开云体育

I want to to do A/B testing with different amps driving my feedline and just a transformer. I want a long lasting relay that will hold up in my outdoor enclosure. I want a 12V DPDT relay. I'd like to keep the price under $5. I had success with a NEC RSP24H2C12H, but I don't find that available anymore. Does anyone have a favorite they recommend? Thanks, Mikek

Has anyone explored ways of building some kind of oscillator into which calibrated levels of phase noise can be introduced? Calibrated here means "value reasonably close to" rather than NIST/NPL traceable:) Clearly the purpose of such a source would be for testing phase noise measuring setups by providing "additional" PN on top of that in the basic oscillator. It's easy to detune the oscillator, generate mixing products and use this to check response to spurii and moderately easy to measure the noise floor using a single oscillator, power splitter and a long and short cable adjusted for quadrature at the DUT and REF inputs (adjusting oscillator frequency for quadrature). In theory the standard measurement processes should be calibratable in steps - but generally there are a fair number of them (measure detector constant, measure the noise bandwidth of the analyser, measure gain of the LNA, measure loss in path to analyser). It would be nice to have a way of being confident that these were all correct. The only approach I can think of would be to use a power combiner to combine a stable low PN 100MHz oscillator with filtered broadband noise from a noise diode (like an HP346 - only good above 10MHz hence suggestion of 100MHz oscillator, not the more ubiquitous 10MHz) and rely on the principle of equipartition to split this noise power equally between AM and PM. This would rely on the phase noise measurement system discriminating against AM (which it ought to do of course). I haven't got a suitable broadband noise source so I haven't actually built one of these as a test but would be interested to know if this was anything more than a fun application of thermodynamics. I also vaguely thought about Leeson's equation for the oscillator but really this only has two useful variables: power and temperature. Varying temperature is going to be a no-go as the frequency will change too and a single oscillator working over a wide enough range of power levels doesn't feel very practical either. Anyone have any ideas? Alan

If you want to avoid monotonicity errors, resistor tolerance (actually matching) has to be within 2**(-n), which would be 0.1% for a 10-bit converter, and 1 ppm for 20 bits. This is achieved in integrated DACs through the nearly perfect matching obtained with lithography. Non-monotonic DACs are really interesting to watch on a spectrum analyzer as you vary the amplitude of a sine wave, and even more interesting if you add a bit of dc offset to the digital waveform. Still, it’s a fun and cheap way to build a DAC of modest resolution and you will learn a lot along the way! Gary NA6O

I bought one on sale from Amazon on a lark for $135. It was "free returns, seller pays shipping". Nothing to lose. I can test it and send it back. As I had done with the last "180 MHz" piece of junk. That's not about to happen. It's fantastic! Even at the regular Aliexpress $157 or the Amazon $169 its the best deal for a real DSO. It runs on 5 V so I'm planning to hack in 18650s, BMS and boost. https://www.eevblog.com/forum/testgear/wow!-$123-hanmatek-dos1102-initial-comments/msg5201193/#msg520119 I've started [email protected] for anyone interested in making them portable as well as other issues and mischief? Have Fun! Reg

I bought an early one which had a FW bug and borked itself. After about 6-9 months I got a replacement one version higher. It also borked itself. The front panel uses an STM32F103. I have 2 front panels and the main unit which I had just modified to 3 wire plug. It works and there is a partially completed FW to drive it from the command line using a "blue pill" that I've not tried. I bought a used Keysight 33622A out of sheer frustration. There's a huge thread on EEVblog about mods. If I take it apart I can probably get it in a small flat rate box. Have Fun! Reg

-- Prof. Thomas H. Lee Faculty Co-Director, SystemX Alliance Director, Stanford-Samsung Research Initiative Allen Ctr., Rm. 205 420 Via Palou Mall Stanford University Stanford, CA 94305-4070 http://www-smirc.stanford.edu

As I teach my college students: there is NO difference between Audio Frequency AC math and so called "Microwave" math. It all about the circuit models. Wavelength is the issue for what is important. In low frequency electronics, component size is extremely small compared to a wavelength - we ignore parasitic elements because their magnitude is much smaller than the lumped element value of the component so we use simple lumped models and ignore parasitics. At 10 kHz, who cares if the resistor leads have 10 nH of inductance. However, at 10 GHz that same inductance represents a reactance of 62.8 ohms, significant if you are in a 50 ohm system. Its the argument of "Lumped" vs Distributed" that defines what is important at the frequency of operation. Not arbitrary designations. After I spent time in the MMW world working on systems for Uncle Sam, I would say "Everything below 1GHz is like DC, everything below 20 GHz is IF, and we get serious at 100 GHz". A lot of this thinking reflects availability of parts, the "make vs. buy" decision. We have the Decimeter spectrum, the Centimeter Spectrum and the Millimeter - nice clear cut decades that define frequency ranges. Where do you think the terms UHF, SHF and EHF came from (look at the military designations). What can we ignore and whats important? We when pass thru the arbitrary boundary of "Lumped vs Distributed" we leave the realm of simple lumped element circuit models that use voltages and currents to define the behaviors of our circuit elements and we enter the world of transmission line field concepts where we no longer look at I & V as carrying the "power", instead we consider field concepts like the Poynting vector and Maxwell-Heavyside mathematics. The fusion of math at this boundary is that of the EE and physicist - they have to provide the same (similar) answers. When lumped element thinking no longer adequately describes the behavior of your circuit elements, you have to go to alternative distributed field theory concepts (harder math) which, by the way, always works independent of frequency, albeit more cumbersome to solve. So the frequency that was considered "FBM" over the years changed as our measurement technology changed. In my engineering days 1 GHz "sort of" was where microwaves started (3 GHz was technically the boundary for centimeter waves), and 30 GHz was clearly MMW. Waveguide would work fine at 100 MHz but would be too costly. However, I saw waveguide at 430 MHz for the planetary radar at Arecibo, when I visited in 1986, so, no, waveguide sizes vs frequency are not the defining issue for use, cost & simplicity are. What was the argument about anyway? Sort of detoured from VNA's. Was it who defines what frequency range as magic vs. plebian ("microwave vs RF")? Depends on what the work is and who is doing the speaking. I have heard UHF TV broadcast engineers "pooh-poohing" AM Radio station engineers "You dont know how tough it is...". Jeff Kruth

Hi, Is there any work-around you can use to measure S12 and S22 using a cheapo VNA which is primarily designed only to provide S11 and S21 data? Thanks. J.

I have an HP 905A sliding load which is complete with all of the interchangeable pins and bodies for male/female N and APC7, but there's a warning about never mating the connector pins with a Precision-N male. Is there a spec like MIL-C-39012C for Precision N? That spec has the standard dimensions for checking pin offset, and runout, but the most recent 1994 version doesn't mention Precision-N. Is that MIL-C spec the canonical spec for N connectors, or is there a higher authority that defines the tolerance and dimensions and includes the Precision-N specs? The manual says I can use IPA to clean the connector bodies (which have the usual greenish decomposed foam guck on them) but nothing like Acetone or other solvents. I'm wondering if a gentle thrash in a warm ultrasonic bath of distilled water and surfactant then an IPA rinse might be acceptable? I can't get the guck off with IPA alone. Can anyone see why using volatile organic solvents on the purely metal parts would be a problem? The airline, load and main body are free of contaminants. the lock and slider works fine. I'm laser-cutting a replacement EVA insert for the box. I hope it won't turn to goo and powder in my remaining lifetime. Last question, the spec says the gap between the shoulder of the N mail pin and the face of the female socket should be 0.001" to 0.007" but the 905A manual suggests pressing the plunger until the pin stops moving, then locking it. That doesn't feel at all sensible to me. I'd prefer to push it gently in, then back it off by a tiny amount in case there's any differential thermal expansion/contraction that might cause excess pressure on the socket/pin. -- Neil https://youtube.com/MachiningandMicrowaves

" signals. I don't like the architecture, but the 5 V TTL compatibility is very valuable. it with 3497A GPIB commands or solder headers in place, connect an ATMega and run bespoke >code on that." I like the idea of using the Atmel uC based GPIB, (built in or grafted on), to do the special task. Is the inguard controller part of the DVM or the 3497A? I have not read the manuals yet. Can the 3497A GPIB be the controller of the 3457A and also receive commands for sequence of actions from a laptop? If so, command the 3457A relay card to use a set of wires, issue commands to relay cards in the 3497A, and get A * B sets of measurement terminals to use by program.

Anyone have any experience modifying/repairing Gore PhaseFlex cables? I have a couple of long (25 ft) Gore cables, unfortunately terminated with TNC connectors, that I found fairly cheap. They are in good shape I would like to cut them into 4 or 5 ft sections and build a few shorter test cables terminated with some high quality N connectors. This is the classic Gore fabric covered cable with the purple color and black stripes and measures about 0.30 in or 7.6MM OD. The cable has great characteristics past 10 GHz. I'm looking for some suitable connectors that would be fairly easy to install, and have good performance up to about 3 GHz. I have a 8 GHz VNA for testing after construction. Not looking for perfection or precise phase stable behavior, just a few more high quality test cables to do mostly scalar S11 and S21 measurements, usually below 1 GHz. Like duplexer and filter tuning. Pre-built Gore PhaseFlex cable assemblies are very expensive. Even used ones. I have two that I was lucky enough to find at an affordable price, but can't determine who made the type N male connectors installed on them. I'm wondering if they are custom, or available somewhere, or perhaps something like a Andrews/Commscope heliax type connector for 1/4 superflex heliax (FSJ1-50) might fit. Can't seem to find a dimensional drawing of the cable cross section. But one simple promotional drawing I did locate, shows an inner spiral metal tape type shield and a spring like outer shield. Not your grandfather's RG8U coax for sure! Any had one of these cables apart? Ideas? I haven't made any cuts yet... Thanks.

My recent, fairly pointless, but lots of fun, project :-) Capacitor substitution box project

Anyone here developed a SCPI command parser for embedding in an instrument? Or know of one well developed. I have googled and never found a parser that appears to be very compliant to the specification. A simple use case might be to have an Arduino UNO analogread() result available through the USB COM port where the UNO appears to be a SCPI instrument. Regards, (Forrest) Lee Erickson

Reginald, for the sake of accuracy, Erik designed the tinySA, inspired by the nanoVNA. The nanoVNA itself was originally designed by eddy555, a Japanese Radio Amateur, whose design was then adopted & adapted by Hugen, a Chinese Radio Amateur, to become the pre-built nanoVNA which is available today. And that design itself has spawned both a wide degree of ownership and the inspiration for further low-cost VNA designs. Erik did use the nanoVNA-H4 hardware as the basis for his tinyPFA. The hardware of both the nanoVNA (-H and -H4) and the tinySA make extensive use of SA612 mixers, SA5351A-B-GT frequency synthesizers, STM32F3xxxx microprocessors, USB powering & communication, a Lithium cell, and ChibiOS for the firmware, plus other bits & pieces. That is a very powerful mixture that can readily be re-purposed, as Erik has demonstrated. [And other ?processors and OSes are both available and very useful, of course.] 73, Stay Safe, Robin, G8DQX PS: Much more at https://nanovna.com/, plus where the links lead. On 14/08/2023 01:50, Reginald Beardsley via groups.io wrote:

For some time I have tried to calibrate a Keithley 261 Picoampere source, but I have found that some parts of the design are inherently prone to drifting. Specifically, the calibration of different ranges involves switching between different 200-ohm potentiometers (R113 thru R119) in series with a wafer switch. The current is approximately 8mA, which means that any contact resistance in the switch or any wiper resistance change in the pots will result in some parasitic resistance change, and the voltage regulator output will drift. As a result, calibration is not stable. The bad news is that the wafer switch and the pots have aged over time and will continue to age. Replacing them, assuming replacement parts can be found, will not correct the initial sensitivity of the design to parasitic resistance changes. I am seriously thinking of redesigning the voltage regulator in a way that is immune to such resistance changes. Has anyone already done this change ? I would hate to reinvent the wheel. Or would anyone like to share details of my design ? Cheers, Joel Setton