*I realized that a one-minute section of the review where I speak about some key specifications is missing from the final edit. The instrument offers some of the best phase noise and harmonic responses in this formfactor. Please see the specification document here:* www.taborelec.com/GenericHendler/GetFile.ashx?D=0&DG=3831
It would be interesting to see the frequency hopping performance - i.e. switching and settling time as viewed on the MSO68B with it's 2GHz RF analysis BW. Maybe use the Freq Settling Time measurement of SVPC on the scope...
New video. No sleep for NJ here :) Most crazy RF PCB I've seen. Also RF interconnect between clock ref board and mainboard is interesting. How are these connectors called?
@@wadehsu2347 I totally agree with you, I never meant to imply it was wrong. It was just incredibly surprising to me to see one in there considering the instrument.
In my opinion, only some of the filters in this generator are SAW filters (those in flat housings). Those with higher housings look like multi cavity ceramic filters.
I think that the 100MHz to 500MHz conversion is implemented by the tiny SOT23 component, it should be a pair of PIN diode and used as a x5 frequency multiplier.
Very interesting video for me personally, as I am working on the signal processing side of another 6 GHz signal generator based on the AD9164 12GSps DAC. The PCB is a little simpler because all the RF is inside of the DAC chip. But you then you deal with all the nonlinear issues that DACs usually have.
Hi dear, you didn't mentioned at all the main feature of this series - an ultra low phase noise, which was a reason of such a complicate RF design from the beginning. The phase noise is one of the lowest on the market for RF synthesizers, especially in such a small form factor.
There was a section which unfortunately was cut in editing. I have added a pinned comments for people to find the datasheet. I will also discuss it in future videos.
@@Thesignalpath Thank you. BTW, I'm actually architect & RF design engineer of that module. These PCBs you show are one of the very first versions of it.. I designed all of them, including Schematics & PCB routing :)) I personally made these threads on mech. walls & added these additional tiny screws (which sure was not easy to remove;) during the bring-up on MAIN PCB..
Shahriar , @22:49 , to see how good phase noise is , is it possible to make derivate on signal and look in logarithmic scald to see small micro level changes ?! in this way it shows the true performance of it
We needed phase coherent frequency source with deterministic phase characteristics during frequency change. It turns out it is hard to do with synthesized sources sharing the same clock due to the indeterministic nature of the ppl lock. I was hoping these sources are capable of such task, but unfortunately it doesn’t seem like it. The only solution for now would be using a multichannel DDS and mixers.
You absolutely right regarding the long time run coherence problems for units locked only with 10/100MHz common clock source only. Actually these sources are truly coherent because as you can see inside the 4 channel units they share common 1, 2.5 & 3GHz common LOs.. There is only one external clock board there.. Believe me, I'm designer of these modules..
Igor Za , I understand that the modules shares the same reference and LOs so they are truly phase coherent. I have no doubt about that. Our application however requires a deterministic and repeatable amount of phase accumulation when the frequencies are jump/ramp back and forth. With PLL synthesizer the lock time is indeterministic by nature. So we get a random phase from shot to shot even when the frequency change is programmed to be the same. With a shared LO and multichannel DDS and a mixer this phase accumulation is repeatable and deterministic, but the frequency and LO rejection are limited by the mixer.
This is a really great video. Thank you for posting it. I do have to note that the AD9208 is not an interleaved ADC. It is a straight pipe. I reckon what you're seeing as Fs/4-fin spur might be the subharmonic of the clock mixing with the analog input. If you connect a 3GHz bandpass filter between the sig gen and the clock input to the eval board and test again, that spur will most likely go away.
We were taught that one of the advantage of modulation is better range ie., RF range increases with increase in frequency but it is actually the opposite as we can see from 2.4g and 5g . Why is this so?
Questions: 1. How are the harmonic spurs these? 2. How is the phase noise? 3. What is the max level output they can do? 4. What power dynamic ranges do they do? 5. Can the 4 channels act on their own? Meaning different power and frequency levels.
@@try_to_guess_it I looked at the spec sheet. Looking for 26.5GHz and 50GHz. Doesn't look like they have any. I'd like to see a 26.5GHz and 50GHz in demo. If possible I would like at least one of each, but I'd like to personally calibrate it and see if it will work for my usage.
I am interested in the lowest power output performance compared to the 866X-004 on his bench .... Add a good LNA and see when the noise floor goes up. My 8665B-004 has the NF at least -150dBm
@@tchiwam 8665B & this 12GHz tiny module are slightly different in all aspects, therefore not comparable. Please compare apples to apples & oranges to oranges..
Hi, Big fan of your videos. I miss the tutorials though. I know its hard to make them and you have said so in your talk with Dave from EEVBlog. But could you please do more tutorial videos, particularly on electromagnetics, antennas etc. I also wish you could you do some videos on microwave design. Please consider including fair amount of theory and math also. That would be immensly helpful. Thanks again!
Would have preferred N-type rather than SMA myself. A little more rugged and longer life unless you keep sacrificial cables permanently attached. But that is all inconsequential as long as lab cat makes an appearance :-)
*I realized that a one-minute section of the review where I speak about some key specifications is missing from the final edit. The instrument offers some of the best phase noise and harmonic responses in this formfactor. Please see the specification document here:*
www.taborelec.com/GenericHendler/GetFile.ashx?D=0&DG=3831
"We could spend hours talking about all the interesting design techniques" - please do!
Yeah please do some videos about actually designing a RF device. For example designing a thing like the ERA synth with all considerations etc.
@Auto Fill what do you mean?
Totally agree
It would be interesting to see the frequency hopping performance - i.e. switching and settling time as viewed on the MSO68B with it's 2GHz RF analysis BW. Maybe use the Freq Settling Time measurement of SVPC on the scope...
New video. No sleep for NJ here :) Most crazy RF PCB I've seen. Also RF interconnect between clock ref board and mainboard is interesting. How are these connectors called?
Can you do a video on how TEM, TE, and TM mode waves are created, please?
Wow, The Raspberry PI was a surprise.
You don't say... WTF moment to say the least.
Charles Dorval , what’s wrong with having a raspberry pi in there? It’s a completely adequate hardware to drive a GUI and it won’t break the bank.
@@wadehsu2347 I totally agree with you, I never meant to imply it was wrong. It was just incredibly surprising to me to see one in there considering the instrument.
Charles Dorval, true. That signal generator probably cost well over 10K.
I saw Raspberry PI's with WWAN modules used in electronic traffic signs.
Great review! The multiple channels surely enabled some interesting and specialized experiments. Thank you for the effort, it was enlightening!
In my opinion, only some of the filters in this generator are SAW filters (those in flat housings). Those with higher housings look like multi cavity ceramic filters.
Yep, I saw that also, only the tiny square silver colored 'cans' were SAWs, the rest were ceramic BPFs. CTS's web site shows a few examples.
I think that the 100MHz to 500MHz conversion is implemented by the tiny SOT23 component, it should be a pair of PIN diode and used as a x5 frequency multiplier.
Very interesting video for me personally, as I am working on the signal processing side of another 6 GHz signal generator based on the AD9164 12GSps DAC. The PCB is a little simpler because all the RF is inside of the DAC chip. But you then you deal with all the nonlinear issues that DACs usually have.
Hi dear, you didn't mentioned at all the main feature of this series - an ultra low phase noise, which was a reason of such a complicate RF design from the beginning. The phase noise is one of the lowest on the market for RF synthesizers, especially in such a small form factor.
There was a section which unfortunately was cut in editing. I have added a pinned comments for people to find the datasheet. I will also discuss it in future videos.
@@Thesignalpath Thank you. BTW, I'm actually architect & RF design engineer of that module. These PCBs you show are one of the very first versions of it.. I designed all of them, including Schematics & PCB routing :)) I personally made these threads on mech. walls & added these additional tiny screws (which sure was not easy to remove;) during the bring-up on MAIN PCB..
@@try_to_guess_it very nice job. I enjoyed reverse engineering and discussing it. :)
@@Thesignalpath Thank you!
very appreciate for such a series of tech video, I was wondering if you have any specific video for a tutorial on designing the analog and RF PCBs
A Cat scan is always a must.
Shahriar , @22:49 , to see how good phase noise is , is it possible to make derivate on signal and look in logarithmic scald to see small micro level changes ?!
in this way it shows the true performance of it
So it is offset loop PLL to achieve low phase noise.
Woah, I didn't even spot how they were showing which channel was selected until he pointed it out!
We needed phase coherent frequency source with deterministic phase characteristics during frequency change. It turns out it is hard to do with synthesized sources sharing the same clock due to the indeterministic nature of the ppl lock. I was hoping these sources are capable of such task, but unfortunately it doesn’t seem like it. The only solution for now would be using a multichannel DDS and mixers.
You absolutely right regarding the long time run coherence problems for units locked only with 10/100MHz common clock source only. Actually these sources are truly coherent because as you can see inside the 4 channel units they share common 1, 2.5 & 3GHz common LOs.. There is only one external clock board there.. Believe me, I'm designer of these modules..
Igor Za , I understand that the modules shares the same reference and LOs so they are truly phase coherent. I have no doubt about that. Our application however requires a deterministic and repeatable amount of phase accumulation when the frequencies are jump/ramp back and forth. With PLL synthesizer the lock time is indeterministic by nature. So we get a random phase from shot to shot even when the frequency change is programmed to be the same. With a shared LO and multichannel DDS and a mixer this phase accumulation is repeatable and deterministic, but the frequency and LO rejection are limited by the mixer.
Any video on RF transmitter and receiver module circuit diagram with explanation ie working of the circuit diagram step by step
This is a really great video. Thank you for posting it. I do have to note that the AD9208 is not an interleaved ADC. It is a straight pipe. I reckon what you're seeing as Fs/4-fin spur might be the subharmonic of the clock mixing with the analog input. If you connect a 3GHz bandpass filter between the sig gen and the clock input to the eval board and test again, that spur will most likely go away.
We were taught that one of the advantage of modulation is better range ie., RF range increases with increase in frequency but it is actually the opposite as we can see from 2.4g and 5g . Why is this so?
Questions:
1. How are the harmonic spurs these?
2. How is the phase noise?
3. What is the max level output they can do?
4. What power dynamic ranges do they do?
5. Can the 4 channels act on their own? Meaning different power and frequency levels.
1. -40dBc max.
2. Ultra low phase noise.
3. +15dBm max.
4. -90dBm min.
5. Yes, indeed.
@@try_to_guess_it I looked at the spec sheet. Looking for 26.5GHz and 50GHz. Doesn't look like they have any. I'd like to see a 26.5GHz and 50GHz in demo.
If possible I would like at least one of each, but I'd like to personally calibrate it and see if it will work for my usage.
I am interested in the lowest power output performance compared to the 866X-004 on his bench .... Add a good LNA and see when the noise floor goes up. My 8665B-004 has the NF at least -150dBm
@@tchiwam 8665B & this 12GHz tiny module are slightly different in all aspects, therefore not comparable. Please compare apples to apples & oranges to oranges..
Hi,
Big fan of your videos. I miss the tutorials though. I know its hard to make them and you have said so in your talk with Dave from EEVBlog.
But could you please do more tutorial videos, particularly on electromagnetics, antennas etc. I also wish you could you do some videos on microwave design. Please consider including fair amount of theory and math also. That would be immensly helpful.
Thanks again!
Would have preferred N-type rather than SMA myself. A little more rugged and longer life unless you keep sacrificial cables permanently attached. But that is all inconsequential as long as lab cat makes an appearance :-)
Nice cameo from Pooch!
Could you please extend the length of those 40-min videos? I would be glad if more of them topped the 45-min limit for Patreon!!!
Amazing as always
This is a real showcase of the kind of ingenuity Israeli engineers are known for.
I love the cat jump , Very funny @31:20
I was about to go to bed.... sleep can wait!
This guy is the Einstein of electronics. I dont know wtf he's talking about half the time but still interesting.
Those aren't circuit boards, they're works of art.
Waiting for the repair video
Amazing
Beau matos ça ! 😲
Great hardware let down by some very rough software. Not really acceptable in instruments (bench versions) costing $10-20k.
I would disagree. The firmware can always be improved....
Thank you for the "great hardware" definition...
Those 40 Minutes went past really fast
Until you get an Auto Dyno your laboratory won't be complete 😛
Roswell, 1947.
This is alien voodoo technology...
Black magic.