TSP

Рет қаралды 20,212

Күн бұрын

In this episode Shahriar takes a close look at the new JouleScope JS220 dynamic energy analyzer:
joulescope.com
The USB-C instrument architecture and details teardown is presented alongside the main PCB X-Ray images. The instrument's unique six orders of magnitude current measurement and switching networks is also described and analyzed. The measurements confirm the six unique shunt resistors (alongside pairing CMOS devices) which can be chosen both manually and automatically by the autorunning software.
The GUI multimeter view is used to verify the accuracy of the unit across all current ranges and voltage ranges using a pair of Keithley 2450 SMUs. The dynamic behavior of the instrument is presented using a wide dynamic range load using various excitation waveforms and frequencies. Finally, a Z-Wave IoT motion sensor characteristics and communication patters is analyzed.
www.TheSignalPath.com
/ thesignalpath
www.Patreon.com/TheSignalPath

Пікірлер: 55

@Joulescope Жыл бұрын

Thank you for the comprehensive review, Shahriar! You did an excellent job demonstrating and communicating the JS220 measurement capabilities and limits in a succinct video! <a href="#" class="seekto" data-time="162">2:42</a>: The 2 x TI ISO7762 isolators use capacitive coupling for up to 100 Mbps data rate. Those lines in the X-ray must be the coupling structures. Cool to see! <a href="#" class="seekto" data-time="186">3:06</a>: We designed the digital GPIO to be separately isolated, too, but the standard product references it to USB ground. With a few trace cuts indicated on the PCB back and external supply, it can be isolated. <a href="#" class="seekto" data-time="414">6:54</a>: This approach to "not miss anything" is our Enwavify(TM) technology, which is new to the JS220. Enwavify significantly improves the JS220's dynamic measurement accuracy from the previous-generation JS110. <a href="#" class="seekto" data-time="702">11:42</a>: Noted: Add over-loaded voltage and current warnings to UI <a href="#" class="seekto" data-time="854">14:14</a>: The design has provisions for a "soft-fuse", which we plan to add early next year. The "soft-fuse" will be software configurable, but we will have a separate static & automatic threshold for fixed ranges to avoid damage. For now, the device does not have much protection from significant over-current in fixed-current ranges. <a href="#" class="seekto" data-time="1062">17:42</a>: Idea: add software trigger on live waveform view when zoomed in. <a href="#" class="seekto" data-time="1210">20:10</a>: The Device Current Ranging preferences only apply to the JS110. The JS220 is all on-instrument. The JS220 has some configuration options which we still need to add to the UI. <a href="#" class="seekto" data-time="1372">22:52</a>: Thank you! <a href="#" class="seekto" data-time="1443">24:03</a>: Note: the current_range signal is non-causal here ;) We will fix this to account for signal path delays in the i/v/p path. <a href="#" class="seekto" data-time="1585">26:25</a>: The initial large spike looks like inrush current, perhaps when the radio's regulator gets switched on? <a href="#" class="seekto" data-time="1671">27:51</a>: While the accumulator is useful and simple, you can also add dual markers over regions of interest to display the charge & energy consumption. <a href="#" class="seekto" data-time="1682">28:02</a>: With the JS220, the GPIO no longer uses the least significant bits. They are conveyed on independent data streams from the JS220 to the host. We need to update the UI documentation. <a href="#" class="seekto" data-time="1690">28:10</a>: Thank you! We will be releasing UI improvements throughout 2023 that will be free to JS110 and JS220 customers. <a href="#" class="seekto" data-time="1705">28:25</a>: I look forward to hearing more about how you put it to use!

@electronicayexperimentos Жыл бұрын

Hi Joulescope. Nice product. Only one question. I know the pc is everywhere at this days. But, could you design an interface with a tablet? It could be very small and a little giant design has more sense. Best regards.

@Joulescope Жыл бұрын

@@electronicayexperimentos Thank you for the suggestion! The Joulescope JS220, like the JS110, is a USB-connected instrument. As of today, we only support Windows, macOS, and Linux hosts. A Microsoft Surface Pro tablet should work with the JS220, but I have not tested it. I think that iOS and Android tablets tend to be locked down preventing general-purpose USB access unless you jail-break them. I suspect that we would need a network interface, like ethernet or Wi-Fi, to work with iOS tablets. I don't think that Bluetooth could keep up with the full-rate sample data, but we could potentially do all the UI processing on the instrument and use Bluetooth. However, we would need a LOT more RAM & processing on the instrument. We could potentially use a small form-factor linux computer, like a Raspberry Pi 4, to provide this functionality using the existing JS220. Do you have any specific tablet platform in mind?

@bansci Жыл бұрын

Great to see active responses from the supplier/designers! Thanks for sharing the product and the responses

@electronicayexperimentos Жыл бұрын

@@Joulescope Hi. No I don't have a specific paltform in mind. But I used to use a pico scope. Nice product, but you have to carry a laptop and cable. In the bench the scope uses a space much bigger than the size of scope. I sold the pico because tektronix presented the 2 series. I belive that is posible an interface that you can insert a Joulescope and a tablet. Something like a tek 2 series. You finish your work, power off and take off the tablet and use it as a cover. The tek 2 series has a very small footprint in the bench. Regards.

@Joulescope Жыл бұрын

@@electronicayexperimentos Thank you for the additional detail. Many customers control their device under test using a computer, so having that same computer host the Joulescope is not an issue. I certainly understand the desire to keep things small and self-contained. Our focus for early 2023 is on the Windows/macOS/Linux experience, and I will keep this in mind for the future. Would you prefer to connect the JS220 to your Tek 2? I am unfamiliar with the Tek 2 architecture, but it's possible that the USB ports are high-speed and connected to an internal Linux computer. If so, Tek could integrate the JS220 as a high-dynamic-range, isolated, 300 kHz BW current probe, which also measures voltage. We provide open-source software to make this easy.

@campbellmorrison8540 Жыл бұрын

Oh my this is complex. Im an old 69 yo engineer and this stuff is just amazing compared to what I was used to. Quite incredible what can now be done in a very small PCB.

@IanScottJohnston Жыл бұрын

Very nice PCB design, and a great explanation of the dynamic switching in both hardware design and how it performs in real experiments. Certainly looks like a great unit for characterising power consumption of battery powered equipment during DEV, and right down into the nitty gritty.

@BogdanO45 Жыл бұрын

Thank you the great presentation. The device is quite remarkable with its design.

@fredflickinger643 Жыл бұрын

I appreciate your pushing the unit to places we might want to go!

@James-wb1iq Жыл бұрын

I have one of the previous models, and it's my favorite thing.

@hnasje66 Жыл бұрын

I am using the "old" version to measure and determine the lifetime of BLE rcu's. Really like the product. Also good price/performance. Another positive thing is the open-source software that comes with the product.

@wickedprotos1937 Жыл бұрын

Excellant narrative!!!!!!!!!!

@pepe6666 Жыл бұрын

i always liked the idea of having different scales for measurement like that. never got round to implementing it in anything. i can only imagine that to do this you must have quite a heavy bandwidth constraint

@NikiBretschneider Жыл бұрын

It would be nice to see the voltage across the current sensing port during the experiment at <a href="#" class="seekto" data-time="877">14:37</a> on the oscilloscope. The voltage should somehow change (in steps) as the device switches its range, because this changes also its internal resistance. If the current ramps relatively slowly compared to the range switching speed, this potentially could cause some relatively steep voltage steps across the current sensing port. I would like to see how steep these steps are (if they are somehow visible), because fast edges lead to high frequency content, which potentially could cause some issues inside the DUT.

@supernumex Жыл бұрын

I have used the NRF PPK 2 (power profiler kit). It's only $100, but works surprisingly well! Would be cool to see some comparisons

@martinguthrie2658 Жыл бұрын

I have compared the NRF to the P1125 (wide dynamic range current measurement power supply). The NRF is similar to this class of current measurement device in that they all use FET switching to get the wide dynamic range. The problem is that FET switching induces artifacts into the measurement, and you can't tell the difference between the artifact and real device current. The P1125 does not use a FET switching design and does not have artifacts. With one of reference currents I use for testing, the NRF started oscillating between two range settings, and didn't stop.

@SvenKillig Жыл бұрын

Thanks for reminding me, they're back in stock, just ordered one :-)

@jholworld Жыл бұрын

I used the ppk and also the first joulescope. It boils down to how low current you are measuring and how much you trust measurements. I compared this with a keithley DMM & source and found Joulescope to be more accurate for my needs of measuring dynamic uA’s. As someone else mentioned the ppk was jumping between two ranges for my application too so the accuracy was not questionable. All in all for over 100uA dynamic current measurements I would be ok with ppk but lower current I more trust the Joulescope or even say a 6.5 dmm with I can easily program etc too using lab view etc. just my opinion though. Ppk2 I haven’t tested.

@2MOSFET Жыл бұрын

Nice video and nice product. Thank you. I am using something very similar: Power profiler Kit II from Nordic Semiconductor.

@t1d100 Жыл бұрын

Thank you

@ogitakasi3030 Жыл бұрын

GOOD

@MrAndri26 Жыл бұрын

Hi, I am interesting on what xray machine you use to inspect the PCB, the resolution is awesome.

@Ziferten Жыл бұрын

I was really interested in picking one of these up after watching your review... then I saw the price tag. 1000 USD will buy you a used Keithley 2400 if you're patient enough.

@James-wb1iq Жыл бұрын

A 2400 is nice - but a Joulescope is 1000x faster. You can pick up microsecond accurate transmit / receive burst timing with it, check switching converter current peaks, and calculate the power consumption of short interrupt handlers. They can also stream samples to your hard drive for days, capture digital inputs, and send digital stimulus to test boards from python scripts. Really fantastic for developing battery powered devices. Also 10x cheaper than an equivalent Keysight solution, and 50x smaller. So don't hold back - you'll love it.

@akosbuzogany2752 Жыл бұрын

Is shorting the wires at the input of the instrument to reduce the the loop area of the leads as opposed to having them shorted at the supply and the load?

@armandraz7725 Жыл бұрын

Hey Shariar, could you also do a comparison with P1125 from Sistemi? Curious how they stack up.

@N2Stooges Жыл бұрын

At around <a href="#" class="seekto" data-time="1441">24:01</a> in the video, why does it appear that the range selection changes from setting #5 to #1 approximately 15 usec BEFORE the leading edge of the measured current? Am I missing something?

@Joulescope Жыл бұрын

Good observation! Unfortunately, we have not invented non-causal prediction. The i/v/p sample path has additional delay compared to the current_range, and we have not yet accounted for this delay. Aligning the current_range signal is already on our todo list!

@cooperlonely Жыл бұрын

👍👍👍

@TheDefpom Жыл бұрын

I reviewed this too about a month ago, I tested on a low power LoRa device, and also showed the test board Joulescope supplied in case anyone is interested in seeing that board.

@xDevscom_EE Жыл бұрын

That X-ray is pure joy. -AD9226 is 12 bit 65 MSPS, not 16 bit- (edit: it's TI ADS9226 16bit 2MSPS SAR) ;). I've noticed that when you switched from overloaded 20uA range with 1A applied back to normal range reading of 1A was quite a bit off and recovering back to normal after. That hints of clamp recovery or possibly overheating in the frontend which shifted values off?

@IanScottJohnston Жыл бұрын

I spotted that also and thought uh oh!.....resistor overheating.

@Thesignalpath Жыл бұрын

The ADC is ADS9226, not AD9226. Unfortunate naming overlap. But the ADC is indeed 16-bits (it has to be, you could not get this performance from 12-bits).

@xDevscom_EE Жыл бұрын

@@Thesignalpath Right, TI part, not ADI here.

@NiHaoMike64 Жыл бұрын

I'm surprised they're not using active I/V conversion on the lower current ranges.

@tamaseduard5145 Жыл бұрын

👍🙏❤

@BartSimpsonHJ Жыл бұрын

This looks interesting but unfortunately doesn't quite meet the requirements I'm looking for. Does anyone know of a power/current waveform analyzer with at least 100 MSPS sample rate and a current resolution of 1 µA? I'm familiar with the Keysight CX3300 but that's way too expensive for the few times per year it would be used.

@Joulescope Жыл бұрын

@BartSimpsonHJ A few questions: 1. What measurement bandwidth do you actually need? 2. What system bandwidth do you actually have? 3. What maximum current do you need with resolution 1 µA? As far as bandwidth goes, the next step up from the JS220 is a dedicated current probe and oscilloscope, but you lose the high-dynamic range capability. Next step is the CX3300. Unless you plan on measuring current & power directly into an IC, the bypass capacitance combined with the cabling impedance often significantly limits the system bandwidth. We actually designed the Joulescope's 300 kHz measurement bandwidth to line up with system bandwidth in the most common measurement conditions. 50 mOhms total resistance (source, cables, connectors, JS220, target) and 10 uF target bypass capacitance is 318 kHz system bandwidth.

@BartSimpsonHJ Жыл бұрын

@@Joulescope I need to be able to measure very short current spikes after clock events. So far we're relying on average measurements and simulation results. The only actual measurement of such spikes I did, was using a CX3300 demonstration model (but even with that, getting a propper measurement without interferences wasn't easy). In that particular case, the current went from a few nA to 140 µA and back in less than 0.5 µs. Our system bandwith should be >25 MHz (estimate) and I think I limited the bandwidth on the CX3300 a bit (so not the full 200 MHz) to get rid of some noise.

@Joulescope Жыл бұрын

@@BartSimpsonHJ That is a very demanding measurement. A Joulesocpe JS220 would measure those short events accurately, but they would be filtered with 300 kHz bandwidth. You may also need to select the fixed 180 µA range or add a small amount of load capacitance to keep it in the 18 µA range. Whatever detail you may want in that 0.5 µs event would be smeared over multiple JS220 samples. However, it sounds like you might not need high-dynamic range. If you want to measure 140 µA max with 0.1 µA resolution, then that is only: log2(140/0.1) => 10.5 effective number of bits (ENOB) Many oscilloscopes are only 8 bits, but you can get 14-bit oscilloscopes, I think that are > 12 ENOB, so oscilloscope + current probe is possible. However, finding an off-the-shelf current probe with low enough leakage, low temperature coefficient, and high enough bandwidth could be very challenging. You could design something custom, but then the CX3300 starts looking like a better deal, especially if you don't already have a suitable oscilloscope. Unfortunately, I don't have any more cost-effective recommendations.

@PlasmaHH Жыл бұрын

I wish it was available for 24V

@physiqueDrummond Жыл бұрын

Down to earth question: banana cables... ours are dying in the students lab... any good brand and/or sources?

@florianhofmann7553 Жыл бұрын

Pomona used to be good, but nowadays everything is going down the drain quality wise.

@supercompooper Жыл бұрын

I'm searching for an AC power analyzer with harmonic analysis and high frequency real time logging... Seems not to exist for under 25k? Anyone know of one? Single phase on.

@robbrown2007 Жыл бұрын

Newtons 4th PPA

@supercompooper Жыл бұрын

@@robbrown2007 ha it starts at 23k list

@robbrown2007 Жыл бұрын

@@supercompooper Take a look at PQube3 also

@dgo42 Жыл бұрын

There is only two optical isolator but there is two digital isolators ISO7762 4/2 channels each, 100Mbps

@TrickyNekro Жыл бұрын

Nice little thing, being isolated is very useful, though for the price, I would go for a second 6.5 multimeter ( DMM6500 ) and not this. I also personally don´t see how that can be all that more useful in the field, you normally wouldn´t do that kind of analysis on the field, unless you need to analyze/repair some niche IoT out in the wild that charges with 10 different methods. Nice thing to have as an extra if you do a lot of IoT and do not know what to buy next, and maybe when you need to service a niche product for niche market on the field as it is always the case with these gadgets, they leave you wanting for more and have you to put the big boys on the table when things go unexpected.

@Maltanx Жыл бұрын

Well, it has a nice set of features and I can see myself using it happily, but I just checked the price... Ouch, 1000 bucks is just way too much for what it does. I opened the link expecting like 500 bucks maximum (and I would have thought that is really on the high side), but 1000 bucks it's just a show stopper in my opinion. It cuts away every possible hobbyist, even those with fairly high budget. If you are not a company you should really have a very, very specific need to buy this instead of an oscilloscope with a setup for current measuring...

@aDaWaN Жыл бұрын

The Nordic Semiconductor NRF PPK2 is a low-cost alternative to the JouleScope.