Voltlog
HTML-код
- Опубликовано: 23 окт 2024
- The Joulescope is a low cost precision dc energy analyser which is currently on kickstarter so check the links below.
→Joulescope on kickstarter voltlog.com/y/a...
→Joulescope website voltlog.com/y/w...
→Support the channel via Patreon, Paypal or Bitcoin
www.voltlog.com...
→Blog Post
www.voltlog.com/
#Joulescope #VoltLog #Teardown
Creator of Joulescope here! Great teardown and review! Here are some answers to your questions:
3:15 You're welcome! Not sure if you had a T15 Torx screwdriver easily accessible
4:00 The evaluation kit is not part of the normal Joulescope purchases
8:04 Resistors are 0.01, 0.1, 1, 10, 100, 1000, Joulescope keeps 20 mV max across the shunt resistor up to 2A.
9:00 Each Joulescope is factory calibrated using precision gear and a custom target ;)
10:55 Yes, and the opto-isolator is much less expensive than the faster capacitive coupled SiliconLabs chip
11:08 These FPGAs are very inexpensive for FPGAs: $3.65 USD each
11:11 Sensor FPGA does do shunt resistance selection and ADC communication - ADCs are U22 & U26
12:15 Agreed. Intend to fix for production units.
13:16 Using high-speed USB. Need 8 MB/s! FPGA implements external memory interface to get data into LPC54608
15:05 No separate driver installation necessary! Uses WinUSB which is included with Windows
15:40 Will add option to hide red min/max lines. Thanks!
17:26 Yes! I will be adding 2xGPI and 2xGPO - will have official Kickstarter update Tuesday
18:10 Great idea. Added to list. Thanks!
18:30 The UI allows you to zoom into the delayed buffer - not always intuitive. If you pan to the right (x-axis 30 seconds), you can see samples with much less delay. No compression.
18:53 Great idea. Added to list. Thanks!
Joulescope I backed your kickstarter. Seems to be a good decision. But of course I have to test it ;-)
Nice device, got any more spare demo units to send to channels?
Thanks for the comprehensive reply with answers to all of my questions!
@@TheDefpom Thanks for your interest! Unfortunately, I have no more Joulescope beta units to give away at this time.
I would really appreciate if you guys would smash that like button.
I noticed it on Kickstarter, and even though I don't really _need_ a precision energy analyzer I became interested. Then I saw its "low cost", had several heart attacks in quick succession, closed the browser tab post haste and that was the end of it.
yes, certainly not something you would get if you don't really need it.
Felt that exact same sentiment(except a lot less intense) when i became interested in the Mooshimeter.When i saw the price i nearly fell out of my chair.
Now, to be fair on the Joulescope, most energy meters around are pretty expensive.So... on that point of view it could be actually considered "low cost".
Thanks for doing the teardown and review. This was very helpful!
I have two things I would have wished you could have included in your review.
1. What is the voltage droop when changing current ranges. That is, when the microcontroller wakes up from sleep, and goes into a higher current mode, what is the voltage droop. The Joulescope has a finite response time while it switches ranges, and it would be good to see that on a scope.
2. What is the accuracy of the Joulescope when seeing current pulses that cross ranges. For example, if you have a microcontroller that wakes for a very short period of time (1s or 10s of us). Can the Joulescope capture that, and if so, what is the accuracy?
Hi Seth! [disclosure: I am the creator of Joulescope]
1. The voltage droop depends upon the system and can be calculated. Joulescope's switching time was designed to achieve < 3% drop @ 3.3V with a 1A step change and >= 10 µF target capacitance. For more detail including how to compute the expected voltage drop for your system, see the "Specialized Equipment" section in this blog post I wrote:
www.joulescope.com/learn/measure-current/
2. Joulescope's stated bandwidth is 250 kHz (4 µs period) with 0.8 µs switching delay typical (1.2 µs max). Joulescope' analog path then has a settling time. The User's Guide currently gives the settling time as 1.5 µs typical and 3 µs max, but I am working on a change that may reduce that to 1 µs max in the production unit. You can find some typical Joulescope measurements from the beta units in the User's Guide under the "Typical Behavior" section. You can download the latest version of the User's Guide here:
www.joulescope.com/download/
With dynamic ranging turned on, Joulescope takes about 2 µs to directly measure an accurate value after an over-range event. If you want to measure very short pulses more accurately, you can select a fixed current range for Joulescope. I will be adding more performance data to the User's Guide over the coming months.
Does this answer your questions? I am happy to add more detail or follow up!
Very nice item
Ben daha iyisini yaparım.
Great review!
thank you !
Very nice instrument and great review.
At 8:06 you mention shunt resistors (0.1 to 1kohm). This would lead to a burden of 10V at 10A. I suspect minimum shunt is 0.01 ohm (10 milli-ohm). At 14:30 specs say 0.01 ohm.
You are definitely right, as Joulescope replied here in the comments, the lowest shunt resistor is 0.01ohm. Thanks!
14 bit ADC only provides for 4 significant decimal digits of resolution. So the 6+ digits they are displaying in their software is nothing but noise after the 4th digit.
In 1983-84 we were building 4 channel 16 bit ADC units for the use with PCs that had 1 MHz sampling and built in memory to compensate for the slower transfer rate of the serial bus. Also the input voltage range of the module alone was over 100v. They were about the same size and price of this unit. We designed software that did digital oscilloscopes, Spectrum Analyzers (via Fast Fourier Transforms), and dataloggers. We later built 24 bit DAC modules that could sample at 8 mhz, after PCs got faster than 8 mhz ;)
We sold hundreds of thousands of these units in the mid-80s. I would like to think that either a superior product could be designed 37 years later or at least one at a tiny fraction of the cost. Granted this one also has built-in discrete current measurements (we had to use a separate module on one of the other 4 sampling inputs to read current as well), but in my mind that makes it worth maybe $80 to extreme techies. I highly doubt they sell more than a few hundred of these over the next couple years at the current price point.
A 14-bit ADC provides 16384 possible values. However, Joulescope averages 1,000,000 samples to compute each multimeter-view update and gains the advantage of noise dithering & averaging to produce a more precise value (standard deviation is reduced by 1 / sqrt(number_of_samples)). If you click the "accumulate" button in the multimeter view, you average the noise over an even longer time such that those digits settle down for static conditions. This allows you to make a more precise measurement, however, it's still only as accurate as the quoted specifications. Averaging noise this way is limited by the actual change of the underlying signal over time. For further information, check out:
en.wikipedia.org/wiki/Allan_variance
If you are just looking at Joulescope as an oscilloscope (like you are), then it's understandable you don't see the value. Joulescope favorably competes against very high-end equipment that costs more than 10 times as much. The key Joulescope features that add cost are dynamic current range (32 effective bits), very low voltage drop, extremely fast dynamic autoranging, and electrical isolation. These features enable you to worry about the challenge at hand, not how to set up your test equipment. People that have this problem are very excited about Joulescope. I have sold several beta units (which was not planned) to people who could not wait for the Kickstarter!
On volume, I think you underestimate how important energy consumption and battery life are to many, many products these days.
Wow, interesting stuff you were building in the 80s and selling them in the hundreds of thousands, you must of been working in a big company.
We were a tiny private startup, working out of a small building under a bridge. It was started by a couple ex-Boeing engineers.
Analog current sense circuit amplifier seems based on x2 cascaded MAX4239 (U14-U15).
Any thoughts how the joulescope compares to the ZS1100A, or at least anything on paper stand-out as important differences?
I don't like how their comparison table is constructed, tries to show advantages for their product where there aren't any. But for anything more than that I would have to get my hands on one for testing.
frankly it is better to buy a Nordic PPK II for a fraction of the price (under 100USD), since its exactly the same thing (done better). 10nA range is nonsense. Other than that, for having extensively studied the state of the art for ULP DUT power profile characterisation and ways to accurately predict how long is the battery going to last in the DUT in operation, which is what actually matters, in my opinion the best topology is what TI patented over 10 years ago, which is basically a software controlled DCDC with calibration controlled by a small mixed domain MCU, and another MCU couting pulses over time with hardware timer. This topology counts the total energy sent to the DUT, without discrete sampling, and it cannot miss anything. You let the DUT run for some days and get a very accurate estimation of your battery life. Unlike all these sampling methods, be it switched shunts or transimpedance amplifier topology, the drawback is the discrete sampling.
Good vid. But you might want to think about locking the exposure when handling something so reflective!
thanks for pointing that out, I need to improve on that.
Interesting project. Looks like it could be useful. Just wish it was cheaper - or that I was richer, LOL.
not cheap for hobby use but for someone who is developing a product and needs to determine the energy consumption when they look around for options, the price will not seem high anymore.
@@johncoops6897 As the creator of Joulescope, I rarely could afford the quality expensive gear, most of which also suffers from some pretty severe limitations... At least until you step up to the Keysight CX3300 family, which starts at $30,000 USD. Joulescope is a quality instrument built with precision engineering, just a new brand, like Saleae for example. Many professional engineers have already stepped up to back the Kickstarter campaign. You are right that the target market is much smaller than from a DMM or a low spec oscilloscope. However, measuring energy accurately is important to A LOT of engineers, developers, and makers. Current probes starting at $1000 USD sell pretty well but offer less accuracy than Joulescope for many real-world applications (which have bandwidth limited by the decoupling capacitors). I wrote a blog article about why, after 200 years of measuring current, it's still difficult to get dynamic range and bandwidth:
www.joulescope.com/learn/measure-current/
Battery life calibration is a really big problem. It heavily depends on things like chemistry, consumer profile, ambient temperature etc. Especially in the circumstances where you would need a unit like this, it is likely that an approximation would be useless.
I think it would be useful to have a base line, like here is your battery life if your device operates like this and your battery life has this capacity. All products on market use some kind of approximation when quoting battery life, which is often the case they do not match with real use because everyone uses the device differently but still they give you a base line.
Nice
$700 is "low-cost" ? I wonder what is normal price for this kind of product, any reference ?
N6705B we have in the lab cost 6k I think.
@@davidbleloch9565 6k sounds about right.
@@davidbleloch9565 that is really not comparable by any standard, this thing is just a switch shunt resistors topology, really not a 4 quadrant sourcemeter, totally not comparable. Just buy a Nordic PPK instead of this, its exactly the same but at the right price.
How does it compare with the Otii device from Qoitech?
Here is what Matt the creator has to say about that, they are somewhat different www.eevblog.com/forum/testgear/joulescope-js110-precision-dc-energy-analyzer/msg2212452/#msg2212452
3 seconds delay?! That's a lot. If their price is 400$ they will never sell it.
Almost no delay, but the user interface allows you to view at a buffer offset from current time. The "current time" is at 30 seconds. I need to work to make this more intuitive. Jack Ganssle recommended a small window of the entire buffer that shows your current view location, like with many oscilloscopes.
www.ganssle.com/tem/tem368.html#article4