TSP  

Comments from SJL Instruments: Thanks Shahriar for the detailed review! Doubly so, as we understand it's difficult to review the software when it's frequently changing. For any viewers with questions, we can be reached here in the comments, on the EEVBlog forum (SJL-Instruments), via email, or via our website. We are very focused on improving our software, and any suggestions will usually get implemented within 2 weeks. There's also a rather long thread on the EEVBlog forum where you can see some of the development history. (Now added to the description - thanks Shahriar!) For students and hobbyists, we offer a 35% non-commercial discount. One of our goals with this product was to make picosecond-scale measurements more accessible to hobbyists. To address the points you brought up during the review: 16:40 - The speed of acquisition is currently a factor of ~30x slower than the theoretical limit. We are working on a firmware revision that carefully optimizes all the internal timings, and expect this to be finished around the end of March. But in general, the acquisition will be somewhat slower than a higher-end device. We are trading bandwidth and timing precision for acquisition speed to bring down the cost. 16:53 - In point mode, the software can of course assume the signal is single-valued and compute the risetime of the averaged trace. This is exactly equivalent to turning vector trace mode on. But if the signal is not single-valued at each point in time, this will produce garbage. To avoid this, we currently only display measurements when vector mode is on. For the next software update, we are working on a way to detect when the signal is single-valued, so that the measurement can be displayed when it is valid. Note that measurements that always make sense in point mode (e.g. RMS, RF power, etc.) are available in all modes. 18:27 - Jitter measurement is scheduled for the next software update. In general, the current feature queue is available in the changelog on our software download page. 22:08 - The cause of the periodicity is somewhat subtle. For details, see our reply #395 to the EEVBlog forum thread. In short, they can be removed by a tweak to the software, and for now they can be avoided by setting some software settings appropriately. The next firmware revision will also significantly improve this issue. 23:28 - Unfortunately, de-embedding is simply not possible given only the probability density function of the waveform. De-embedding on a multivalued signal such as an eye diagram is only possible with a real-time oscilloscope, without a pattern trigger. With a pattern trigger and some software tweaks, though, the GigaWave could be used to analyze ISI with S-parameter de-embedding. 25:30 - Thank you for the suggestions. We will definitely add jitter measurements. For a TDR, we think that it's possible to simultaneously increase the speed by ~100-1000x and reduce the cost significantly, since the device controls the trigger source. This is something we have under development.

This has to be one of the most underrated technical channels on youtube. Thank you for providing us with such a wonderful resource and sharing your wealth of knowledge with us with each video Shahriar!

We have released software update v2.5.12 which significantly improves the speckle issue. The manual has also been updated to revision H12 (now H15), which adds detailed guidance on choosing appropriate CDF sampling settings. Jitter analysis is available as of v2.6.0. All measurements shown are now available in point mode. We have also released firmware revision 14, which significantly improves the acquisition speed.

I enjoyed your review. I'm glad you held off as SJL is moving at a fast pace.

Seeing the description of the 1-bit ADC made me think more of logic analyzers and how bad the low-end equipment field is for them. But in terms of the features of this device I could see using it for the eye diagram functionality, but it's out of my price range for my small hobby projects. A thousand bonus points for actually having a price and "buy now" on the website, though, and not just "request a quote" like so many manufacturers do.

Very cool. I saw Joe Smith's review of this and have been waiting for your review. Thank you!

Very good explanation and demonstration thank you. And the price difference between the 2,4 and 8 channel is surprisingly reasonable

My signals and systems exam I am currently studying for is going to benefit from this video.

Long time no see you. We all miss your magistral explanations. Welcome back!

Absolutely awesome work my friend

Thank you for sharing, Shahriar !

Fantastic review. Thank you

Very good video. I love the information provided.

Thank you for the great video! I also love the quality of the overhead (microscope?) camera. How can it display the chips names so visible? What is the model number? Thank you!

спасибо, отличный контент.

Interesting scope. Though that it uses a 12 bit ADC seems a low res, though 8.4 effective bits is even more lackluster. Not a major complaint but it does partly take away one of the bigger advantages for sampling scopes. Though, at this price point one will not find a real time scope that competes in bandwidth nor resolution. It also feels like the device could use a dedicated DC power input, USB-C is great and all but these seems fairly power hungry. (especially the 8 channel version.) Input protection could be nice, but I also understand not having it from a bandwidth standpoint. But 6 GHz isn't all that fast so surely one could have something protecting that 1 volt-ish capable front end. Range switching would be a wonderful feature creep, but relying on external attenuators has its own advantages.

10:38 I'm having troubles to understand the name of the USB signal generator. What's the name of it again? I'm looking for a 200MHz clock generator and apart from the Stanford research clock generator CG635 I couldn't find a cheaper solution.

Hi can this device do sampling for a sine wave? Or it just can take samples of a pulse wave???

Hello In user manual of this device have been written that we need thousand of millions of triggers to make a waveform So if the input signal be a 10khz repeatitive pulse but very short (

*Abstract* This video review examines the Gigawave 6400, an affordable sampling oscilloscope designed for high-bandwidth applications. The discussion covers the fundamental differences between sampling and real-time oscilloscopes, the Gigawave 6400's internal design, and its practical use cases. Key strengths include its ability to analyze eye diagrams in backplane testing scenarios, while limitations such as speckle effects and lack of S-parameter de-embedding in certain modes are noted. The review highlights the instrument's potential, particularly if firmware refinements and a more targeted software approach for specific applications (e.g., serial analyzers) are implemented. **Keywords:** sampling oscilloscope, Gigawave 6400, eye diagrams, backplane analysis, jitter, CDF, PDF *Summary* *Intro* * *00:00:00* Introduction to the Gigawave 6400 sampling oscilloscope, its price point, and its goal as an affordable, high-bandwidth oscilloscope. *Real-time vs. Sampling Oscilloscopes* * *00:00:30* Key differences between real-time oscilloscopes and sampling oscilloscopes. * Real-time: Data converters bound by the Nyquist criteria for comprehensive waveform capture. * Sampling: Focus on front-end sampling; useful for repetitive sequential data. * *00:02:30* Sampling oscilloscopes remain valuable for their cost-effectiveness in analyzing high-bandwidth data, especially regarding data communication. *How Sampling Oscilloscopes Work* * *00:03:00* Sampling oscilloscopes reconstruct waveforms slowly from individual data points captured between clock cycles. * *00:04:00* Basic sampling oscilloscope architecture: comparators, triggers, delays, and a DAC for setting thresholds. * *00:05:00* Sampling oscilloscopes generate a Cumulative Distribution Function (CDF) rather than a Probability Density Function (PDF). This allows for higher trigger rates with less data. * *00:05:30* Limitations of sampling oscilloscopes: slower signal construction with less trigger events and potential speckle effects in the final measurement. *Gigawave 6400 Hands-On* * *00:06:00* User-friendliness and build quality are key questions to be evaluated. * *00:06:30* Design and build quality: * Solid construction and RF absorbers for isolation * Separate circuitry for the trigger channel * *00:07:00* Internal Design : * FPGA, similar architecture for all channels, equation for PDF/CDF conversion included on the internal board. * *00:07:30* Components overview: * Ultra-fast silicon germanium comparators (with no ESD protection) * Flip flop for trigger reset * Delay controlled ring oscillator for precise timing * One to four clock distribution for simultaneous slicing * DACs for setting comparator thresholds *Measurements and Testing* * *00:09:30* Channel return loss and isolation measured using a network analyzer. * *00:10:30* Testing with the Leo Bodnar Pulse Generator: * Measuring instrument's inherent rise/fall time. * Assessing trigger behavior with infrequent pulses. *Using the Gigawave 6400* * *00:11:30* Understanding the instrument's limitations is crucial: * Can't capture data right at the trigger edge due to hold-off time for DACs. * Suggestions provided at the end for potential improvements. * *00:12:00* Finding a Signal: * Adjusting trigger threshold to capture the signal. * Setting trigger rate consistent with signal frequency. * *00:12:30* Analyzing the Signal: * Using Power User Mode for more control over CDF range and timebase. * Adjusting CDF sampling options. * *00:13:30* Finding the Edge (in Default Mode): * Setting appropriate timebase and hold-off to locate the edge. * Changing resolution for faster sweeping (low resolution for searching, then increase). * *00:14:00* Analyzing the Rising Edge: * Capturing sharp transitions. * Built-in measurements for rise time (20-80%, 10-90%). * Vertical measurements for values like peak-to-peak. * Adding cursors to visualize jitter (currently no built-in jitter measurements). * *00:15:30* Analyzing with "Dot" Visualization * Switching to "Dot" mode for eye diagram analysis. * Color grading indicates the statistical likelihood of signal presence. * *00:16:30* Limitations: * Need external delay to view the edge at the trigger point. * Measurements are unavailable in "Dot" mode. *Comparison with Keysight DCX Oscilloscope* * *00:17:30* Gigawave 6400 results are validated against a high-end Keysight DCX (primarily for ground truth, not direct performance comparison). * *00:18:00* Similar waveforms captured by both instruments. * *00:18:30* Consistent jitter and rise time measurements between the two. *Analyzing Real-World Signals* * *00:19:00* Ideal use case: Multi-channel analysis for backplane or SerDes applications. *Experiment: Bit Error Rate Tester (BERT)* * *00:19:30* Setup: * BERT generates data (3 Gbps) * "Eyesight channel" for backplane emulation * Oscilloscope channels: * Ch 2: Direct data from BERT * Ch 3: Data after Eyesight channel * Ch 1: Trigger * *00:20:30* Results: * "Vector" mode is not useful for PRBS data * "Dot" mode reveals eye diagrams * Demonstrates the Gigawave 6400's value in analyzing signal degradation and jitter in backplane scenarios. * *00:22:00* Discussion of 'speckles' in measurements: * Caused by CDF differentiation errors. * Not always present across all measurements. * Developer feedback suggests this could likely be fixed. *Eye Diagrams and Backplane Analysis* * *00:22:30* Gigawave 6400 demonstrates value in analyzing eye diagrams: * Shows clear differences in eye opening. * Histograms provide signal distribution information. * *00:23:00* Further eye degradation can be simulated. * *00:23:30* Key use-case: Analyzing issues within PCB design, backplane design, connectors, etc. * *00:24:00* Limitations: * S-parameter de-embedding not currently supported in eye diagram mode (hoped for future addition). *Comparison with DCAX* * *00:24:30* Gigawave 6400 results closely mirror those from the high-end Keysight DCAX. *Two-Tone Measurement Experiment* * *00:24:30* Setup: Two tones (1 GHz and 1.1 GHz) fed into the instrument, measurement on one channel. * *00:25:00* Results: * Shows beat frequency as expected. * Fourier transform reveals both tones correctly. * This functionality allows for frequency information and signal shape analysis. *Potential and Refinement* * *00:25:30* Points to consider: * Developers seem passionate and provide a calibration certificate. * Understanding the instrument's scope is vital. * Focusing on specific applications could be beneficial (e.g., backplane and serial analyzers). * An 8-channel version with jitter analysis, eye diagrams, equalization, and TDR capabilities would be a major improvement. *Conclusion* * *00:26:00* The Gigawave 6400 has significant potential, especially if firmware issues are addressed and the software becomes more targeted. Disclaimer: I used gemini to summarize the video transcript. This method may make mistakes in recognizing words and it can't distinguish between speakers.

I just can't follow your explanation of the 1-bit ADC... :-(

I noticed the title, looked over at my Tektronix TDS-820 that I just bought. Then I looked up the price of the GigaWave and realized it was like 4x what I paid for it.

the us navy rules. x2