the signal level going to the preamp and will improve the indicated
>>> RL greatly, which will not give me the true RL of the preamp itself,
>>> which apparently can not be done with the NanoVNA I have. While the
>>> NanoVNA is great for dealing with bandpass filters and other passive
>>> devices, it just doesn't seem to be the right tool for the job when
>>> testing some preamps that can't deal with the -10dBm output level of
>>> the NanoVNA I have.
>>>
>>>
>>>
>>
>>
>>
>
>
>

Some examples using a tinySA

Amp : https://tinysa.org/wiki/pmwiki.php?n=Main.OnedBc
Mixer: https://tinysa.org/wiki/pmwiki.php?n=Main.Mixer

Well, it's Tuesday. Thanks to the ebay seller who never shipped my attenuators, I have to buy them from someone else. So I have nothing new to report as to their effect when placed in series with Port1. Grrr.

My 10dB attenuators finally arrived. Today I connected them in series to Port1, then used the supplied cables and calibration components to calibrate the unit after a reset, so it covered 100-900 MHz with 101 points. The calibration seemed to be successful although I didn't actually look at any of the traces.

I then connected a UHF preamp to the NanoVNA. It reported 13.25dB gain, which is at least in the ballpark of reality, as the preamp has 12-14dB gain as measured with my much more expensive spectrum analyzer.

Because of the 20dB of attenuation, I was unable to measure the input SWR of the preamp. The NanoVNA reported a VSWR of 1.03 or 36.6dB return loss. I think most of this is due to the attenuators. The actual preamp has a much poorer return loss, as measured with my separate return loss bridge using a -30dBm source level.

My conclusion is that the pads are useful for measuring the gain of a preamp but useless when trying to measure the input return loss (SWR) due to the -10dBm output level of the NanoVNA causing conduction in the preamp's transistor. For passive networks or antennas none of this would ever be an issue.

I think something is likely wrong with the calibration if you get 36.6dB return loss. What return loss do you see when you leave the cable open (after calibration, with attenuator + cable connected to port 1)?

With 2x10dB at port 1 the reflection measurement may be well too noisy to be usable (1x10dB is probably fine), but it definitely should not read high.