I'm thinking about a new product line that will need a sinewave input signal in the 1 MHz to 3 GHz range. 0 dBm should do.
We can test initial units manually, with a big signal generator, but if the idea is successful we'll want to automate testing. So can anyone recommend an affordable programmable signal source? Super low phase noise is not important.
There are some cute little USB boxes around; I just want a recommendation for one that people like.
We'd talk to it from a big Python program that talks to other instruments too.
--
John Larkin Highland Technology, Inc
lunatic fringe electronics
I was thinking about buying a box, all done. My test department would want a few.
The outputs of the gadgets would need to be measured too. Which we could do an a fast oscilloscope, or maybe one of those USB RF power meters or spectrum analyzers.
I guess an untriggered sampling scope is an RF power meter.
The idea is to not spend > 100K to generate and measure some clock signals.
--
John Larkin Highland Technology, Inc
lunatic fringe electronics
On a sunny day (28 Jun 2019 10:39:29 -0700) it happened Winfield Hill wrote in :
AFAIK it does not go below 35 MHz There is a lot of spurious around the selected center frequency. That chip is much like the ADF4350, using it, and wrote test software for it:
formatting link
I would do it this way (for low budget): Buy 2 universal LNBs with PLL control, almost all cheap new LNBs have a
9.75 and 10.6 GHz on chip oscillator PLL locked to a 25 MHz xtal.
I have a setup here where I lock it to an external synthesized source that is crystal locked TCXO.
formatting link
that circuit has been changed now, old circuit diagram.
So take one 10.65 GHz LNB, and synthesize an about 25 MHz reference for second LNB to stay between
13.6 GHz and 10.7 GHz, mix down the 2, lowpass above 3 GHz, and you are done. (The 25 MHz is internally multiplied in the LNBs by 390 or 426 to get 9750 MHz or 10650 MHz, if you use a different reference and not the 25 MHz xtal in that other LNB, then you can usually (depends on the LNB chip) steer away a lot from those frequencies,
and that is the trick, the difference frequency will be your output. Mixer chip... 2 LNBs, synthesizer board with PIC micro and ethernet chip for remote control and software I already wrote, Was published here this year, use it receive QO100 satellite.
Total < 50 USD. Time...
I do not know about simple cheap USB boxes that can do 1 MHz to 3 GHz sine wave output. My raspberry can do 150 kHz to 500 Mhz square wave xtal locked, so that will not work.
I have one of those (AliExpress actually). Beware though; the chip contains multiple VCOs and goes squirrely when it changes over and when it range switches until the VCO locks again. Plus you have the square-wave thing.
Another way to do what Jan talked about is to use its faster brother the ADF4356 somewhere above 3GHz and mix that with the (less adjustable but cheaper) output of an ADF4156, LPF at 3GHz. That gets you continuous coverage from 0..3GHz, but you still can't get clean continuous sweeps.
On a sunny day (Fri, 28 Jun 2019 20:42:45 +0100) it happened Tom Gardner wrote in :
formatting link
The simple 10 dollars solution, not tested,
2 twisted wire oscillators each at 10GHz or higher, fed into a dual gate MOSFET mixer. Output is f1 - f2. Frequency control: one oscillator is driven higher and the other lower from a diff amp. add varicaps or whatever have you to increase range.
Optionally a 2 USD (hehe) prescaler chip and a phase comparator, to make it PLL locked to some reference, else potmeter.
People that still use the Nipkov disk for viewing may have problems reading this.
I dreamt this up last night after watching a Riddick movie where he had to survive on some alien planet hunted by bounty hunters and all sort of creatures... Story of my life..? LOL
So YMMV. JoLa will have no problem with improving this scheme I am sure.
We have a new pulse generator - almost done - that uses two of the LMX2571 synthesizer chips, which are amazingly good. We see 1 ps RMS jitter, which is our measurement floor.
But I just want to buy a few boxes all done with USB and all, but not pay multi kilobucks. I don't need super signal quality.
--
John Larkin Highland Technology, Inc
lunatic fringe electronics
If you have one of th older HP sig gens in the lab with ieee, why not use that ?. You can get a network to ieee adapter at low cost now, and there will be open source software to drive it from Linux or FreeBSD. Either that, or look at the HP benchlink (?) lab control software, some of which is free.
PTS Instruments PTS-xxxx shows up on Ebay often. I have a 160, a 25O, and a dual 300. The words i would use are Awe Inspiring. Not suitable as an AM/FM modulated box, but the 1 Hz tuning step is awesome, as is the continious pbase.
But I think there are some ebay stuff using the same IC.
That's about as "inexpensive" as it gets for a ready-made version
formatting link
formatting link
Still in the vicinity of ? 4k, either of them.
The cheap "single chip" ones typically intended for ham radio use don't go low enough (35 MHz and up) and the output level is uncalibrated.
An Atten AT8010D only goes up to 1.1 GHz, so not high enough. A GA1483 or GA1484 covers the full range, but costs even more and may difficult to find now (not sure if they are still being made, hard to tell).
Keysight and R&S ones with similar specs cost about double the Siglent or Rigol ones. Rigol and Siglent cost about the same for the hardware alone, but Siglent includes pulse modulation functions while for Rigol you would have to buy them extra and they are somewhat expensive.
For USB control, you might take a look at the Valon synthesizer hardware at
formatting link
. He's just down the road from you in Redwood City. I have a few of the Valon modules around here and can recommend them. Good bang:buck ratio. Not as cheap as the latest shaky Shenzhen specials on eBay, but much nicer hardware. I had to scratch a similar last week, coincidentally enough. I put together a nifty multichannel generator using only common household items such as an AD9959 DDS eval board, some THS3491s, and an Arduino. :-P
Sounds like the application was pretty similar to yours if you do end up needing automation. It's basically a SCPI-controlled signal source for automated factory tests of up to four DUTs at once. I didn't want to require the operator to connect the device under test to a specific port, or to have to enter the port number manually after hooking up the cables.
Fuzzy photo taken in hotel room in Fremont while working on the test script:
formatting link
The box responds to typical SCPI commands over the usual USB COM port. What's different about it is that each channel can be assigned a separate "default" frequency and amplitude. When you send OUTput:CW or OUTput:DBM to specify new settings for a given channel, you can also include an OUT:TIMEout command that tells the generator to fall back to those defaults after a specified time.
So, given different default frequencies in each channel, the test software can infer what port the DUT is connected to simply by commanding the DUT to measure the frequency it sees at the beginning of the test cycle. This is easy enough given that frequency measurement is basically what the DUT is supposed to do.
During the test, commands are periodically retransmitted with OUT:TIMEouts of
10 seconds, and the box ignores any commands that match its existing settings in the channel. But if the test software or the PC goes down for any reason, the signal generator will time out by itself and revert to the frequencies used for port identification in plenty of time for the next test cycle. The software never has to initialize or reset the generator, or make any assumptions about its current state.
This box only covers 10 kHz-200 MHz so the details wouldn't be useful in your situation, but I'm now a big fan of this "timeout" technique. Worth considering for any new test hardware that needs to be used by non- specialists in a PC-driven environment, which is common enough these days. Obviously most DUTs don't measure the applied frequency, but if estimating it is something you can do for free in the firmware, it could save some headaches and operator training time.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.