Keyboard Shortcuts
ctrl + shift + ? :
Show all keyboard shortcuts
ctrl + g :
Navigate to a group
ctrl + shift + f :
Find
ctrl + / :
Quick actions
esc to dismiss
Likes
Search
detection bandwidth in NanoVNA
|
Looking at the code in dsp.c, nanovna.h, and nanovna.c,
It looks like the two synthesizers are set for an offset frequency of 5kHz - that is, the LO to the mixers is 5kHz higher than the frequency coming out the CH0 port. The stock edy555 version samples for 48 samples at 48 ksps, so it grabs 1 millisecond. That's multiplied by sin and cos at 5kHz (in a table in nanovna.h) integrated in dsp_process() to produce the values that get processed into the measurement. So, to a first order, the detection bandwidth is 1kHz (= 1/T). The effective band pass filter is a sin(x)/x kind of shape (I didn't see any windowing or weighting). so interfering signals (i.e. broadcast band interference), if within, say, 20 kHz, of the test frequency might have an effect. The "selectivity" (in radio receiver terms) is set by the audio low pass filter formed by and the input Z of the ADCS (10k?, set in tlv320aic3204.c). I didn't look long enough to figure out if edy555 programmed the chip to have a low pass filter (it has some ability to do FIR filtering) |
|
The front end of that 'receiver' is WIDE OPEN!
toggle quoted message
Show quoted text
RF rejection of that signal 20 kHz, 20 MHz or even 200 MHz away is ZERO! Kent WA5VJB On Thursday, July 30, 2020, 9:21:24 AM CDT, Jim Lux <jimlux@...> wrote:
Looking at the code in dsp.c, nanovna.h, and nanovna.c, It looks like the two synthesizers are set for an offset frequency of 5kHz - that is, the LO to the mixers is 5kHz higher than the frequency coming out the CH0 port. The stock edy555 version samples for 48 samples at 48 ksps, so it grabs 1 millisecond.? That's multiplied by sin and cos at 5kHz (in a table in nanovna.h) integrated in dsp_process() to produce the values that get processed into the measurement. So, to a first order, the detection bandwidth is 1kHz (= 1/T).? The effective band pass filter is a sin(x)/x kind of shape (I didn't see any windowing or weighting). so interfering signals (i.e. broadcast band interference), if within, say, 20 kHz, of the test frequency might have an effect. The "selectivity" (in radio receiver terms) is set by the audio low pass filter formed by? and the input Z of the ADCS (10k?, set in tlv320aic3204.c).? I didn't look long enough to figure out if edy555 programmed the chip to have a low pass filter (it has some ability to do FIR filtering) |
|
On Thu, Jul 30, 2020 at 07:21 AM, Jim Lux wrote:
No FIR filtering in tlv320aic PS In my version firmware i use 96k ADC (in last beta 192kHz) and 12k IF, so for 48 samples detection bandwidth is 2kHz (4kHz on 192k ADC), and can be chaned by n*samples measure. |
|
On 7/30/20 7:28 AM, KENT BRITAIN wrote:
The front end of that 'receiver' is WIDE OPEN! Actually, no, any more than your standard HF receiver is "wide open" - most good receivers have limited filtering before the first mixer (because it's loss and would degrade the noise figure). The NanoVNA is a standard superhet receiver - with a 5kHz IF. The mixer LO is 5kHz above the desired frequency, there's some analog filtering after the mixer (RC), and then a digital filter with 1 kHz BW in the dsp code. No, it's not going to make Rob Sherwood's top 10 list as a receiver, but it's not wide open, like a lot of the earlier VNAs (TAPR VNA from TenTec) that used the Analog Devices detector chips, or the MFJ-259 type devices. The selectivity isn't awesome, but if you're tuned to 4 MHz, it will probably reject 1MHz pretty well, as long as you don't overload the mixer (like any receiver). Obviously, it has no image rejection, so if you're tuned to 4 MHz, and there's interference at 4.010 MHz, it will screw up the measurement. I don't have a signal generator here to test it, but it would be an interesting measurement to try. Some of the later firmware versions apparently jack up the sample rate. If they do that, and they implement either a different offset frequency and/or a filter on the ADC samples and/or capture more samples than the current 48, then the selectivity would improve. On Thursday, July 30, 2020, 9:21:24 AM CDT, Jim Lux <jimlux@...> wrote: |
|
On 7/30/20 7:38 AM, DiSlord wrote:
On Thu, Jul 30, 2020 at 07:21 AM, Jim Lux wrote:No FIR filtering in tlv320aic Yeah, I figured that's the case. I'm not sure it makes a huge difference overall - it would drive down the noise floor somewhat to use more samples, and getting farther away from DC is usually good. It kind of gets down to "what do you need" - I see the nanoVNA as a 40-50dB sort of device, not a 80-100dB device. |
|
I fully understand the layout of the Nano's.I currently own 5 commercial Network analyzers andmy vintage 141T does have a tracking filter when usedas a Scalar Analyzer.
toggle quoted message
Show quoted text
My point is that when you connect one to a BIG outside antenna, AM Stations, FM Stations, TV Stations, etc can all easily introduceenough RF Voltage to drive it non-linear. I see the same issue with SDR users.?? They think they can eliminatesignal overload by just telling their laptop to ignore that frequency. No, when the first simiconductor device goes non-linear, your screwed! So, if you own a 40 GHz HP VNA, why are you messing with a Nano?They are cute and much easier to carry!?? In fact, I own 3.?? Kent On Thursday, July 30, 2020, 9:46:50 AM CDT, Jim Lux <jimlux@...> wrote:
On 7/30/20 7:28 AM, KENT BRITAIN wrote: ? The front end of that 'receiver' is WIDE OPEN! Actually, no, any more than your standard HF receiver is "wide open" - most good receivers have limited filtering before the first mixer (because it's loss and would degrade the noise figure). The NanoVNA is a standard superhet receiver - with a 5kHz IF. The mixer LO is 5kHz above the desired frequency, there's some analog filtering after the mixer (RC), and then a digital filter with 1 kHz BW in the dsp code. No, it's not going to make Rob Sherwood's top 10 list as a receiver, but it's not wide open, like a lot of the earlier VNAs (TAPR VNA from TenTec) that used the Analog Devices detector chips, or the MFJ-259 type devices. The selectivity isn't awesome, but if you're tuned to 4 MHz, it will probably reject 1MHz pretty well, as long as you don't overload the mixer (like any receiver).? Obviously, it has no image rejection, so if you're tuned to 4 MHz, and there's interference at 4.010 MHz, it will screw up the measurement. I don't have a signal generator here to test it, but it would be an interesting measurement to try. Some of the later firmware versions apparently jack up the sample rate. If they do that, and they implement either a different offset frequency and/or a filter on the ADC samples and/or capture more samples than the current 48, then the selectivity would improve.
|
|
On Thu, Jul 30, 2020 at 07:49 AM, Jim Lux wrote:
No in range 1-100Mz nano vna allow get -100dB Here is attenuator measure in this range on last firmware 
1-100 10dB.png
1-100 50dB.png
1-100 80dB.png
1-100 90db.png
1-100 100dB.png
1-100 110dB.png
Noise floor.png
|
|
On 7/30/20 7:54 AM, KENT BRITAIN wrote:
I fully understand the layout of the Nano's.I currently own 5 commercial Network analyzers andmy vintage 141T does have a tracking filter when usedas a Scalar Analyzer.I'm not sure that you'd actually see enough power from BC transmitters to saturate the front end - I guess it depends on where you are relative to the transmitter. The MFJ259 are notorious for problems, but that's because they have no selectivity, and the diodes are sensitive to being burned out. The NanoVNA has a fair amount of attenuation between the two inputs and the mixers. I think you'd need to see over 0dBm at the input before it's a problem. I don't know that I'd try to measure the Z of an antenna at Field Day with multiple transmitters going, for instance. I'd be interested to see some actual measured data. Yes, the SDR folks do get hosed by this occasionally, as do people using lab spectrum analyzers and take out all the attenuation to see that weak signal. |
|
On 7/30/20 8:23 AM, DiSlord wrote:
On Thu, Jul 30, 2020 at 07:49 AM, Jim Lux wrote:No in range 1-100Mz nano vna allow get -100dB Well, that's certainly the case that you can *measure* 100dB loss, but I think that when you're talking that kind of attenuation, there's other factors that will limit the accuracy - for instance, is your cal kit well known enough. Is the instrument stable enough? Is the connector mate/demate uncertainty low enough. What you get with the fancy VNA in a lab is stability, repeatability, and accuracy, as well as immunity to interfering signals (as well as a lot of handy performance features like being able to run power sweeps, etc.) Not that you can't do those things with a NanoVNA (and software, and other hardware), but in the lab box, it's all right there. The NanoVNA is very impressive, but it's a "horses for courses" kind of thing. And what I find really cool is that folks are pushing the limits and coming up with new features, and new post processors. I've been thinking it would be nice to have one that doesn't have the reflection bridge, and just brings out the stimulus, and has the 3 receiver ports. Then you could have an "add on" unit that would, say, do frequency translation up to microwave frequencies (like 10 or 24 GHz), while staying well within the "comfortable" range of the basic box, without needing to do the harmonic thing. |
|
Jim,
Have a look in the file section under Hardware Mods... /g/nanovna-users/files/Hardware%20Mods/Senzu%20-%20nanovna%20mod.pdf This person is on twitter, do a search for 'nanovna' and see what comes up. ... Larry |
|
On 7/30/20 10:08 AM, Larry Rothman wrote:
Jim,Yes, that's the sort of idea. Time to buy another nanoVNA to experiment with. |
to navigate to use esc to dismiss