SJL Instruments GigaWave 6400: 6GHz, 4-Channel Sampling Scope
Вставка
- Опубліковано 17 лют 2024
- Review of the GigaWave 9400 6GHz, 4-channel sampling oscilloscope
EEVBLOG, discussions about the GigaWave 6400:
www.eevblog.com/forum/testgea...
SJL Instruments
www.sjl-instruments.com/ - Наука та технологія
Its an amazing little device and they did not miss the mark that so many do with not having a usable API for production automation, i could see this being very useful for factory testing modern high speed cabling
It is also a fantastic price for what it can do, however at the cost that it is i would have really liked to see USB based firmware updating, or at least bring a JST connector to the rear. At least its small enough to not be annoying to send back for factory upgrades.
P.S. Until we implement USB DFU, you can request the necessary firmware flashing hardware from us at no charge. See the firmware upgrade instructions in the software section of our website. There is no need to send back the device unless you prefer that option.
USB DFU is now implemented as of firmware revision 14. The firmware can now be updated in ~30 seconds via the software. This update also significantly improves the acquisition speed.
Great review Joe, but I watched it a little too late in the evening not looking to see how long it was before I got into it. I’ll need to review a couple of the items that you brought up. I ordered one after I watched yours and Shahriar’s videos along your discourse with Samuel on eevblog. This device has a lot of potential. Thanks again.
It's very possible that some of what were mentioned during the reviews has been addressed, or is in the process of being addressed.
good jobbie for repeatable pulses
You can or course have some random signals, for example, I demonstrated up to PAM8 with it using pseudo random sequences. The clock does need to be synchronized.
What is the voltage on those kelvin leads when measuring the DC resistance? My 4 wire is 815mV which pretty close to the limit. Great review.
I assume you are asking about the voltage across the current sense outputs of the meter (of course with leads attached to DUT). Less than 1mV.
Software is pretty good. Works without issues on Linux. It is in Python, and Python API available too. Would be nice to have source code, as there are still some things that software could be improved on in the GUI, and the driver provided is closed source. But it should be possible to recreate it open source, as the serial protocol looks to be well documented (and it is very simple and strighforward, so writing one from scratch can be really just one evening to get usable data and entire own data pipeline).
And price is nice.
If only there was also some decent real-time scope with similar approach. Would love 500MHz 4-channel scope with similar interface, and similar price.
Good to hear that you're enjoying the software! If you do try to write some open-source software, we're happy to provide guidance on all aspects (and relevant code snippets from the official software if needed).
If you have any suggestions for the GUI, we implement user feedback in 1-2 weeks.
Could you clarify what you mean by "driver"? The scope uses the standard USB serial protocol, and drivers for the relevant chip (CP2102) are included by default in most modern Linux distros. The Python API is fully open-source - it uses the standard Python serial library, and you can adapt the code for your own needs.
Cool!
Joe can you test Megger, new Hioki and Flir multimeters?
About your delay line and stuff: why aren't they using something akin to regular scopes FIFO pipeline (with a sampling twist)? Sure, time intervals between samples will become random (especially if they add random waits between runs for safety), but it can be fixed in post. Like, for example for something really simple, interpolate (or fit a curve) and then resample at regular intervals. Ancestors couldn't do it because it was too computationally expensive, but now the problem can be unloaded onto a PC... Or the acquisition will take forever even if they can sample every 11ns? Or the problem is with measuring exact time offset between the trigger and current data frame?
The hardware has no way to measure delays - it can only generate a very precise delay, after which the sample is taken. This comes down to cost.
Many random sampling scopes do take samples at uncorrelated times and later reconcile them in software; but such an approach would push up the price by 5-10x.
Thanks for the review, Joe! My first question is why SMA? Wouldn't N-Type have been way better?
If I had to guess, size. If they used anything other than SMA the case size would have to be greatly increased. Think they wanted to keep it shirt pocket size.
@@mikesradiorepair Yes, portability ended up being the deciding factor. There is no other scope on the market with a comparable bandwidth and form factor.
If the SMA connectors are a deal-breaker, we do take custom requests via email. Lead time for an N-type version would be ~4-6 weeks at no additional charge.
@@SJLInstruments Nice to see a company that takes product improvement suggestions from users and works with customers that may require a different interface. Like I said in the reply to the original poster, I figured the choice of SMA connectors was probably based on size. Mike KC3OSD
Hi what is the difference between sampling heads and this device?
The device can be used as a sampling head by using Channel 1 as a dedicated trigger input. The applications overlap quite closely. We can offer more specific guidance if you share additional details about your application.
Is your resistance measurement equipment ground isolated or is there a ground loop via the USB that could be affecting the readings?
Isolated... I guess it depends on your definition. For all practical sense, the 34401 is isolated and it is not the cause of the shift.
Promo-SM