EEVblog Electronics Community Forum
EEVblog => EEVblog Specific => Topic started by: EEVblog on July 09, 2018, 11:15:04 pm
-
Review of the Omicron Labs Bode 100 50MHz Vector Network Analyser / Frequency Response Analyser and some experiments measuring bypass capacitors and characterizing a quartz crystal.
https://www.picotest.com/products_BODE100.html (https://www.picotest.com/products_BODE100.html)
https://www.youtube.com/watch?v=66s9easZKxU (https://www.youtube.com/watch?v=66s9easZKxU)
-
Omicron just makes nice gear. I'm quite familiar with their protection relay testgear. That comes in at least 10x the cost of the Bode but for what it does it is worth it.
-
Nice to see more network analysis stuff. This is a nice peace of kit and seeing you get usable results without ( i think) working through the manual and many, many hours practice, the software must be user friendly.
VNA's are great for impedance around 50 ohm. For very high or very low impedance you use IV measurements (impedance meters/analysers use that) and for some VNA's (like the SDR-kits VNWA) there are adapters to do that. The software needs to support that (like the VNWA does) if you do not want to do the math. I did that once just for fun with a 1968 build HP VNA (the ESR of a 1nF cap) See the attached photo. But there are also tricks like measuring a 100 K resistor parallel to a 50 ohm or a 0,001 ohm in series with a 50 ohm. The 50 ohm can be your cal load or just measure it, a 50 ohm resistor can be measured extremely well with a good calibrated VNA. In both cases I use the custom trace function so I do not need top do the math.
I would expect the Monicron supports IV measurements in the bodeplot function and "normal" wave parameters for the VNA part ?
Dave, the results of a vna are at the best as good as the calibration.
This is often a rather time consuming thing and needs a very good calkit (a male and a female based kit and then the things you used fort that fixtures) and very advanced software implementation to do n-term error corrections. That is why VNA's have very extensive calibration functions and things like de-embeding and port extension functions. Even a hobby VNA as the SDR-kits VNWA can do that. I missed that in your review.
Are the fixtures and their cal-kit parts already characterized in the software ? I would expect a cal certificate with the numbers
The resistor value does not have to be a precision one, as long as you know the resistance and are able to tell that to the VNA. Things like open and short are more important (OK, for 0-50 MHz probably everything will do ;) ) I use my SDR-Kits VNWA often in the span 1 KHz tot 10 MHz and calibration there is not really a big thing. Above 500 MHz it becomes the most important part of the measurement.
I like the fact it can make bode plots and do s-parameters, that combination is very nice.
The DUT fixtures look cheap if you compare them with the keysight ones. But I do not know the prices from those, probably much higher.
http://literature.cdn.keysight.com/litweb/pdf/5968-5329E.pdf (http://literature.cdn.keysight.com/litweb/pdf/5968-5329E.pdf) this pdf tell a lot about the things to know when using fixtures.
When I started playing around with VNA's there was almost no information for non RF graduated EE's That is why I made a series of tutorials. They are not flawless but I tried to make it easy to understand and for most practical. I used the SDR-kits VNWA for it, back then more or less the only choice.
For those who are new in this field, check them out, they are free pdf's. http://www.pa4tim.nl/?p=1594 (http://www.pa4tim.nl/?p=1594) All in English and some are translated in French and German by other VNWA users.
Keysight has a free PDF, the impedance measurement handbook. https://literature.cdn.keysight.com/litweb/pdf/5950-3000.pdf (https://literature.cdn.keysight.com/litweb/pdf/5950-3000.pdf) very interesting and "easy" to understand
-
Thanks Dave - great first video - looking forward to the others. Flexibility looks off the scale.
-
https://www.eevblog.com/forum/eda/bode-plots-come-from-hendrik-bode-who_s-name-is-pronounced/ (https://www.eevblog.com/forum/eda/bode-plots-come-from-hendrik-bode-who_s-name-is-pronounced/)
... Just saying...
-
Hi,
I would say that the creation of cal kits is a very tricky task in the GHz region, but not really at 50MHz where the Bode 100 is operating. I have made a lot of cal kits for ham radio use, and I calibrated them by comparison with a Keysight E5071C using a Keysight eCal module. Below 100MHz 30dB or better can be reached even with homebrew open, shorts and loads. Even leaded components could be used.
The accessories that come with the Bode 100 are all made using very classic construction techniques and no Gigahertz trickery. Everything more seems to be luxury.
-
The isolation transformer seems to use a regular Hammond type aluminum cast enclosure. Perhaps $8 a piece or something. The main unit seems to use a standard enclosure as well. Nothing against it in terms of performance but I'd think for the selling price something custom could be made to stand out better.
-
The isolation transformer seems to use a regular Hammond type aluminum cast enclosure. Perhaps $8 a piece or something. The main unit seems to use a standard enclosure as well. Nothing against it in terms of performance but I'd think for the selling price something custom could be made to stand out better.
Do it :)
-
As I mentioned in another, similarly named thread, I broke down the entire BOM and came up with about $110 for a single unit. I'm building one and will characterize it - then probably start selling them because, why not?
-
Much luck ! IMHO you will be able to reproduce the hardware, but all the software and GUI is another story. This could well be a few mythical manmonths minimum ... >:D
-
Much luck ! IMHO you will be able to reproduce the hardware, but all the software and GUI is another story. This could well be a few mythical manmonths minimum ... >:D
This. And the promise that you will be available when something is wrong. And the design work that you do not have to earn back. And the samples that you don't have to send to Australala...
-
... that you don't have to send to Australala...
La la land is on the west coast of the USA. ;)
-
Doh!
I complete missed the AD834 multiplier in the receiver can.
The part number wasn't really visible direct on
(https://i.imgur.com/iTkc4y3.jpg)
-
Using that analog multiplier is kind of a surprise: Its an expensive part and helps to reduce spurious signals due to harmonics of the LO. However its also limited accuracy and stability. So the 24 Bit ADC used later on does not make that much sense anymore. At least the analog multiplier (likely as a mixer) reduces the need for filtering. So maybe they use some of the CMOS switches for the possibility to do extensive internal adjustment runs to compensate for the limitation of the analog multiplier. It does not even look like they could bypass the AD834 for the low frequencies, where there is no absolute need for a mixer.
-
Photos are up:
https://www.flickr.com/photos/eevblog/albums/72157669079722607 (https://www.flickr.com/photos/eevblog/albums/72157669079722607)
-
Ups - I think the last few posts should have gone to the tear-down thread and not here.
-
Looks like so, I ended up here as someone mentioned in the other post...
JS
-
Sweet bit od kit. The price point might be a bit steep, but dang does that software look good. It does exatly what its suposed to without exesive messing about, setting stuff, reading manuals etc. If only otger test gear came with software that us even half as good as this.
Seing a mutiplier makes perfect sense to down convert to an IF that the ADC can sample, and we can make very very nice ADCs for 100KHz due to audio. Mixer harmonics can be compensated for with calibration and the 24bit ADC gives you exelent dynamic range(they are not too expensive ether)
I do have a Agilent Vector Signal Analyzer that does DC to 10MHz with similar capability(signal gen, dual input) but the software is simply not there to do much more than gain/phase and spectrum analysis(tho has exelent dynamic range). It could do all this can do too with some extra test fixtures, but i neved had a big enugh need to make such software using GPIB, nor would it be anywhere near as nice as this.
-
Thanks for this post, Dave!
If anyone would try to re-eng the input receiver channel schematics, or have some ideas how to build one, please, share.
Couldn't afford such an expensive FRA, but would like to try to make something similar for homelab.
-
Sweet bit od kit. The price point might be a bit steep, but dang does that software look good. It does exatly what its suposed to without exesive messing about, setting stuff, reading manuals etc. If only otger test gear came with software that us even half as good as this.
+1
-
Thanks for this post, Dave!
If anyone would try to re-eng the input receiver channel schematics, or have some ideas how to build one, please, share.
Could not afford such an expensive FRAs, but would like to try to make something similar for homelab.
I am also looking for Venable Industries FRA 32xx series analyzers older versions of software (vers. 4 for WinXp).
The first half of the receiver part should be similar to an scope input stage, and a few extra switches to choose between the input and an internal cal signal / the source. For the rather low frequency range a simple mixer, some AA filtering (not that much with an SD-ADC) and that the good resolution ADC could be all it takes. A difficult part would be to get a well defined frequency response for all gain settings - because of this I would prefer to have some of the gain switching behind the mixer. 50 MHz is not that high in frequency, but a IF in the 10 kHz range is even easier to handle. Both Channels seem to share a single LO that is likes driven from the same master clock. For a low cost version I would imagine one could get away with a switching mixer, maybe CMOS switches.
There are a few DIY VNA projects around, that also use DDS chips and high resolution ADCs - as the software is about the largest part of the project, it might be easier to start with such a circuit so that maybe one could have a start on the software. It may take changes with the input stages to go all the way to the DC coupled range.
The other alternative would be something like the red pitayar or analog discovery HW, using the faster ADCs and direct sampling. This common HW might help to find some open source SW, at least for a start.
-
I realise the Analog Discovery has nowhere near the breadth of capability the Omicron has but it would be interesting to see a few comparisons.
Does the Analog Discovery FRA give comparable results, at least within it's comfort zone?
-
Using that analog multiplier is kind of a surprise: Its an expensive part and helps to reduce spurious signals due to harmonics of the LO. However its also limited accuracy and stability. So the 24 Bit ADC used later on does not make that much sense anymore. At least the analog multiplier (likely as a mixer) reduces the need for filtering.
It is a delta sigma ADC, this type of ADC typically have hardware very narrow band selectable digital filter, which greatly reduces requirements for external analog filters. This comes at the expense of increased conversion time though.
-
The isolation transformer seems to use a regular Hammond type aluminum cast enclosure. Perhaps $8 a piece or something. The main unit seems to use a standard enclosure as well. Nothing against it in terms of performance but I'd think for the selling price something custom could be made to stand out better.
Do it :)
Sorry do what?
-
(https://i.imgur.com/iTkc4y3.jpg)
What are your guesses why MELF resistors were used?
-
Mouser and Vishay singing audiophile praise to MELF resistors, with the familiar "no distortion" thingie.
High-end audiophiles want their analog-to-digital converters (ADC) and digital-to-analog converters (DAC) and amplifiers to exhibit the lowest noise possible and be ultra-stable. MELFs delight audiophiles because they deliver low noise performance, uniform component design, excellent stability and precision, and best of all, they provide nearly ideal resistor performance – introducing essentially no discernable signal distortion to pass through.
https://www.mouser.com/blog/reating-a-sound-people-cant-resist-melf-resistors (https://www.mouser.com/blog/reating-a-sound-people-cant-resist-melf-resistors)
-
Form the color code those resistor seem the be 5.1 K and 1 K with 0.1% tolerance. They might be used for there low tolerance, maybe in a kind of differential amplifier between the AD834 mixer and the ADC.
Compared to smaller SMD resistors the MILF form factor might offer a slightly better excess noise because more resistive material is used, though I doubt this would be a factor here.
-
There are 0.1% low thermal PPM SMD resistors, you do not have to use MELF ones .
-
"MELF Resistors - The World’s Most Reliable and Predictable, High-Performing Film Resistors"
Long-term stability results in a superior maximum resistance change (ΔR/R) of 0.05 % after 8000 h operation for high precision MELF resistors.
If MELF does indeed have superior long term stability, maybe that was a deciding factor for them.
-
Yes i was also thinking about better aging characteristics as a possible reason
-
Given the target customer for their gear, Omicron was very brave to give this unit for an Eevblog review!
They intimated exactly this in their recent newsletter:
"...And this is the moment when you have to take risks. Risks like sending a Bode 100 to Dave at EEVblog."
:)
-
The isolation transformer seems to use a regular Hammond type aluminum cast enclosure. Perhaps $8 a piece or something. The main unit seems to use a standard enclosure as well. Nothing against it in terms of performance but I'd think for the selling price something custom could be made to stand out better.
Hi,
now I have the opportunity to test the B-WIT100 and a functionally equivalent homebrew transformer on a Bode 100 and on the sharpest knife of LF VNAs, a Keysight E5061B-3L5 (35k$, 5Hz to 3GHz). A few € wont do the job, but with about 70€ you can homebrew something decent:
https://electronicprojectsforfun.wordpress.com/injection-transformers/
at the bottom.
-
The VAC T60006-L2030-W514 core appears to be something exotic, $30 Canuck bucks here at Mouser plus shipping, i will probably pass.
-
I picked up a T60006-L2030-W423 from the Mouse for about 1/3 that much. I want to play around with a 2:1 winding.
-
The VAC T60006-L2030-W514 core appears to be something exotic, $30 Canuck bucks here at Mouser plus shipping, i will probably pass.
I picked up a T60006-L2030-W423 from the Mouse for about 1/3 that much. I want to play around with a 2:1 winding.
Please tell me about the results. My prediction is:
Als long the coupling inductance is in the range of 200mH and strays are low (less than 1uH) results will be comparable. A look at Omirons SPICE model gives you some idea:
* BWIT 100 WIDEBAND INJECTION TRANSFORMER MODEL
* CREATED BY OMICRON Lab ON APRIL 7, 2011
* VERSION 1.0
*
* CONNECTIONS: PRIMARY SIDE SIGNAL
*
| PRIMARY SIDE GROUND
* | | ISOLATED OUTPUT 1
* | | | ISOLATED OUTPUT 2
* | | | |
* | | | |
.SUBCKT BWIT 1 2 3 4
* CAPACITORS
CPRI 1 2 14e-12
CPRISEC 1 3 85e-12
CSEC 3 4 14e-12
*
RESISTORS
RPRI 1 5 0.36
RCORE 6 2 35e3
RSEC 8 3 0.362
* INDUCTANCES
LLEAKPRI 5 6 0.85e-6
LLEAKSEC 7 8 0.85e-6
LMAG 6 2 199e-3
* IDEAL 1TO1 TRANSFORMER
E 9 4 6 2 1
F 6 2 VM 1
VM 9 7
.ENDS BWIT
Probably a lot of similar cores with the same material would do.
-
I realise the Analog Discovery has nowhere near the breadth of capability the Omicron has but it would be interesting to see a few comparisons.
Does the Analog Discovery FRA give comparable results, at least within it's comfort zone?
Could not compare Analog Discovery (=AD) FRA's performance directly against any of those kilo-$ FRAs like Omicron or Venable. It would be interesting to know if anyone has a chance to test it.
But I tried AD to measure Gain-Phase response of a very noisy DIY AC-DC swithcing PSU. The DUT PSU had a high enough conducted EMI level so all my attempts to measure it's frequency responce with USB PC scope-based FRAs failed.
Surprisingly, AD was the ONLY device among other low-cost FRAs that gave a well-defined FRA responce of DUT PSU.
-
Can you show us some of your plots. I am interested in seeing your results.
Robert Bolanos
-
The isolation transformer seems to use a regular Hammond type aluminum cast enclosure. Perhaps $8 a piece or something. The main unit seems to use a standard enclosure as well. Nothing against it in terms of performance but I'd think for the selling price something custom could be made to stand out better.
Hi,
now I have the opportunity to test the B-WIT100 and a functionally equivalent homebrew transformer on a Bode 100 and on the sharpest knife of LF VNAs, a Keysight E5061B-3L5 (35k$, 5Hz to 3GHz). A few € wont do the job, but with about 70€ you can homebrew something decent:
https://electronicprojectsforfun.wordpress.com/injection-transformers/
at the bottom.
EEVblog APPROVED! :-+
(https://electronicprojectsforfun.files.wordpress.com/2018/10/bode100_2xvact60006-l2030-w514transformerpictureoutsidebox3.png?w=656)
-
Dave, you need to look forward to my "Nude Virgin Current Injector" !
Stay tuned
Wolfgang
-
Not nude but injector :D
Parameters are similar or mostly better like original ;)
[attach=1]