NanoVNA V3 (6GHz)

[Mechatrommer]

2. Official V2 units have epoxy blobs as an anticopy-mechanism, which decrease our ability to repair.
 - https://groups.io/g/NanoVNA-V2/message/1151
3. OwO talks about anticopy protection of the v3 design.
 - https://www.eevblog.com/forum/testgear/nanovna-v3-(6ghz)/msg3815639/#msg3815639

i dont think epoxy blob can do any good at protecting from reverse engineering, they must use even better method. such as some sort of encryption in comm-lanes etc.. PCB can be easily R.E. with proper equipment even hobbiest can afford. scratched ICs can be guessed or find alternatives. the other protection is legal/patent/copyright but that can cost millions $ i guess.. thats why afa R.E is concerned, there's not much a small budget company can do.. unless they have some secret recipe that only them can imagine. my bet (their last stand) will be on the FW/SW encryption/mangling/etc

ps: i de-potted a very thick (about an inch) potted bike's ECU (damaged) last time for the sake to see whats inside or if i can repair it, i managed to break some traces and components because i want it to be quick. i can do a perfect depotting if i really wanted to but i ended up it being in the trash. the damaged board i still can see many things to learn or whatever..

[nctnico]

In the end copy protection mechanisms just hurt the end user. It needs to be paid for. A better approach is to come up with good software and keep adding new features over time. That way the competition (cloners) are always behind.

[tautech]

If a reputable brand, like Siglent were selling something similar for $5-6k, I would have bought that already. I'm looking for used stuff in the sub-$2k range.

I've heard some Chinese whispers...... ETA unknown.

[mawyatt]

Custom ICs in a modern advanced small geometry process are a good way to protect IP, because they are extremely difficult to reverse engineer. If one intends to protect IP by means of using a custom chip and designs for such, then RE becomes almost impossible within a reasonable timeframe and cost. Here is where "IP Thieves" will revert to trying to get access to IC design files. Using insider techniques, like hiring away a original designer, or paying off company employees, paying off fabrication employees, blackmail, even using an IP spy which will get hired into the company with the purpose of stealing select valuable IP. Industrial espionage seems common place today, and many companies go to great efforts and costs to protect valuable IP.

Of course the downside of a custom IC is the non-recurring cost due to the initial design and fab for prototyping is quite expensive, and may not be viable for a small $ market. The upside is the recurring chip cost is small and the IP well protected and better suited for high $ markets.

Best,

[jjoonathan]

OwO, you stuck with the good cause longer than I did, and as a result also left more of a positive impact on the world than I did. I'd be a hypocrite if I asked you to stick with it, so instead I'll say thanks for what you have done 

[nctnico]

Custom ICs in a modern advanced small geometry process are a good way to protect IP, because they are extremely difficult to reverse engineer. If one intends to protect IP by means of using a custom chip and designs for such, then RE becomes almost impossible within a reasonable timeframe and cost. Here is where "IP Thieves" will revert to trying to get access to IC design files.

That will only work if the chip does something really unique that can not be achieved by using standard parts. All in all not that useful for IP protection (there are many people that have fixed broken equipment by creating circuits to replace an unobtanium chip).

A microcontroller with a verification mechanism inside that checks whether valid software is running and use encryption on firmware images is a better option. But there is no doubt that can be broken as well.

[mawyatt]

That will only work if the chip does something really unique that can not be achieved by using standard parts. All in all not that useful for IP protection (there are many people that have fixed broken equipment by creating circuits to replace an unobtanium chip).

Try reverse engineering one of the later System on Chip, or one of the older ones we did, good luck with that effort

The only unobtanium chips that were replicated with other chips and/or circuits were likely very old chips used in old products from long ago and thus fabricated in old processes and likely not much modern IP involved, if any.

I've seen lots of chips get replaced by a modern like chips, sometimes having to deal with the differing pinouts, but no modern ones I can recall where the actual chip was reversed engineered and duplicated. If you have any examples please provide those.

The replica that does come to mind is the Tektronix U800 type used in many older analog scopes, this was copied and duplicated, but this wasn't a custom chip, it was a simple hybrid which was easy to copy and replicate. Also vaguely remember an old HP chip that was replaced with some discrete JFETs and maybe a microcontroller.

A microcontroller with a verification mechanism inside that checks whether valid software is running and use encryption on firmware images is a better option. But there is no doubt that can be broken as well.

That works for IP that is firmware based, but of no use for hardware IP protection. How long do you think it would take to copy OwO unique IP if it's just hardware based using just microcontroller verification??

As mentioned by others just erasing the IC identification won't hold back someone long, as it's relatively easy to figure out the function and trace out the circuit, and you can even probe the circuit to check things out. If the entire function were integrated, that copy effort becomes much more difficult, and if the original design was implemented to prevent IP theft all the more difficult. Probing an IC outside just the I/O pins isn't easy either, and with sub-micron lines deeply embedded (one simple IP preventive measure) is something that will be difficult indeed!! Then the reversed engineering effort must be verified with another custom chip fabricated, hoping that something wasn't missed, this is expensive, time consuming, and highly risky.

Not saying a custom IC is the ultimate IP hardware theft prevention, but can make life miserable for those IP thefts out there

Best,

[nctnico]

Nowadays hardware isn't interesting. And nobody needs reverse engineering a chip; a functional replacement is good enough and cheaper as well given de engineering costs. Why bother with breaking into the front door when the back door is unlocked? Be creative. Many years ago some burglars went into my employer's building. They found a locked door but instead of breaking the door open, they simply kicked through the plaster board wall.

Nowadays the secret sauce is always in the software. Just look at modern low cost oscilloscopes. They all use the same hardware. The differentiator is in the software. IOW: if you care about IP protection, make sure the secret ingredient is in the software and make it hard to probe the software for its functionality.

[mawyatt]

Nowadays hardware isn't interesting. And nobody needs reverse engineering a chip; a functional replacement is good enough and cheaper as well given de engineering costs.

This misses the whole point of IP, if one can't duplicate or surpass the performance then one must "steal" it, or bail. IP isn't about me-too performance that can be easily replicated, it's about unique ability that isn't easy to do, otherwise it would have already been done by many!!

Just look at modern low cost oscilloscopes. They all use the same hardware. The differentiator is in the software. IOW: if you care about IP protection, make sure the secret ingredient is in the software and make it hard to probe the software for its functionality.

Yep those scopes do all use basically the same hardware! Why, because it's decades old me-too technology, and nothing unique IP to protect!!

Don't think the latest high performance scopes use the same hardware tho, know Keysight spent quite a bit on the new ADCs for their high end scopes, so did LeCroy, and likely Tecktronix. Why develop full custom ADCs at considerable expense, and likely have significant IP involved, when they could have used off-the-shelve ADCs from TI or AD and just use software to differentiate their products?  They couldn't achieve the performance levels required without custom ADCs and unique hardware IP, even with the software wizards at work, so they opted to develop full custom ADCs with considerable IP to help differentiate their products.

Software can't always fix flawed or performance limiting hardware, and this is where hardware does matter!!

Best,

[joeqsmith]

Right now all we have is the VNA core prototype (built out of many small modules). The final design could be anything at this point, display/no display, 4.3 inch vs 4 inch, capacitive vs resistive touch, full two port vs T/R + transfer switch. You can post your ideas and suggestions, I can't promise anything but I will read this thread occasionally and keep all the suggestions so far in mind.

In the mean time please do watch these masterpieces (might be available on Netflix):
https://myanimelist.net/anime/10087/Fate_Zero
https://myanimelist.net/anime/11741/Fate_Zero_2nd_Season
https://myanimelist.net/anime/9756/Mahou_Shoujo_Madoka%E2%98%85Magica
https://myanimelist.net/anime/11981/Mahou_Shoujo_Madoka%E2%98%85Magica_Movie_3__Hangyaku_no_Monogatari (this movie is the sequel to the TV series)

First, let me say I use the V2Plus4 from your store a fair amount and have not had a single problem with it.  This includes the firmware.  It's a great product for the price.   

Assuming the V3 works as intended, I could care less if you do not release the hardware or firmware to public domain.  If you want to stay with an open source design, determine the development costs and run a kick start.   Make sure the design goals are clear.  I would gladly pay up front and have no problems waiting for it's release.  If you don't meet your goal on kickstart, just don't continue with the development.   I think you have enough clout that people would trust you to pull it off.

I never use the V2Plus4 without a PC and have no use for the display, buttons or battery.  As long as there is a way to update the firmware if there's a problem, I don't how it's accomplished.   The software interface must be well documented.   Stick with the same protocol as the V2Plus4 and make my life easier.    I'm not a fan of USB but it would with staying with it.  SMAs are good.   While the full 2-port would be nice, for myself the use would be limited with with sqarewave drive.   I'm not expecting it to compete with that CMT I mentioned.   

I would like to be able to use it for narrow band work.  For us, it will come down to how low we can set the start frequency, noise, dynamic range, how much data and how fast it can sweep and be offloaded.   I would like to see 20kHz for a start freq with 110 dB dynamic range.  The rest is just icing on the cake.   

We are looking forward to seeing what you come out with. 

[Mechatrommer]

If you want to stay with an open source design, determine the development costs and run a kick start.   Make sure the design goals are clear.  I would gladly pay up front and have no problems waiting for it's release.  If you don't meet your goal on kickstart, just don't continue with the development. I think you have enough clout that people would trust you to pull it off.

the problem is.. he's not the only one developing this and there is some "corporate/strategic/market" decision going on behind. so this is not entirely up to him. i guess OwO is just a design engineer and not more...

[all_repair]

Kickstarter or crowdsupply is a good suggestion.  And it does not have to a big bang, the release can be in few iterations and so development effort can be recovered through few releases which have a good interval in between.  Or the potential buyers shall be slowly reduced through those who do.  Since then, there are LibreVNA and liteVNA.   Hugen is making themselve a vehicle for designer of tinySA, libreVNA (6GHz)and liteVNA(6GHz with LCD).     
 

[OwO]

all_repair: one of the designers you mentioned does not receive any royalties for their design, and the other designer receives peanuts. This is a warning for all potential designers about letting vendors access to your design who have a track record of not respecting the original developer's wishes and just steals your design if negotiations fall down. There are many better vendors to cooperate with to take your design to market, like Seeed studio or Elecrow. Clones or not, I'm doing far better now than if I had accepted a deal with that vendor. Engineers, please do not undervalue your work and make sure you demand fair pay for it.

[OwO]

I never use the V2Plus4 without a PC and have no use for the display, buttons or battery.  As long as there is a way to update the firmware if there's a problem, I don't how it's accomplished.   The software interface must be well documented.   Stick with the same protocol as the V2Plus4 and make my life easier.    I'm not a fan of USB but it would with staying with it.  SMAs are good.   While the full 2-port would be nice, for myself the use would be limited with with sqarewave drive.   I'm not expecting it to compete with that CMT I mentioned.   

I would like to be able to use it for narrow band work.  For us, it will come down to how low we can set the start frequency, noise, dynamic range, how much data and how fast it can sweep and be offloaded.   I would like to see 20kHz for a start freq with 110 dB dynamic range.  The rest is just icing on the cake.   

We are looking forward to seeing what you come out with.

I will keep this in mind. The low frequency support is in the plans and I'm starting to prototype various ways to do it. I think the main competitors will be the lower end of the professional offerings (copper mountain, picotech, etc) rather than NanoVNA related products. There are actually some aspect where we may be able to beat the professional offerings, like temperature drift and low frequency performance. But with the budget constraints things like sine wave output and wide adjustable power output probably won't make it.

[mawyatt]

Engineers, please do not undervalue your work and make sure you demand fair pay for it.

Very good advise.

Most older engineers have had times when they were taken advantage of, myself included. Most engineers are not good business folks, myself included, and an opportunity for others to take advantage.

Good engineers always want to do the best possible, it's ingrained, and others take advantage of that characteristic!!

Good luck with your endeavors,

Best,

[joeqsmith]

... The low frequency support is in the plans and I'm starting to prototype various ways to do it. I think the main competitors will be the lower end of the professional offerings (copper mountain, picotech, etc) rather than NanoVNA related products. There are actually some aspect where we may be able to beat the professional offerings, like temperature drift and low frequency performance. But with the budget constraints things like sine wave output and wide adjustable power output probably won't make it.

I don't see you competing with the professional products but instead forging a new niche area.   

All of these profession units will be sine wave drive.  Some support log sweep, bias Ts, adjustable level.    Freq range, tempco,  noise, dynamic range for a few low cost (<$25,000 USD) portable units:   

Pico Technology PicoVNA 106
0.3 MHz to 6 GHz at 10 Hz resolution
± 0.005 dB RMS ± 0.02 dB/°C. Measured at maximum bandwidth.
Transmission: ±0.1 dB, ±1°
Reflection: ±0.5 dB, ±4°
Dynamic Range: Typically: 118 dB @10 Hz
Guaranteed: 108 dB 10–4000 MHz 100 dB 0.3–6000 MHz

Copper Mountain S5065
0.01 MHz to 6.5 GHz at 1 Hz resolution
± 0.0021 dB RMS ± 0.02 dB/°C. Measured in 3 kHz bandwidth.
Transmission: ±0.1 dB, ±1°
Reflection: ±0.4 dB, ±3°
Dynamic Range: Typically: 130 dB @1 Hz
Guaranteed: 125 dB 0.3–4800 MHz 120 dB 0.3–6500 MHz 75 dB 0.1–6500 MHz

Tektronix TTR506A
0.1 MHz to 6 GHz at 1 Hz resolution
± 0.008 dB RMS ± 0.015 dB/°C. Measured at maximum bandwidth.
Transmission: ±0.2 dB, ±2°
Reflection: ±No spec. dB, ±°
Dynamic Range: Typically: 122 dB @No spec. Hz
Guaranteed: ≥124 dB typ. 2–200 MHz ≥125 dB 200 MHz–3 GHz ≥122 dB 3–6 GHz

Anritsu Shockline MS46322A-008
0.05 MHz to 8 GHz at 1 Hz resolution
± 0.006 dB RMS ± 0.02 dB/°C. Measured at maximum bandwidth.
Transmission: ±0.08 dB, ±0.5°
Reflection: ±0.37 dB, ±2.2°
Dynamic Range: Typically: 115 dB @10 Hz
Guaranteed: ≥100 dB 10 MHz–8 GHz

[Just_another_Dave]

Engineers, please do not undervalue your work and make sure you demand fair pay for it.

Very good advise.

Most older engineers have had times when they were taken advantage of, myself included. Most engineers are not good business folks, myself included, and an opportunity for others to take advantage.

Good engineers always want to do the best possible, it's ingrained, and others take advantage of that characteristic!!

Good luck with your endeavors,

Best,

Pricing your own work can be quite difficult (especially consulting services in which you can’t use the cost as a reference) and, at the same time, the best engineers I’ve known were always willing to collaborate and openly discuss about their projects or knowledge. That makes it easy for others to take advantage of them

Additionally, when you’re starting it is difficult to realize that some seemingly uninterested questions might actually be a way to get information on how to solve a problem that a commercial product has without paying a consultant

[all_repair]

Never been easy on pricing, and market judgement.  Most enginners which likely had worked or still working in big companies err wanting to price for high price high volume (which only monopoly like Apple can play).  The stable equilibrium is always either high price (=high profit) low volume, or low price high volume.  No room for other combination.

[mawyatt]

Regarding pricing for consulting, believe a good reference price is to use 2X the salary one would get working for a company. Most companies I've been involved with charge 3~5X salary for engineering time, so the 2X individual consulting rate is reasonable. At the senior advanced consulting level this is what folks seem to do, at lease the ones we know.

Also, don't short change yourself on travel time and cost, attorneys do it and consulting engineers are equally entitled. If your travel is extensive, like out of town and over night or longer, then a flat daily rate for travel + incurred cost (airfare, hotel, car rental, parking) is usually acceptable.

Best,

[optotester]

Would be interested too. Not sure if it was reported, but it is not possible to subscribe to the waiting list from an iPhone. The button just does not work and I never got the "OK" message, so maybe that could explain why the waiting list is filling slowly too.