Jun 23 19:08
Indeed, those tiny MOSFETs are going to be dissipating about a watt each @ 20 A. If you haven't made arrangements to remove that heat, they're going to burn up.
 

 Electrical Engineering

A place to talk with friends from the EE community about vacuu...
Mar 26 20:12
After actually rereading that answer, the other aspect is the ability to pull a DSSS signal from below the thermal noise floor, which is simply impossible for FHSS. So in that sense, DSSS does have a definite advantage even with white nose. "Coding gain" doesn't exist for FHSS -- a signal that's below the noise floor will still be below the noise floor after despreading, so there's no way to find it and lock onto it.
Mar 26 18:39
I was just trying to condense all of that down to a succinct sentence.
Mar 26 18:39
@NickAlexeev: Yes, it's true that DSSS has no specific advantage WRT white noise. I was just speaking generally in terms of the fact that it spreads individual payload symbols over a wide bandwidth, which makes it more robust to most forms of noise and interference. In contrast, a single FHSS hop can be taken out by a narrowband interferer, which requires more elaborate burst error correction to get the same level of overall reliability.
Jan 2 14:23
But yes, in general, it's a good idea to keep circuits compact and connections short in order to minimize external influences.
Jan 2 14:22
I'm not sure where you're going with this. That question was about building a filter with differential input and output, and my answer was a way to convert the single-ended output of the existing filter back to differential.
Dec 28, 2024 13:08
Without knowing the details of the filters, it's hard to say. My guess would be that one path is optimized for voice recording (8 kHz sample rate, 3500 Hz bandwidth) and the other is optimized for music (32+ kHz sample rate, 15+ kHz bandwidth).
 
Mar 13 12:09
Yes, but why would Vcc change sign?
Mar 12 20:31
That isn't how "chat" works. You post something, then when I have time, I post a response. Since I needed to draw a schematic, I've written an answer to your original question.
Mar 12 15:03
Hello. I assume this has something to do with the power transistor question.
 
Dec 18, 2024 17:45
Oh, sure! I work with image sensors and high-definition video signals, and I've used both DDR I/O and SERDES I/O to handle high-speed signals. But each new interface requires significant engineering to implement.
Dec 18, 2024 17:19
The point is, the FPGA's physical hardware is capable of doing what you want, but you're not going to find existing IP to do it for you. It's going to involve generating an internal 800 MHz clock from the MCU's memory clock, phase-aligning all of the signals to that, and then also doing some sort of "passive" timing calibration on the individual lanes of data. And all of this assumes that you haven't already totally destroyed the signal integrity on the memory bus by adding your stubs.
 
Apr 10, 2024 13:19
I pulled the code from your comment, but the StackExchange Markdown formatter made changes that make it difficult to understand. Please edit the question yourself and paste the actual code into the area I created. In any case, loading, incrementing and storing a 64-bit variable on a 32-bit CPU are non-trivial operations. Again, look at the generated code to see what's actually going on.
Apr 10, 2024 13:19
Show us the code of your while (1) { ... } loop. Better still, look at the generated code and count the number of instructions executed in one iteration. A 200 MHz CPU is not going to execute anywhere close to 200 million iterations per second. If you're using something like digitalWrite() to toggle a pin, there's a lot of overhead in that kind of function call, and your numbers sound about right regarding the number of loop iterations per second.
 
Jan 4, 2024 20:11
I understand the premise, but not the motivation for it. What advantages does it have? What about drawbacks, such as the need for the balls to be transparent (i.e., not steel)? What is the user experience like -- for example, how does the ball signal that it is closer to the jack than the other balls?
Jan 4, 2024 20:11
Probably dreaming. I would think that an optical ball-tracking system based on a few cameras placed around the arena would be a better approach. Such systems are commonly used these days in the TV coverage of baseball, tennis, golf, etc. and they are surprisingly precise.
 
Dec 31, 2023 15:51
But note that the clearance on the PCB, between the pads, is much less than the clearance between the pins on the transistor case. You really should spread those pins and pads out. The high-voltage MOSFETs that I've used in the past come in much larger packages with wider pin spacing to begin with. Given that you have the transistor standing perpendicular to the board, the most direct solution would be to bend the drain pin "up" (in the current view) and away from the other two.
Dec 31, 2023 15:51
Also, looking at your layout, the creepage distances between high voltage and low voltage traces don't look like anywhere near enough to my eye. Especially between the gate and drain of Q1, but other places as well.
Dec 31, 2023 15:51
I would start looking at voltage stress on Q1. Universal motors are noisy. Have you tried putting an R-C snubber across the motor (in addition to D2)?
Dec 31, 2023 15:51
But you're rectifying the mains in your circuit, which means that you lose the benefit of any reactive component in the motor. But you said that you have tried running at 100% PWM while building the circuit, so I guess that's OK.
Dec 31, 2023 15:51
Is that how it works if you manually short drain to source? It doesn't blow the fuse? Because of its high internal resistance of 3.88 °C/W junction-to-case, note that Q1 can only dissipate an absolute maximum of 32 W even with a perfect heatsink.
Dec 31, 2023 15:51
If you didn't use thermal grease on both sides of the mica, then I'll bet that you've got way too much thermal resistance, and that the root cause is Q1 overheating.
Dec 31, 2023 15:51
Note that the highest stress on Q1 occurs when running at partial power, not at full power. Usually the switching losses are significantly higher than the I^2*R losses. Do you have a good heat sink on it? I'm wondering whether the value of R1 should be reduced a bit. I didn't look up the gate charge of Q1, but if it's slowing down the switching transitions too much, it vastly increases the switching losses in the transistor.
 
Oct 28, 2023 00:14
No demand for them?
 
Apr 10, 2023 02:16
@BlueRaja-DannyPflughoeft: A metal basically is a dense cloud of electrons, held together by the lattice of positive nuclei. (I can't think of any other way to create such a cloud.) The reason a metal reflects light is that it contains a sufficient number/density of electrons that are free to move about, and it these electrons that interact with the EM field of the light in a way that leads to reflection.
 
Jan 21, 2023 17:25
With that much current available, you basically have an arc welder. Can you think of any way to stop an arc welder other than separating the electrodes or removing the power source?
 
Jan 21, 2023 06:20
base64 would only require 24 characters, and base85 would only require 20 characters. Both are designed to transfer arbitrary binary data over text-only links.
Jan 21, 2023 06:20
But why does a UID require a decimal representation in the first place? Why not use hexadecimal, like we commonly do with other large numbers such as encryption keys, hash values, CRCs, etc.?
Jan 21, 2023 06:20
Sounds like a good application for double dabble. It can convert binary numbers of arbitrary length to decimal while working on as little as 4 bits at a time.
 
Nov 7, 2022 12:56
Well, yes, but as others have pointed out, in an embedded system, you also need to be concerned with the state of the external hardware, which might have changed while the power was off. You need to work out a way to re-synchronize the state of the software with that.
Nov 7, 2022 12:56
What a laptop does is called "hibernation", and it takes a lot of behind-the-scenes work to do it successfully. After detecting imminent battery failure, it saves the state of the CPU to RAM, and then it saves the entire contents of RAM to the hard drive. The entire process is then reversed on reboot.
 
Jul 2, 2022 08:44
I have always assumed that the name "assembler" had something to do with putting together the information from multiple text fields in the line of source code to produce the corresponding numeric fields in the opcode and operands. But I have no references to cite for this -- just ~50 years of experience reading and writing assembly language, and producing a few assemblers and disassemblers along the way.
 
Oct 14, 2021 05:14
Wait, you're counting on the interior of the car being hot in order to power the system that will cool it? I can't even begin to imagine how that's supposed to work. Why not just use solar panels to run a small fan, like everybody else?
 
Oct 11, 2021 16:32
Direction finding is typically done with a circular array of antennas that are scanned at a fairly high rate. Comparing the phase of the PM signal created by this with the phase of the switching tells you the direction the signal is coming from.
Oct 11, 2021 16:27
The individual antennas are amplitude-modulated. This generates sidebands, and the sidebands from adjacent antennas combine in a way that looks exactly like the sidebands created by phase modulation.
Oct 11, 2021 16:13
To the OP: Everything you say is true, but you have neglected to ask an actual question. You're going to have to establish what is acceptable performance for your system and that will then tell you the acceptable value of d. Perhaps try a few examples in simulation.
Oct 11, 2021 16:13
@DKNguyen: The "new frequencies" are generated during the crossover from one antenna to the next, when you're effectively summing two signals that have different phases and varying amplitudes.
Oct 11, 2021 16:13
@DKNguyen: Hence the term pseudo-Doppler. It does work, and can be used for receiving, too. Some VHF direction-finding equipment does it this way. I assume you understand the relationship between frequency modulation and phase modulation.
 
Aug 24, 2021 16:21
It's quite a bit more. I didn't expect that. Still not sure I can explain it.
Aug 24, 2021 16:17
OK, fair enough. What is the actual base current of Q8 vs. Q7?
Aug 24, 2021 16:17
Nothing. The same fixed current flows from base to emitter. Physically, some electrons from the emitter may end up in the collector region, but they eventually migrate back to the base region as well. The net flow between base and collector across that depletion region is zero.
Aug 24, 2021 16:17
@jonk: Are you sure? I'm trying to think of any mechanism by which the Vbe for a transistor with a "failed" load would go down instead of going up or just staying the same, and I can't come up with anything. What is your reasoning?
Aug 24, 2021 16:17
Sure, that's usually the point. Each arm of the current mirror is connected to a different load. The mirror can't tell what's on the other side of that load, as long as it has the correct polarity with respect to the mirror's reference node.
 
Aug 22, 2021 12:44
Andy is correct. A flyback transformer generates high voltage by storing energy in its magnetic field during a relatively long "charging" phase and then releasing it quickly during a "discharge" phase. If you draw energy out of the secondary during the charging phase by full-wave rectifying it, the charging of the magnetic field never occurs. If you are instead using something like a neon sign transformer, then you shouldn't have used the the term "flyback" in your question.
 
Aug 16, 2021 14:31
A switched capacitor can function as a variable resistor whose conductance is directly proportional to the switching frequency. This technology is well-known in analog filtering circles.
 
May 28, 2021 04:59
What is controlling TB? TB must go low a definite amount of time before UPD clocks the output latch, in order to guarantee that the inputs to the latch are not changing at that time. With an asynchronous (external) clock, controlling TB is the only way to make that guarantee.
 
May 26, 2021 13:31
I've added it to my answer.
May 26, 2021 13:22
I have a circuit based on a 555 timer that is guaranteed not to oscillate. I'm sure I've got it in an old answer somewhere here on EE.SE, but I can't seem to locate it.
May 26, 2021 13:14
The efficiency of what, exactly?
May 26, 2021 13:04
Well, all I can say is that the symbol you use in your schematics explicitly shows the body diode, so you just need to learn to pay attention to those kind of details.