Jump to content
Electronics-Lab.com Community


  • Posts

  • Joined

  • Last visited


gg4rest's Achievements


Newbie (1/14)



  1. I don't know if their indian language site works better, but none of their links on their english site work so I can't give you much information on that. Are you creating a PCB? Wireless, from what I've heard, is difficult to make reliable for a hobbyist. Especially if it isn't on a PCB. I've had difficulty occassionally with a wired system and that should be relatively simple. The problem I think with wireless is, if you don't have a spectrum analyzer you can't really tell if your signal is being transmitted in an easy manner. I've done very little with RF so maybe someone else has a more informed opinion.
  2. Most microcontrollers have a serial interface. You connect this to a MAX232 chip (or equivalent) in addition to some capacitors (check the data sheets). This chip is available through samples from many companies. It basically does the level conversion from the micro to the PC. You then connect this to a DB-9 connector with the right pin configuration (typically only require 3 pins TX, RX, and GND). You should then be able to talk to a PC through a serial port. This is also usually how your will program the internal EEPROM or flash on the microcontroller. I suggest you look at getting a microcontroller with onboard electrically erasable (EEPROM) or flash memory as this makes development easier than UV erasable or external EPROM memory. On the PC side, you will need a program that will read data from the serial port and then save it to disk.
  3. You will also need an instrumentation amplifier (like an INA338 from TI) before the UAF42 ICs. The signals will go from the sensor to the instrumentation amplifier and then to the UAF42 filter and then to the microcontroller. I've never interfaced to a flash drive before, I just would output to a PC for my system. But if you want portability, I guess that an industrial flashdrive is what you are looking for. Sandisk has an article on interfacing a 80C51 (or 68HC11 with some modification) microcontroller to a flashdrive. You might want to check the price on these drives. http://www.sandisk.com/industrial/flash-drive.asp http://www.sandisk.com/pdf/oem/AppNote80C51FlashATAv1.0.pdf I'm partial to the HC11 because I've used it before but it is up to you. I think that you can get 80C51's as samples from some sites. Let me know which type you want. Are you making a PCB or will you just be using sockets and wirewrap or solder? What is your budget?
  4. So what you are thinking of is a blood glucose monitor. Most of these currently extract blood from the person and test it using some sort of sensor. Doing a quick search I did find one non-invasive one that used photonic measurements of the eye to determine blood sugar. Some use Spectroscopy. It looks like a rather complicated process from my quick scan of articles because of the small concentrations involved in the blood. There seems to be no electronic sensor that you can buy unless it is built into an already build glucose sensor. I guess you could try to detect the sugar in the liquid by measuring boiling point of the liquid very very accurately but I don't know how much it would vary since the amount of sugar in the blood is in the mmol range. Here a a block diagram of a blood glucose system; however, the "invasive/non-invasive sensor" is not specified. http://focus.ti.com/vf/docs/blockdiagram.tsp?family=vf&blockDiagramId=2003 Sorry I couldn't help more. Maybe someone else has some suggestions.
  5. Do you mean its concentration in a liquid? Or how much granular sugar there is in a container?
  6. I went to the University of British Columbia in Canada. Does the piezoresistive sensor actually measure force or are you trying to correlate the angle of the piezoresistive sensor to the force? As I recall, piezoresistive sensors change their resistance with angle of bending which doesn't necessarily mean there will be a force generated by the muscles (ie the elbow can be bent at an angle without using any muscle force). How are you orienting the sensors around the elbow to detect force generation (a picture would help me if possible)? Are you trying to determine force at the hand or at the actual joint? Do you have more than one sensor (one for flexors and one for extensor muscles)? I think your sampling rate may be too high if you are simply looking for gross human motion. The sampling rate you should choose is dependent on the maximum velocity of the joint in question. This will save you on memory storage if that is important to you. From a quick experiment on myself, I can swing my elbow at full velocity through its full range (0 degrees to around ~120 degrees and back to 0 degrees) at a rate of 2.1 times per second or ~252 degrees/sec. Therefore a sampling rate of 1000 Hz will give a resolution of 0.252 degrees. This works out to around 476 data points for the motion of the arm from bent to straight at near maximum velocity. Most motions will be much slower than that. As for circuits, you will need an instrumentation amplifier with a gain range of around 100 to 200 to bring you 0-50 mV signal to 0-5 V or 0-10 V (depending on your ADC range). This circuit should be placed as close as possible to the piezoresistive sensor to minimize amplification of transmission noise. You can get a whole instrumentation amplifier in one IC through samples from TI or Analog (or some others). Following that, you will need a low pass filter at a cutoff frequency of approximately 5 to 10 times lower than your sampling frequency (for 1000 Hz it will be around 200 to 100 Hz). This ensures you will absolutly not have aliasing when you sample the signal. I typically use UAF42 ICs from TI for my motion applications in a Butterworth configuration. There is a DOS program to help you design your filter and select resistor values). I suggest you breadboard these circuits and look at the output from an oscilloscope. You can then see how much detail the system is giving you and adjust your filter cutoff frequency. If the signal is "noisier" than you would like, lower the cutoff frequency to make the signal more smooth. If it is is too smooth, increase the cutoff frequency. There is a tradeoff because the lower you can make the cutoff frequency, the lower you can make the sampling rate and thus save memory but the tradeoff is less detail in the measured signal. Let me know if you have any more questions.
  7. So are you actually measuring the electromyogram signal (the electrical signal sent to the muscle)? If so, this is similar to my MASc. So I'll be more than glad to help out. What university do you go to? I need some more answers to help you. Are you sure the signal isn't bipolar? EMG signals are bipolar unless you rectify them. How much detail do you want in the signal? Do you just want to general amplitude or do you want all the detail of the signal? I can give suggestions on reducing the sampling rate and filter cutoffs if I know this. You may have memory problems with 1000 kHz. At 1000 kHz, you will fill your memory in a very short time and have extreme difficulty processing and storing the data and will likely require a DSP and a hard drive. At 1000 kHz, you would fill 512 MB of memory in about 9 minutes (without compression, assuming 8-bit). Are you sure you don't mean 1000 Hz?
  8. My TVs all have only 2 prongs.
  9. So you decided against a wired link? Texas Instruments has the ICs you are looking for. I don't know if they supply sample parts to India though. http://focus.ti.com/analog/docs/articles.tsp?articleType=brc&templateId=5266&familyId=367&path=templatedata/cm/brc/data/200210_rf_ismhome&DCMP=RF+NotApplicable&HQS=NotApplicable+OT+ismrf
  10. Hi Garvey, You may want to contact that company and ask their pricing for 1 or 2 chips. They say $2.40 and $3.85 in quantities of 100K but you won't be using anywhere near that amount. It might be too costly. Do you care if different people can say the phrase and gain entrance? Or can it be anyone to say the command? Assuming the IC is too expensive. Here is what you could do. For different amplitudes, you could record the signal to memory, scale the signal to its maximum (ie if they said it quietly it would still use the full range). You may also have difficulty with people talking slower or faster. To solve this, you would need to "shrink" or "stretch" the signal to fit the same length. To do that, you will need some sort of indicator to tell the person when to speak (like a tone). That lets you line up the signal to be compared with the one in memory. I don't know how you will determine the end of what they are saying. Following all this processing, you then perform a correlation between the pattern signal and the signal just entered into the system. You do this by multiplying each sample together (which is why they should be the same length), and then adding all the multiplications together. This step is what typically kills the system because all those multiplies is processor intensive. Following the correlation, you simply need to "tune" the system (ie if the correlation is over X then open door, else, quit). The lower you make X, the more likely the door will open for any noise. The higher, the door will not open at all for any sound. I believe this method will have problems discriminating between different people though. You can use fuzzy logic or a neural network algorithm to get more specific. Hope some of that helps.
  11. I assume you want to record the voltage as well as display it right? Anyway, to do this you will require several things. 1. A low pass filter (LPF) and amplification 2. An analog-to-digital converter (ADC) 3. A microcontroller (and some sort of flash memory) or a connection to a PC. The low pass filter is to prevent aliasing in the signal when it is sampled by the ADC. The amplification is to size the input voltage to the voltage range of the ADC. The ADC is used to convert the voltage to digital format. Some microcontrollers have onboard ADCs which makes that part a little easier. Plus microcontrollers typically have a method of communicating with a PC which would make the data storage much simpler. How fast is your voltage signal changing? How fast do you want to sample the signal (ie how many measurements per second)? The limit on how fast you want to sample the signal is limited by how fast your link to the PC is as well as the maximum sampling rate of the ADC on the microcontroller. Assuming that your data transfer to the PC is 9600 bits/second and an 8-bit ADC, your absolute maximum sampling rate can be (9600 bits/sec)/(8 bits/sample) = 1200 samples/sec. You then need to verify that the ADC can in fact measure this quickly (usually it will). So the maximum sampling rate is 1200 Hz (it will usually be lower due to processor overhead). This means that you need a low pass filter cutoff at a maximum of 600 Hz to prevent aliasing. Aliasing is roughly high frequency signals "folding over" into your sampling range and appearing as a different lower frequency signal. You typically want to go 10X lower than that for good resolution. So the LPF cutoff is now at 60 Hz. Now to determine the amplification required between the LPF and the ADC. Assuming the ADC takes voltages from -5 to 5 V and your input voltage signal ranges from -2.5 to 2.5, you would require a maximum bipolar amplification stage of 2. Typically you set it slightly lower than that to ensure no clipping of the signal. This amplification is important to reduce quantization noise. For example, say your signal ranged from 0 to 1 V and you have a 2 bit ADC (unrealistic but this is an example) so 00b = 0 V, 01b = 1 V, 10b = 2V and 11=3V. Most of your measurements would be either 00b or 10b and you would miss all detail. Whereas if you amplified the input signal to 0 to 3 V. You would get the whole range of the ADC. Let me know if you need any help. I can give you more specifics on the components if you tell me more specifics on what you want to do.
  12. There is high voltage inside a TV set to launch the electrons through the CRT with enough velocity to cause the phosphor coating to light up. There is also a high voltage for the steering coils to cause the electron beam to curve to allow it to hit the top and bottom rows of the set and to scan across each row. I assume TVs don't have a ground pin because they likely have an isolation amplifier inside them which would separate the TV set from the household power. This would prevent noise from the household power from recking the tv signal. I'm not sure though. The reason that Air Conditioners, irons, and refrigerators would not, is because noise travelling to those devices is typically not as big a problem so an isolation transformer wouldn't be necessary.
  13. This site lets you search for various types of sensors. Most manufacturers will have datasheets for the sensor you are using and likely has circuits in their applications notes. http://www.globalspec.com/ As for books, "Principles of Engineering Instrumentation" by Ramsay has general information on types of sensors and the basics of how they work.
  14. The easiest and probably cheapest examples I can think of are using velocity sensors or accelerometers to control DC motor speed. If you want to get more complicated, you could use some position sensors as well and use two motors and a couple of linkages to make a simple planar robot. Depends really on how much controls experience you have.
  15. Hi audioguru, Where abouts do you live? When I drove from Vancouver to Calgary, there were no FM stations once I left the two major cities. So I was forced to listen to AM because I forgot to put songs on my mp3 player. It was kind of funny because there were some religious talk shows and funny old songs (from the 40s and 50s). And the announcers didn't seem to care if anyone was listening because they said whatever they wanted. As for the drive from Regina to Calgary, it is just about as bad. As for Andbor's problem. It is also possible that if he is using a digital FM tuner, it won't let him select the actual frequency his FM circuit signal is transmitted on. Or am I wrong?
  • Create New...