SDR, RTL, NWT4000, HackRF... What?

[masterx81]

Hi! I'm searching a decent (or at least the best bung/buck) spectrum analyzer in the cheapest range... I don't need to project antennas or similar (in that case nanoVNA will be the answer). I need only sometimes to check if a trasmitter is working and check it's frequencies (in the 2.4ghz range), and maybe analyze the DVB-T2 frequencies/power levels .
I've seen a lot of projects for this scope under the 200usd, but i can't choose what's better.
The sdr-rtl seem to work only up to 1.7ghz, and from what i've read are the best thing for work in the dvb-t range (and are quite cheap).
Remain uncovered the 2.4ghz...
I've seen the nwt variants, and other projects (like hackrf that have only 8bit resolution), i've read a lot but can't decide...
Someone have some suggestions?
Thanks!

[radiolistener]

I bought this one: https://www.aliexpress.com/item/33032121416.html

At a glance it has some little input impedance issue, needs to do a little modding.
Probably remove 50 Ohm resistor between RF IN connector and mixer should fix it.

But such mod will require about 10 dB attenuator between RF OUT and RF IN. Because RF OUT has about -5...0 dBm. This is too much for IAM-81008 mixer.

I didn't tried yet such modding.

According to my measurements, actual dynamic range is about 50 dB up to 1.5 GHz and about 30-40 dB up to 4.4 GHz.

You can see picture which shows result taken for RF-IN=RF-OUT and RF-IN/RF-OUT loaded with 50 Ohm dummy loads.
The first picture is math correction after calibration with 10 and 40 dB attenuators.
And the second picture is RAW ADC data taken for the same measurement.

As you can see from RAW ADC data, there is some RF leak at about 2.7 - 3.3 GHz range. At a glance the root of cause of this RF leak is ground plane layout. Because it is changing when you touch ground on SMA connectors.

When it is used in spectrum analyzer mode, it's noise floor is about -47 dBm. So, it needs LNA amplifier in order to see low power signals.

The third picture is a spectrum of GSM phone during call received on RHD-771 antenna from 30 cm distance from the phone.

And the fourth picture is a spectrum of WiFi router taken with the same RHD-771 antenna placed between router antennas at several centimeters distance.

[masterx81]

Whoa, so high noise floor... This seem a clone of the nwt4000, from what i've seen the "genuine" nwt4000 will perform bit better (and costs 4x/5x more).
And also, seem that the genuine nwt4000 has no aluminium case, while some other nwt4000 (clones of the clones of the...) does have it.
Thanks!

[radiolistener]

I bought it primary as a signal source for 33-4400 MHz range. But it also works good as a simple spectrum analyzer and bode plot.

For example, here is my hand made BPF tuned for 145.3 MHz measured with this LTDZ device:

[MikeLud]

radiolistener

I also just bought one, which software are you using. Occasionally when running a 0 db sweep I get the below spikes. I was also looking to do the filter mod and found the newer versions of the LTDZ has it already

[RoGeorge]

I have a PlutoSDR, designed by Analog Devices as a learning/demo for their 6GHz transceiver (on the pge sais 3GHz but can be unlocked to 6GHz for both Rx and Tx if you don't care much about accuracy).
https://www.analog.com/en/design-center/evaluation-hardware-and-software/evaluation-boards-kits/adalm-pluto.html
Nice toy w a 6GHZ ADI transceiver, 12 bits ADC, a dual core Zynq FPGA, open source, runs Linux, has support for Matlab, GNU Radio, etc.

There are some applications like SATSAGEN that can make a PlutoSDR act as a very basic spectrum analyzer or generator upto 6GHz:
https://www.eevblog.com/forum/rf-microwave/adalm-pluto-as-simple-spectrum-analyzerw-tracking/
http://www.albfer.com/en/2020/10/11/satsagen-0-4/

If you want it for learning SDR, too, there is a free book from ADI and some uni classes, too:
https://www.analog.com/en/education/education-library/software-defined-radio-for-engineers.html
https://pysdr.org/

[radiolistener]

I also just bought one, which software are you using. Occasionally when running a 0 db sweep I get the below spikes. I was also looking to do the filter mod and found the newer versions of the LTDZ has it already

I'm using my own software, it has better calibration, but limited functions

Note that 0 dB sweep leads to overload LTDZ input, because it has a little higher output than needed for mixer saturation.
So, I recommend to use it through 10 dB attenuator.

Recommended attenuators for calibration are 10 dB and 40 dB.

I attached current version of my software for LTDZ.

If possible please run it on your LTDZ device and give me screenshot with ADC output for the following cases:
- RF-IN connected to RF-OUT
- RF-IN connected to RF-OUT through attenuator 10 dB
- RF-IN loaded with 50 Ohm dummy load and RF-OUT loaded with 50 Ohm dummy load

If you don't have dummy load, just leave both connectors open for third test.
Use the following sweep parameters:
- start freq: 35000000
- stop freq: 4400000000
- points: 9999

I'm interesting if there is a difference in hardware performance for different LTDZ instance.

===========
Some instructions:

1) unzip and run
2) Select proper COM port at the top of window and click "Connect" button
3) Select "ADC" for Mode\Data combo box
4) Press Sweep button

for calibration:
1) press button on LTDZ to enable tracking generator
2) press "Calibrate" button
3) connect RF-IN to RF-OUT through 10 dB attenuator and press OK
4) connect RF-IN to RF-OUT through 40 dB attenuator and press OK
5) press Yes to save calibration (or no to not save, in such case calibration will remain in memory, but not be overwritten on the disk)
6) select Mode\Data = Magnitude and press Sweep to check the calibration

Notes:
- 9999 points (maximum for LTDZ) is recommended for calibration
- LTDZ has min frequency 33 MHz and max frequency 4400 MHz
- you can use your own values for attenuators in calibration procedure, just enter it's value below Calibrate button
- offset field can be used to add some offset in dB on the chart
- calibration file is saved to the file _refLogA.xml, if such file exists it will be loaded at startup. Otherwise it uses default values.
- calibration file consists of two datasets with raw ADC data in hex format with comma delimiter, taken for two attenuators.
- md5: CFF3BF1C63067D02F4EFB5553B11F39D

[masterx81]

I have a PlutoSDR, designed by Analog Devices as a learning/demo for their 6GHz transceiver (on the pge sais 3GHz but can be unlocked to 6GHz for both Rx and Tx if you don't care much about accuracy).
https://www.analog.com/en/design-center/evaluation-hardware-and-software/evaluation-boards-kits/adalm-pluto.html
Nice toy w a 6GHZ ADI transceiver, 12 bits ADC, a dual core Zynq FPGA, open source, runs Linux, has support for Matlab, GNU Radio, etc.

There are some applications like SATSAGEN that can make a PlutoSDR act as a very basic spectrum analyzer or generator upto 6GHz:
https://www.eevblog.com/forum/rf-microwave/adalm-pluto-as-simple-spectrum-analyzerw-tracking/
http://www.albfer.com/en/2020/10/11/satsagen-0-4/

If you want it for learning SDR, too, there is a free book from ADI and some uni classes, too:
https://www.analog.com/en/education/education-library/software-defined-radio-for-engineers.html
https://pysdr.org/

Interesting device. The other sdr devices that i've seen arrive at only 2ghz (airspy 1.7, sdrplay 2.0, seem that most of them are derived from dvb-t systems).
But i think that the use of plutosdr is more indicated for a more rf instructed person than me 😅😅

[MikeLud]

radiolistener

Thanks for the software. I ran the tests, the descriptions for each test is in the file name. I ran all of the tests uncalibrated.

To see what is happening when I get those spikes I scoped the ADC and RF_OUT and what I found out is when the spike occurs  there is no signal coming out of the RF_OUT. I am thinking the Firmware I have V1.19 has a bug with the SPI timing. What version firmware does your LTDZ have.

[radiolistener]

Thanks for the software. I ran the tests, the descriptions for each test is in the file name. I ran all of the tests uncalibrated.

Thanks. Calibration doesn't affects ADC chart results, because ADC chart shows RAW ADC data taken from LTDZ.

There is definitely some difference between our devices. What I found:

- through (direct IN=OUT) on our LTDZ produce the same ADC value 460

- isolation (IN=load and OUT=load) produce a little different level. On my LTDZ ADC value is 89 vs about 65 on your. May be this is due to different dummy load?

- my LTDZ don't have any drops on 108 MHz. I can run a lot of sweeps and it's pretty stable. See screenshot from 35 to 200 MHz.

- your LTDZ has definitely higher RF leak between IN and OUT above 760 MHz (isolation screenshot). My one has RF leak above 2 GHz.

- your LTDZ has more flat response for through connection. My one has significant drops above 2.63 MHz

I'm not sure what is the root of cause for these difference.
May be this is due to different component quality or different firmware.

By the way, did you clean your PCB from flux? I clean it with ethanol

Do you have installed 50 Ohm resistor on the ground at RF in connector (after capacitor)?

To see what is happening when I get those spikes I scoped the ADC and RF_OUT and what I found out is when the spike occurs  there is no signal coming out of the RF_OUT. I am thinking the Firmware I have V1.19 has a bug with the SPI timing.

I don't have such drops. What is the frequency? 108 MHz?

What version firmware does your LTDZ have.

I don't know. My LTDZ reports firmware version FW:1.19:V10 (Version: 77, Variant: 0a).
But it doesn't means that our firmware is really identical. Needs more deep check.

I will try to read frimware from my LTDZ and will report the results later.

What I found during debugging my software is that my LTDZ firmware has a bug for a sweep with no delay. If I send command for a sweep with delay 0 it just hungs and needs to power off to reset. This is why I'm using sweep command with delay. It is used in SNASharp and WinNWT.

So, there is possible some difference and bugs in firmware. Needs to learn it more deep.

Also it's not clear why it use just 1/8 of ADC scale.
STM32F103C8T6 has 12 bit ADC, so the resolution can be 8 times better by using full ADC scale.
So, there is a perspective for improvement.

[radiolistener]

also measurement with different attenuators:
- isolation = both connectors are open
- through = IN connected to OUT through 32 cm RG174
- 10 dB, 20 dB, 30 dB, 40 dB, 50 dB - IN connected to OUT through 32 cm RG174 and selected attenuator.

I'm not sure if my attenuators working good above 2 GHz. I'm use popular Chinese SMA attenuators for 6 GHz.

[radiolistener]

LTDZ firmware is not protected. I can read it.

Does anyone know if LTDZ firmware source is available on the internet?

[DH1AKF_Wolfgang]

Also owner of a LTDZ analyzer, I now made the modification and soldered out the 50 Ohms resistor at the receiver port. Thanks for the hint.
Now the result, measured with my NanoVNA V2:

[radiolistener]

Also owner of a LTDZ analyzer, I now made the modification and soldered out the 50 Ohms resistor at the receiver port. Thanks for the hint.
Now the result, measured with my NanoVNA V2:

Do you know if LTDZ firmware source is available on the internet?

I can read firmware from my LTDZ bought on aliexpress, you can see it in attachment. But I want source to make mods...

I've found this github project for D6: https://github.com/joseluu/D6_firmware
It looks that D6 firmware works with LTDZ, but it uses 12 bit ADC format and different NWT command style, so it is incompatible with existing LTDZ software, but it can work with WinNWT.

I'm wondering where I can download source code for original LTDZ firmware?

[DH1AKF_Wolfgang]

Hi Radiolistener,
only this link collection I could find:
http://www.kh-gps.de/analyzer.htm

[MikeLud]

By the way, did you clean your PCB from flux? I clean it with ethanol

I cleaned it with 90% alcohol

Do you have installed 50 Ohm resistor on the ground at RF in connector (after capacitor)?

No, I remove it

I don't know. My LTDZ reports firmware version FW:1.19:V10 (Version: 77, Variant: 0a).

My firmware reads the same

But it doesn't means that our firmware is really identical. Needs more deep check.

This is true, I did a memory dump of my firmware and compared with your dump and some parts were the same. So I took a chance and flashed your firmware. After flashing I tried your software with the newly flashed LTDZ and found it does not work. I then reflashed with the dump I did and now the board does not work at all with any software.

I've found this github project for D6: https://github.com/joseluu/D6_firmware
It looks that D6 firmware works with LTDZ, but it uses 12 bit ADC format and different NWT command style, so it is incompatible with existing LTDZ software, but it can work with WinNWT.

After flashing your dump and my dump I figured what do I have to loss so I flashed Joseluu D6 firmware and now my board only works with SNASharp http://alloza.eu/david/WordPress3/?p=1014  also no more spikes. I tried several versions of WinNWT none of them worked worked with Joseluu D6 firmware

[radiolistener]

This is true, I did a memory dump of my firmware and compared with your dump and some parts were the same. So I took a chance and flashed your firmware. After flashing I tried your software with the newly flashed LTDZ and found it does not work. I then reflashed with the dump I did and now the board does not work at all with any software.

This is strange, the only reason why it cannot work is a firmware lock to STM32 serial number or something like that. I don't think that someone will do that. So I think most of all there is some mistake in reflashing procedure.

I reflashed attached firmware after testing other firmware and it works again, the same as before.

Did you used STM32 ST-LINK Utility for flashing?

Can you give me your LTDZ STM32 dump?

[MikeLud]

Did you used STM32 ST-LINK Utility for flashing?
Can you give me your LTDZ STM32 dump?

Yes, I used STM32 ST-LINK Utility for flashing

Its 3:00am here in the US, I am going to get some sleep.

[radiolistener]

Yes, I used STM32 ST-LINK Utility for flashing

Its 3:00am here in the US, I am going to get some sleep.

This is not a full dump of STM32F103C8T6. It consists just first 4 kB.

At a glance you're set wrong chip memory size in ST-LINK Utility. From the text file it reports checksum for
[0x08000000:0x08001000] - this is 4 kB 0x1000

but STM32F103C8T6 has 64 kB 0x10000 flash and some version may have even 128 kB 0x20000.

So, this is the reason why it didn't works after reflashing. This is not a full memory dump.

I compared it's content and it has different code from my one.
It's a pity that you didn't make a complete dump of your version.

Try to setup 0x10000 length and save memory dump, may be area above 0x1000 still remains your original firmware. If you didn't reflashed it with other firmware.

After that you can reflash firmware that I posted above, it should work for you.
But this is not the same firmware as you used out of the box.
Just don't forgot to setup memory size 0x10000 (4 zero! not 3!)

[MikeLud]

Radiolistener

The original firmware is toast. I ordered a replacement from amazon and should have it by Thursday. I retested your firmware with your software and the RF In works somewhat (ADC numbers looks off) see attached 100Mhz -10 dBm sweep. The TG does not work at all (I did turn it on with the button)

[Bicurico]

May I recommend my blog: https://vma-satellite.blogspot.com/

It basically discusses the original question of masterx81 and includes my own software for this range of devices, as well as some comparisons between models, etc.

Bottom line is:

NWT4000 is better than D6 is better than LTDZ.

There are mods published for D6 to improve performance, I don't know of any mods for the LTDZ. It has basically a too high noise floor.

I have looked into the firmware and I don't see any dramatic improvements being possible just by changing the FW (except what my software already does anyway).

Regards,
Vitor

[HB9EVI]

NWT4000 is better than D6 is better than LTDZ.

+1

the LTDZ is quite a shabby device; neither with LinNWT nor with SNASharp i got it properly working on my linux workstation; but what else can be expected for the price

[Bicurico]

It is quite poor a device, however it can still be useful for some applications.

You get what you pay for!

I think that it is a good device to start learning about spectrum analysers. If you break it, the loss is not so great.

[radiolistener]

NWT4000 is better than D6 is better than LTDZ.

There are mods published for D6 to improve performance, I don't know of any mods for the LTDZ. It has basically a too high noise floor.

I have looked into the firmware and I don't see any dramatic improvements being possible just by changing the FW (except what my software already does anyway).

Where I can find the source code of LTDZ firmware?

[Bicurico]

Sorry - I meant the D6 firmware!

The whole SDk can be downloaded here: https://www.mediafire.com/file/j0kx8hndiieweak/Simple_Spectrum_Analyser.rar/file

Note that I am not hosting this file nor did I put it together. I just received this link when I bought my D6. I have nothing todo with its contents! Use at your own risk.
 
Regards,
Vitor