|
Ehhh, acronym clarification time...  UWB - Ultra Wide Band? DBS - To me, this stands for DataBase System (MySQL, Oracle, Access) LNB - Large, Narrow Band?
|
#
?
Dec 5, 2010 16:40
|
|
|
|
| # ? Jun 11, 2024 03:45 |
|
|
Krenzo posted:I'm trying to do time of flight stuff: send out a single pulse from a transmitter and time how long it takes to reach the receiver. I have several FPGAs and have built a test circuit that can time the travel time of a square pulse through a wire to within about 100 picoseconds (I'm planning to reduce the noise to get a more accurate timer). The next part is to construct a UWB transmitter and receiver to replace the wire that the pulse travels through. However, the difficult part I'm having to deal with is figuring out how I'm going to troubleshoot any problems I may encounter with such narrow pulses (ones that will register in the >3 ghz range). To even be able to see the pulse with an oscilloscope would take a very expensive one. I'm currently trying to ask electrical engineer grad students at my school to figure out if the university has any fast scopes I could get some time with. I hate the idea of getting into this project and not have any concrete feedback as to what's going on, ie if things are really operating how I planned them to. Scopes that can do reasonable time domain measurements on ~100 ps wide events will be extremely expensive - even probing signals that fast will be challenging. This also won't translate very well to RF - you would need a very, very powerful pulse in order to transmit something that narrow any appreciable distance (like across a room). Even a 1 megawatt transmitter would only give 100 uJ of energy at 100 ps wide. Any reason you don't want to use pulse compression techniques, like in radar? What type of range resolutions are you trying to measure?
|
|
#
?
Dec 5, 2010 17:35
|
|
|
ValhallaSmith posted:Why do you need UWB? You don't have to worry about reflected waves interfering with the signal. The first pulse to arrive is the original pulse if there's a line of sight to the transmitter. With a continuous wave, not only do you have to deal with self interference, but you also have to deal with the ambiguous nature of the signal when using correlation and have to add some modulation to serve as a marker. The Scientist posted:UWB - Ultra Wide Band? http://en.wikipedia.org/wiki/Ultra_wideband http://en.wikipedia.org/wiki/Low-noise_block_converter SnoPuppy posted:Scopes that can do reasonable time domain measurements on ~100 ps wide events will be extremely expensive - even probing signals that fast will be challenging. Actually, I was planning on using 40 ps wide pulses.  I'm planning on range being at a max of 15 feet. I think the attractiveness of UWB is how precise you can get the time estimate. I haven't researched a whole lot with pulse compression, but it would appear that you couldn't achieve the same level of accuracy.
|
|
#
?
Dec 5, 2010 20:45
|
|
|
crabrock posted:Problem 1: Semiconductors work with dynamic resistance. This includes diodes, LEDs, base-emitter junctions of BJT transistors, etc. That means that you will never actually calculate the resistance of the junction, you just assume that it has a 0.7V or whatever drop across it. What you're asking exactly is kind of ambiguous, and you didn't answer the other guy's question that I saw. Do you want an entire circuit to only work when the input is above ten volts, or do you want a circuit to get whatever voltage is above the ten volts? If it's the former case, use an op-amp in a comparator configuration. For the former, use a zener diode. Zeners work like regular diodes when they're forward biased. When they're reversed biased, they work kind of the same way, except the forward voltage is going to be 8 or 10 or 24 volts or whatever it says on the datasheet for that particular diode. crabrock posted:Problem 2: Lower resistance allows higher current to flow, which would make the capacitor leak current faster. I don't know the kind of configuration you have this in, so I can't really say any more. crabrock posted:Problem 2.5:
|
|
#
?
Dec 5, 2010 23:30
|
|
|
Is there any disadvantage to using an oversized inductor in a buck converter? Some of what I've read seems to suggest the inductor value is a minimum rather than something that needs to be matched.
|
|
#
?
Dec 6, 2010 01:56
|
|
|
BattleMaster posted:Is there any disadvantage to using an oversized inductor in a buck converter? Some of what I've read seems to suggest the inductor value is a minimum rather than something that needs to be matched. You don't want saturation to occur. Then your inductor behaves like a piece of wire, and bang goes the driver. There shouldn't be any problem with a larger inductance until you get into current handling issues.
|
|
#
?
Dec 6, 2010 02:56
|
|
|
Krenzo posted:You don't have to worry about reflected waves interfering with the signal. The first pulse to arrive is the original pulse if there's a line of sight to the transmitter. With a continuous wave, not only do you have to deal with self interference, but you also have to deal with the ambiguous nature of the signal when using correlation and have to add some modulation to serve as a marker. So are you planning on using two receivers and calculating time difference of arrival? Not to pee on your parade, but I'm not sure it will be possible or practical to send a 40ps wide pulse over the air. There's a reason why all modern radar systems use some form of pulse compression. It's very difficult to send a narrow pulse, and hard to tell when a pulse starts and stops after you send it through space. You said you had it working with wires - how well does it perform when you inject noise? What about phase/gain distortion due to channel, TX, or RX effects? I'd try to mock it up/simulate as much as possible before trying to build any RF system - you want to prove it is workable on paper before you try to build it.
|
|
#
?
Dec 6, 2010 18:40
|
|
|
I'm looking for an entry level oscilloscope (<$100) on craigslist. Can someone give a rundown of brands? BK Precision, Kikusui? 60MHz, 35MHz? I know nothing about oscilloscopes. Why do I need one? I probably don't, but I want one and am going to learn how to use it.
|
|
#
?
Dec 6, 2010 19:36
|
|
|
Scopes are for capturing voltage levels at any particular point in time. If you don't know wha they are, you probably don't really need one. I'd buy an Arduino and find some code online to turn the ADC into a scope
|
|
#
?
Dec 6, 2010 19:51
|
|
|
I'm trying to analyse a LCR circuit, where the resistor is in parallel with the capacitor. I've solved it for the initial conditions(non trivial). I'm trying to simplify it but not really getting anywhere. There are probably 3 conditions, damped, under and critically damped regions, that I'm going to have to simplify for. I've looked online but can't really find anything that's much help. It seems like if I was working with AC it would be fine and I could just use impedances but I'm kind of stuck using DC. I have herd that I could just assume it was for AC and then use an infinite sum of sins to get DC. Is that a reasonable method? I'm just looking at the transient response, the time period would be on the order of the frequency.
|
|
#
?
Dec 6, 2010 21:00
|
|
|
Unparagoned posted:I'm trying to analyse a LCR circuit, where the resistor is in parallel with the capacitor. I've solved it for the initial conditions(non trivial). I'm trying to simplify it but not really getting anywhere. There are probably 3 conditions, damped, under and critically damped regions, that I'm going to have to simplify for. If you want the DC operating point, just treat capacitors as opens and inductors as shorts. If you're trying to get the step response, I'd figure out the frequency response (laplace) of the system and multiply that by the frequency response of a step function and then inverse it.
|
|
#
?
Dec 6, 2010 21:21
|
|
|
Thanks to whomever tried to help me, I made a little bit of progress but I'm still not really sure how to do what I want to do. I thought maybe some pictures would help. I've been using one of the little simulators that was linked in here. Basically I'm trying to accomplish these things: 1. Source fills up capacitor. The voltage/current is variable, sometimes being low, sometimes being high, and sometimes being off. 2. Basically I want it to add up til a certain level (that Z diode thingy) and then discharge. 3. When it gets to that level and discharges it goes through the other side of the circuit and lights up the LED for a brief time until it falls below the required value again and has to charge up. 4. I'd like it if the capacitor source was "off" the charge leaked from the capacitor but not through the LED. I hope this makes a little more sense. I'm still pretty confused on how some of this works. I feel like I've read the same intros to electronics a billion times and they're not really helping.
|
|
#
?
Dec 6, 2010 22:08
|
|
|
crabrock. What you want to do sounds fairly similar to a 555 circuit. Software questions: What's the best software to use. I like Falstad Circuit Simulator but it's slow and and very limited. From what I can gather is I want is some kind of spice but with a decent interface. I think I want spices power with a Falstad circuit simulator interface. Any advice
|
|
#
?
Dec 6, 2010 22:20
|
|
|
I thought you guys might find this interesting. Here's a true look inside an XBox 360. I took a 360 Mother Board to a place I do business with and they were kind enough to let me Xray it and send me the photos. You can get an idea of what you're looking at by the little window on the left of these screenshots. Here's the GPU:  Close up on the BGA balls:  The CPU:  RAM:  Close up on Wire Bonding:  Zoomed out on that same chip:  You may notice some of the traces look really squiggly. This is most likely to match trace lengths for speed sensitive nets.
|
|
#
?
Dec 7, 2010 00:06
|
|
|
crabrock posted:Thanks to whomever tried to help me, I made a little bit of progress but I'm still not really sure how to do what I want to do. I thought maybe some pictures would help. That's pretty much the design of various BEAM robotics circuits (called "relaxation oscillators). Google those terms and you should get lots of good info. The one thing about those circuits is that they normally trigger at much less than 10v. There is a Zener based one which can be used with a 10v Zener but apparently this particular circuit is kinda finicky. The ones based on the 1381 voltage detector are much more stable but the 1381 is only available in variants that trigger at about 5v. If you use the tc-54, the European equivalent, you can get trigger voltages up to 7.7. Will that work or do you need 10v specifically? Circuit here: http://www.solarbotics.net/library/circuits/se_t1_mse.html
|
|
#
?
Dec 7, 2010 00:08
|
|
|
Hillridge posted:I thought you guys might find this interesting. And you were correct! thanks. quote:1. Source fills up capacitor. The voltage/current is variable, sometimes being low, sometimes being high, and sometimes being off. http://www.solarbotics.net/library/circuits/se_t1_1381.html EDIT: gently caress. Mill Town has a faster posting finger than I!
|
|
#
?
Dec 7, 2010 00:16
|
|
|
I don't need it to run at 10V, i just pulled that number out of the air. I will look into this BEAM stuff, thank you! Unparagoned posted:crabrock. What you want to do sounds fairly similar to a 555 circuit. Not quite what I wanted but it does answer one of my future questions so congratulations on predicting the future.
|
|
#
?
Dec 7, 2010 00:29
|
|
|
Unparagoned posted:Software questions: What's the best software to use. I like Falstad Circuit Simulator but it's slow and and very limited. From what I can gather is I want is some kind of spice but with a decent interface. I think I want spices power with a Falstad circuit simulator interface. Any advice LTSpice has a pretty reasonable interface, is very powerful, and is free. http://www.linear.com/designtools/software/
|
|
#
?
Dec 7, 2010 01:51
|
|
|
standardtoaster posted:I'm looking for an entry level oscilloscope (<$100) on craigslist. Can someone give a rundown of brands? BK Precision, Kikusui? 60MHz, 35MHz? I know nothing about oscilloscopes. Why do I need one? I probably don't, but I want one and am going to learn how to use it. It's useful for seeing how voltage changes with respect to time. With a voltmeter, you only know what the voltage is right now. With an oscilloscope, you can see what voltage is over time like if you have constantly changing voltage source. You can hook it up to a capacitor and see how long it takes for the capacitor to charge and discharge for example. It's handy. I use a USB scope (http://www.linkinstruments.com/mso19.htm), and they're cheaper because they don't have to include a display and other things that your computer can handle.
|
|
#
?
Dec 7, 2010 02:38
|
|
|
SnoPuppy posted:So are you planning on using two receivers and calculating time difference of arrival? No, I will send the pulse at regular intervals that the receiver knows when they were sent. It will compare the time it knows it was sent to the time it was received. Eventually, I will have multiple receivers in order to construct 3d position based on their individual distances from the transmitter. SnoPuppy posted:Not to pee on your parade, but I'm not sure it will be possible or practical to send a 40ps wide pulse over the air. There's a reason why all modern radar systems use some form of pulse compression. It's very difficult to send a narrow pulse, and hard to tell when a pulse starts and stops after you send it through space. I'm going off of research papers and books that are written on the subject. They have actual products out now that work on this technology. I would probably have given up by now if there were no papers and examples for me to work from. Here's a good thesis paper I found on the subject: http://scholar.lib.vt.edu/theses/available/etd-09192006-230733/unrestricted/Anderson_Final_ETD_Version.pdf SnoPuppy posted:You said you had it working with wires - how well does it perform when you inject noise? What about phase/gain distortion due to channel, TX, or RX effects? Well, it's digital in principal so noise should not have a large effect. The detection circuit isn't tripped until the pulse reaches the specified voltage level (ie digital logic high in my test circuit). I'm planning on using a variable comparator to detect the leading edge of the pulse and having to modify its trip voltage level based on amplitude so that it trips at the same fraction of the pulse's height. Interference and noise from other signal sources may be a problem, but I'm hoping that will be limited in the >3 GHz band. If it's not, I will add some bandstop filters in any interfering bands. If there is significant noise, the research papers talk about using a match filter to find the pulse, but I hate the thought of trying to make a match filter in analog circuitry, ugh.
|
|
#
?
Dec 7, 2010 02:56
|
|
|
Krenzo posted:No, I will send the pulse at regular intervals that the receiver knows when they were sent. It will compare the time it knows it was sent to the time it was received. Eventually, I will have multiple receivers in order to construct 3d position based on their individual distances from the transmitter. Under what circumstances would the signal not reach the receiver according to C (speed of light)?
|
|
#
?
Dec 7, 2010 07:04
|
|
|
crabrock posted:Thanks to whomever tried to help me, I made a little bit of progress but I'm still not really sure how to do what I want to do. I thought maybe some pictures would help. Is this just to do it or do you have an application in mind? This is basically a switched capacitor LED driver. There are versions that flash LEDs at various speeds as well.
|
|
#
?
Dec 7, 2010 09:21
|
|
|
randyest posted:I work for NEC Electronics (now Renesas Electronics) in a different division (semiconductor design,) but I have pretty extensive intranet access and some decent connections with the display group (which, as you might guess, has pretty much completely transitioned to LCD design and support these days.) Thank you so much, unfortunately I doubt any of the people I talked to actually tried to find it though. There are some other people here in Norway who might have worked on monitors, but it's out of my hands now. At this point I doubt we'll be able to use the monitors so we're mostly looking for a replacement that includes complex analogue circuits, switch mode power supply etc. like a CRT does. The only appropriate devices we have are a few audio amplifiers and CD players, but they're simple enough that anyone actually interested in fault finding will know them like the back of their hand in a few weeks. Any ideas there would be greatly appreciated standardtoaster posted:I'm looking for an entry level oscilloscope (<$100) on craigslist. Can someone give a rundown of brands? BK Precision, Kikusui? 60MHz, 35MHz? I know nothing about oscilloscopes. Why do I need one? I probably don't, but I want one and am going to learn how to use it. Oscilloscopes are fun to play with, I would recommend finding a 20+ MHz analogue CRT scope, they should be plentiful on the used market and far more robust than any DIY digital solution. Over here Leader is commonly found used, but Tektronix, Agilent/HP and Philips-Fluke are likely better. I'm pretty pleased with my HP 1740a 100 MHz but it's not exactly compact and uses some custom ICs for important functions that will be difficult to replace. I used a PM3350 in school, that was good too but probably still expensive. longview fucked around with this message at 13:24 on Dec 8, 2010 |
|
#
?
Dec 7, 2010 18:19
|
|
|
The Scientist posted:Under what circumstances would the signal not reach the receiver according to C (speed of light)? Umm, if the receiver is too far away or something is blocking the transmitter? I don't know what the speed of light has to do with it. I must not understand your question.
|
|
#
?
Dec 8, 2010 01:31
|
|
|
Krenzo posted:Umm, if the receiver is too far away or something is blocking the transmitter? I don't know what the speed of light has to do with it. I must not understand your question. He's wondering why you are measuring the time-of-flight when you know the signals are going at the speed of light, more or less. I think he's missing that you are working up to having multiple receivers so you can do triangulation.
|
|
#
?
Dec 8, 2010 02:34
|
|
|
ohhhhhHHHHHHHHHhhhhhhhh.
|
|
#
?
Dec 8, 2010 03:30
|
|
|
So I want to get 12V from a D-cell battery. LT1073 (http://cds.linear.com/docs/Datasheet/1073fa.pdf) looks great, but no one stocks it. Is there something else that makes more sense to use? I need to pull about 100mA edit: is there a relevant IRC channel? vvv cool man thanks, I think one of those will work Fish Ladder Theory fucked around with this message at 03:47 on Dec 9, 2010 |
|
#
?
Dec 9, 2010 03:12
|
|
|
Will this work? It require 2V but it is in stock! http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=LT1107CS8%23PBF-ND EDIT: Here is a 1.8V one http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=LT1301CS8%23PBF-ND EDIT 2: Here is a 1V min that puts out 12V and is in stock http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=LT1110CN8-12%23PBF-ND AnomalousBoners fucked around with this message at 03:33 on Dec 9, 2010 |
|
#
?
Dec 9, 2010 03:25
|
|
|
Unparagoned posted:Software questions: What's the best software to use. I like Falstad Circuit Simulator but it's slow and and very limited. From what I can gather is I want is some kind of spice but with a decent interface. I think I want spices power with a Falstad circuit simulator interface. Any advice Why everyone isn't using MicroCap is beyond me. Krenzo posted:I'm going off of research papers and books that are written on the subject. They have actual products out now that work on this technology. I would probably have given up by now if there were no papers and examples for me to work from. Here's a good thesis paper I found on the subject: http://scholar.lib.vt.edu/theses/available/etd-09192006-230733/unrestricted/Anderson_Final_ETD_Version.pdf Also, this is really neat! Cyril Sneer fucked around with this message at 04:55 on Dec 9, 2010 |
|
#
?
Dec 9, 2010 04:37
|
|
|
Cyril Sneer posted:Why everyone isn't using MicroCap is beyond me. Krenzo posted:I'm going off of research papers and books that are written on the subject. They have actual products out now that work on this technology. I would probably have given up by now if there were no papers and examples for me to work from. Here's a good thesis paper I found on the subject: http://scholar.lib.vt.edu/theses/available/etd-09192006-230733/unrestricted/Anderson_Final_ETD_Version.pdf Man, this is a really good paper. Never heard of a vivaldi antenna, but it works better than anything I've seen for UWB. Gonna save this for later. Also what the hell is a DOCTOR OF PHILOSOPHY in ELECTRICAL ENGINEERING? ANIME AKBAR fucked around with this message at 15:01 on Dec 9, 2010 |
|
#
?
Dec 9, 2010 14:57
|
|
|
ANIME AKBAR posted:Also what the hell is a DOCTOR OF PHILOSOPHY in ELECTRICAL ENGINEERING?
|
|
#
?
Dec 9, 2010 15:16
|
|
|
I thought that electrical engineering was applied science, not philosophy.
|
|
#
?
Dec 9, 2010 15:21
|
|
|
Maybe he works for google
|
|
#
?
Dec 9, 2010 15:45
|
|
|
PhD Stands for Doctorate of Philosophy
|
|
#
?
Dec 9, 2010 16:23
|
|
|
Have people generally had good experiences with buying parts from Asian ebayers? I've been playing around with some Arduino projects and certain components can be bought for a fraction of the cost via ebay as long as you don't mind 2-4 week ship times I guess. For example I was looking at getting some DS18S20's, but mouser/digikey want $6 each plus shipping for them, whereas I can get them from china for 5/$10 with free shipping. Seemingly the only difference between the two is that the ebay ones are probably the non-ROHS compliant version. The sellers have tonnes of good feedback so I'm not worried about actually getting the parts, but are they likely to be cheap knockoffs that may or may not work?  http://cgi.ebay.ca/5-pcs-DS18S20-DS1820-1-Wire-Digital-Thermometer-Dallas-/390221418978?pt=LH_DefaultDomain_0&hash=item5adb0251e2
|
|
#
?
Dec 9, 2010 22:53
|
|
|
Golden Mongoose posted:For example I was looking at getting some DS18S20's
|
|
#
?
Dec 9, 2010 22:57
|
|
|
I would like this bass amp to get back to working order. RB75 Randall Bass combo:  It's not a bad little unit, but a while ago, when plugged into a electric drum kit, bizzare sound came from it, and then an acrid smell. It has not worked since. I opened it up just recently, and this is what greeted me against the heat sink:  Now, i've had a good google around for a schematic for this amp, but to no avail. Can any one identify these components? And is it worth while tooling around with it to try fix?
|
|
#
?
Dec 10, 2010 01:13
|
|
|
Looks like a couple of power transistors blew up. If the writing on the casing is intact you can figure out what exactly it is and replace it with the same or a similar part. That might not be the only damage, though. Transistor Q11 also has a scorch mark under it, and who knows what else got damage in the same event.
|
|
#
?
Dec 10, 2010 01:26
|
|
|
ANIME AKBAR posted:Never heard of this before. How are its component libraries? What version are you using? http://www.spectrum-soft.com/index.shtm I'm using the latest version. Its component library is great; but it's its interface that really makes it shine.
|
|
#
?
Dec 10, 2010 06:58
|
|
|
|
| # ? Jun 11, 2024 03:45 |
|
|
BattleMaster posted:Looks like a couple of power transistors blew up. If the writing on the casing is intact you can figure out what exactly it is and replace it with the same or a similar part. Add Q13 and Q8 to that mix. They both went up too. I vote scrap it.
|
|
#
?
Dec 10, 2010 07:06
|
|