[Bart Schroder] is busy designing a high-voltage variable speed motor driver, and regrets that a standard oscilloscope cannot visualize the waveform around the switching transistor. This is because the three-phase characteristics of this type of motor are driven by three current waveforms, 120 degrees out of phase with each other, where the current flows between each pair of winding taps without reference to the common concept of grounding. However, the average oscilloscope on your workbench must be ground-referenced, so visualizing such waveforms is a bit cumbersome. Then there is the fact that the motor runs at hundreds of volts, and the prospect of using your precious bench-top instrument for detection is a bit nerve-wracking to say the least. The solution to the problem is obvious. Build your own isolated high-voltage oscilloscope. Here is the development journey of Cleverscope CS448, which will bring you a pleasant viewing experience.
The oscilloscope itself is in terms of specifications, nothing is too flashy, it is the isolated channels that make it unique. However, it does have some advantages, such as 8 additional 100 Mbps digital inputs and a convenient 65 MHz signal generator. In addition, don't reach out for your wallet for the time being, because this is a special tool, and its potential user base is smaller than that of ordinary oscilloscopes, so these devices are quite expensive. Having said that, the focus here is not the existence of scope, but the process from problem to solution that we are most interested in. Many things can be learned from [Bart]’s journey, for example, where should the front-end ADC be placed? Isolated side or not? The noise floor of the signal chain determines the former.
The real trouble in building a multi-channel isolation instrument of this nature is the isolation from the front end to the public back end. The front end needs power, so this is fed forward through a beautifully designed two-wire wound toroidal transformer, which is a complete story in itself. According to JESD204B, data is taken out of the ADC through a pair of 4.375 Gbps compressed serial channels, using a customized fiber link module. [Bart] A lot of issues were explained in the process, just to show that building something like this is not always the first way you try. Really smart (range) things
Do I still need a high current probe? We have solved the problem for you. Since we are talking about isolation, how to use an isolated USB serial board to protect the safety of your laptop?
Would love to see the 4-channel design that I can actually buy.
Just bought one for work from Tektronix. All four channels are isolated. It is very suitable for working on both sides of a flyback power supply.
TPS2014B 100mhz and TPS2024B 200mhz
I was very disappointed when I learned that common inexpensive oscilloscopes do not isolate ground on their inputs. This sounds great!
When I found out, I was disappointed and had to replace some circuit components!
@Greg A said: "I was very disappointed when I learned that common cheap oscilloscopes do not isolate and ground their inputs."
Unless otherwise specified in the user manual, always assume that your oscilloscope channel ground is common at the instrument input connector, regardless of the cost of the oscilloscope. Isolated grounds are not standard, and obtaining them usually means that an option must be specified. The only reason there are separate ground connections at the end of each probe is to maintain signal integrity. These grounds are usually connected together at the probe connector. If you need a floating or differential input, use two channels, invert one and add the other. I have never seen a useful oscilloscope without a standard built-in dual-channel inversion and addition function. Even if the oscilloscope has an isolated ground function (that is, differential or isolated input), these channels will be able to use a common ground (that is, single-ended mode), and will also provide inversion and addition functions.
This really makes me wonder how Tektronix solved the isolation problem in its THS720 handheld oscilloscope in the 1990s. I used one in my work a long time ago to observe the waveform of a motor, which has a low-side switch armature and an excitation coil switched by an H-bridge. When the controller switches the direction of the motor, it is really cool to see that the voltage of the H bridge just flips, because knowing that a normal oscilloscope will malfunction at that time.
https://w140.com/tekwiki/wiki/THS720
"When you encounter an oscilloscope that shows "calibration error, DC balance" fails to calibrate, and there is a continuous DC offset that increases with the attenuator setting, please try to replace U45 and U46. They are HP CNR201A analog optical isolators, now Manufactured by Avago as HCNR201-300E. They send a low-pass filtered version of the input signal through an isolation barrier to provide a DC level that cannot cross the isolation barrier of a high-frequency transformer."
1. Transformer used for isolating AC high-frequency materials. 2. Opto-isolator used for DC offset (probably a lower frequency).
TEK THS720 oscilloscopes appear on the second-hand market regularly. They still go from $300 to $500. Compared with modern DSOs, they are not functional enough, but their portability and isolation make them still very useful.
I have been watching these on eBay recently. There are a few things to be aware of. One is that LCD monitors like to break. The contrast is so bad that you can't see the image. Obviously, the front polarizer was broken for some reason. The cost of replacing the LCD is usually about $150. Or, you can pull out the LCD, take it apart, peel off the polarizer, clean the glass, and then attach a new polarizer. You can buy it online for about $5. For people who read Hackaday and use an oscilloscope, it may be feasible.
Another thing to note is that they come with original oscilloscope probes. They are special probes and there is no exposed metal on the BNC connector. Since the oscilloscope is not connected to a safety ground, they do not want your oscilloscope probe ground to be connected to any bare metal that may be in a high-voltage state. I often see these oscilloscopes sold on eBay with other random probes with metal BNC connectors. As long as you don't connect the ground to high voltage and touch the BNC connector, they should work properly and won't shock you. I personally don't want to believe that I remember not to touch the connector when using it, so I want the original probe.
An oscilloscope with a good LCD, original safety probe, battery, charger base, and wall wart will cost $500 or more.
If you can still see through the polarizer (that is, you can see the screen) but it does not work like a polarizer, you can try placing another polarizer in front of it. Not an ideal solution, but it will be very challenging to remove the adhesive on the LCD without damaging the LCD.
The TEK222 oscilloscope also has two isolation channels. Although the oscilloscope is so old and so limited, when I need to view two signals that need to be isolated at the same time, I still use my oscilloscope—for example, the sample PWM motor driver in the linked article.
Sadly, they are too slow (10MHz) to cut with modern electronic equipment.
I am considering using a smart oscilloscope to satisfy my desire for isolation, but I cannot hack the USB oscilloscope. If it is integrated into a single instrument, I will buy one on the spot!
Yes, you can buy a standalone oscilloscope (Fluke, etc.), but I *hate* you should use the knob button. The closest I have so far is the MicSig STO1000 series... knob, battery powered (it can float), but all channels have one thing in common. They do have a trigger, so they might be daisy-chained to the second and third units...
Picosope has a series of independent oscilloscopes!
Unfortunately, the Picoscope product is a USB oscilloscope...portability goes beyond the window-now you need to carry not only the oscilloscope, but also the laptop. Use with bench tops-I want less clutter on my bench, not more!
There is a limit to the usefulness of high-voltage isolated oscilloscope inputs: most oscilloscope probes only specify approximately 150 V CAT I isolation between their ground and the user's hand. This applies to better connectors with insulated BNC connectors.
But of course 150V is sufficient in many cases.
Isn't Fluke Scopemeters not enough?
Please note that I have seen on the Internet that not all Fluke oscilloscope channels are isolated from each other. Some do, you have to check the model. Any battery-powered oscilloscope allows you to not be grounded through the power cord, but if you want to perform the tricks I mentioned elsewhere in these comments, I connected a probe and the two channels also need to be isolated from each other. TEK handheld oscilloscope Connect to the motor on the H bridge, and use another channel to monitor other ground reference signals on the same board.
Why not use two channels in diff mode and leave the ground unconnected? The smaller the distance between the signal source and the oscilloscope, the better for me.
Many reasons (far from exhaustive-exactly what I have dealt with in recent memory. Read the linked article for a better introduction): You still need to provide a ground reference for each input, when you have differential signals floating When this is difficult for the same reason, hundreds of volts of resolution lose additional noise input impedance (lumped impedance and capacitance)
Oversimplified example: Imagine a 5V differential signal floating at 800V in the 1000V range (100V/div on many oscilloscopes). If the input resolution is 10 bits (easy count), the input is 1.0V/count, providing a range of 5 counts for the required signal. The isolated input allows the use of 10V range and provides 10mV/count, which is more useful. At the same time, there are advantages in capacitance and input impedance.
Although I really admire the determination and expertise of the original publisher, I may continue to use my differential probe to deal with any ground loop-prone problems.
Fluke Scopemeter with relevant probe (600V CAT III rating) can solve the problem. I used the old 199c to measure the switch mode power supply and the results were very good. BR Peter
This is what I said! A portable oscilloscope like this does not have a normal ground connection available for use.
Well, regarding the initial task, isn't this a classic high-tech sledgehammer cracking nuts? Or, since this is Hackaday, can't we use much lower-tech hardware to get something satisfying?
First of all, for motors, we are definitely talking about relatively slow PWM signals on the order of 10 kHz – 1 Mhz. Any higher value, you will drive the H-bridge as an analog amplifier. The simple way to measure the current through the H-bridge Fets is to measure the temperature, which can be easily done using the relevant temperature IC, and then perform some mathematical operations
Measuring the high-side gate drive voltage is the most difficult part, but it will certainly not be complicated because it is a low-Z measurement, because any useful FET will have a large input capacitance, so the gate drive will be a few amperes and the order of 10 Voltage at <1 MHz. For this task, you can use analog mode fast optoisolators connected to the drive as you do in many isolated PSUs. Connect the output to a fairly competent A2D converter and complete the work. A more fancy arrangement might use a digital isolator, such as the Si8640, as the original ADC driven by the floating power supply of the high-side driver.
Is it really a problem to measure Vout1 and Vout2? Both are low Z values and can be measured with reference to the ground. The current sensing resistance is almost irrelevant, but you can subtract their values if you want!
The z value of the current sense resistor is also low, so it can be measured with reference to the ground.
Cleverscope is not a hobby project. It is a commercial product derived from specific needs.
For certification testing, the output must be measured directly, rather than using mathematical inference, even if the mathematics is credible and the components are well understood.
At this price, compared with Tektronix TPS2024B, this is a tough sale. Tektronix TPS2024B is an independent oscilloscope with higher bandwidth, but only 8-bit resolution, and the price is only 1/4 of that...
I am also confused as to why they choose a two-wire wound toroidal transformer. This greatly reduces the leakage inductance, but increases the coupling capacitance; I would rather use a separate winding to minimize the coupling capacitance (and the capacitive load on the circuit reference point), and only design the power supply to handle the leakage inductance (which is not at low power) As painful as under high power) power).
> "If Zload is inductive-like most motors-when a voltage is applied, the current will slowly increase or decrease."
Wait, what? What is a non-inductive motor?
Does anyone have the part numbers of these optical transceivers? I can't find it anywhere.
https://corporate.murata.com/en-eu/more_murata/techmag/metamorphosis20/productsmarket/module has been terminated
Compared with using high-voltage differential probes on your existing oscilloscope, what are the advantages?
"The second method is to use a differential probe-but even a good probe has a poor common-mode rejection ratio (CMRR) at high frequencies. For example, the Tektronix P5200A has a CMRR of 30 dB at 3.2 MHz. Now 3.2 MHz The rise time is tr = 1/ f = 100ns. Many modern switches are faster than this. If you have a 680V bus (typical for three-phase = 2 x √2 x 240V), the probe will generate 680V equal to -30dB = 0.032 x 680 = 21.8V spurious response. This almost obliterates anything interesting. Why measure the gate drive you mentioned? After blowing up a lot of IGBT modules, I can tell you that the gate drive must be correct."
Please be kind and respectful to help make the comment section great. (Comment Policy)
This website uses Akismet to reduce spam. Learn how to handle your comment data.
By using our website and services, you explicitly agree to the placement of our performance, functionality and advertising cookies. learn more