ttl / vna2 Goto Github PK

Hi, thanks for sharing!

I have read your articles and hopefully one day I'll be able to replicate your work.

Wouldn't it be possible to implement shielding cheaper than custom milling / bending with the help of 3D printed covers coated with EMI-shielding copper-based spray paint? There are several products specifically for this use and I have seem them claim about 60 dB attenuation for 50 µm coating.

Thanks for sharing! From the PCB, I have got the impedances in bridge parts are different and not equal to 50ohm. Could you please explain the reasons for me? Thank you!

First of all, amazing work! I would like to try my hand at cloning this project, and seeing what I can do to increase the frequency range and/or extend to 4 ports. I am in electronics reliability, and often need cheap relatively-accurate RF bench test and in situ (environmental stress) test equipment in the L, S, X, and K bands. Due to business accounting, it's often easier to build engineering boards rather than slug through the arduous capital acquisition steps for equipment that is far more than what I need. My most recent activity requires a stable differential S-par characterization pre/post stress up to 26.5GHz. Obviously I'm not going to get that all in one shot/design/project, and will require an iterative and more complex (and higher part count) solution. This would try to rival the basic capabilities of something like the 5242 4-port PNAs from Keysight/Agilent. If this could be done at 5-10% of the cost, that would be a phenomenal achievement. I would then probably quit my job and make PNA's for a living :-) I think a more reasonable goal would be to target something functional through X band. Any block diagram / recommendations on where to focus for this project would be greatly appreciated. One idea I was thinking about was replacing the baluns with diplexers or filters. This would be much more lossy (particularly in the cross over region) and noisy, but might start me on the path of combining dedicated blocks to handle lower and higher frequency bands without having to source extremely broad band components. As a side note, I'm also aware that calibrating such a beast will be a wholly new project (perhaps a second switched calibration board). I do fully admit that the convenience of Agilent Ecal may be giving me a false impression of how achievable this endevor is.

Also, I had some questions about possible ways to improve performance / noise on your mark-2 VNA design. It seems to me that there is a lot of unused space on the bottom layer. Would it be more efficient to move components that generate noise interference such as receivers/switches to the backside? My thinking is that the internal ground planes should provide some free receiver shielding from the source components and FPGA. Also, this might shorten some of your lines, or place via rows in between some signal lines that are the worst cross-talk offenders (if known). Clearly this would make soldering more of a challenge, but the benefits might outweigh the assy challenges. I was also wondering if you couldn't extend the edge of the board and place the couplers on it? You could pull back the copper grounds for better isolation, and it would cut down on the connectors. The ferrite bead sub assembly is hand solder regardless, and it's one less setup / board fab cost.

Thank you for putting so much effort into making this concept such a successful reality. I'm new to GitHub, and just thrilled that you are sharing the fruits of your labor to perpetuate greater access to advanced analysis tools.

Hi Henrik,
1.
I sent your geber data to OSH to make the PCB, but later I found out that there is your finished product on OSH（https://oshpark.com/profiles/Chynthlis/page/1）. I don't know what is the difference between the PCB I am doing and yours?
2.
Thank you very much

Hi Henrik,

Great job on the VNA design! I am trying to correlate between an FMCW radar and a VNA. In the FMCW radar, the ADC digitizes the IF signal into I and Q samples.

In your VNA design, the ADC samples the IF signal into I and Q samples. How did you convert the low frequency IF data into RF S parameters? I am reading your sparam.py file and in iq_to_sparam() there is a formula and algorithm which I am not able to grasp.

Thanks for your help on this.

Hi Henrik,
I'm little confused about the difference between coupler BOM on Github and photos on your blog.

What I see on the photo is 3pcs of 5943000911, 2pcs of 5977000101 and 1pcs of B64290P0687X046.
On the BOM is 2pcs of 5943000911 and 3pcs of 5977000101.
Did You make a change in assembly without posting it on your blog or is it mistake in the BOM?
Thanks, Michal

Dear friend, do you have the BOM for VNA2, could you please upload here, thanks

Hi Henrik,
I am impressed by your VNA design, and want to made one for myself.
Do you recommend any improvements over what you have on Github?
You mentioned:

The board had unfortunate resonance at 4.2 GHz, which I think comes from the poor grounding of the multiple ground planes. I think I'll need to order a new PCB with better grounding.

Do you plan making new revision of the PCB?