Post by thread:[PIC]: Questions regarding the PIC16C77X and its b

Hello, I have a few questions in regards to ADC capable PICs (in regards to the project mentioned below). Unfortunately, my knowledge is lacking although I am trying to correct that), so bare with me. I am wanting to sample input from a telephone, encode that input into mu-law audio format, and then send that data over a LAN to another machine. I found a company that sells boards that use the CS8900 ethernet chips (preassembled and kits), as well as a tutorial on how to send UDP packets over an ethernet link using it. Right now however, I am more concerned with sampling the signal from the telephone, and to make sure my basic design for that is correct. Since human speech usually is below 4000 hz, the minimum sample rate I would need is 8,000 samples per second. Since the signal itself's highest frequency should be around 4000 hz, a 12 bit resolution (4096 possible values) should work. Is this correct? I have been looking at the PIC16C77X MCUs.. I was also looking at the PDF datasheet at: http://www.microchip.com/download/lit/pline/picmicro/families/16c77x/30275a.pdf So here are some more somewhat basic questions: The pinout diagram shows pins 11 as AVdd, pin 12 as AVss, pin 4 as VRef- and pin 5 as VRef+. These are all of the pins I will need to use specifically for the ADC part of things, correct? I wish to have a telephone connected via a standard telephone cord to an RJ11 jack on a standalone unit (with among other things the PIC in it). A 9V power supply will be used to power the phone line. Since 9v is way too much to connect to the AD on the PIC (correct?), a resistor will need to be used to get it below the 4.096V level (which is the max voltage level for AVDD,AVSS right?). So, basically I was thinking the negative terminal on the battery can be connected to the RING pin on the RJ11 jack, and the TIP pin on the jack can be connected to the Avdd pin on the PIC, and have the AVss pin connected to the positive terminal on the battery... I would need a resistor inbetween the RJ11 TIP pin and the AVDD pin (correct?).. I have the bad feeling that I have something (lots of things?) backward however (hence the reason for this post).. I also don't know what to use for VREF + or VREF -... In the datasheet it mentions "The analog reference voltages are software selectable to either the device's analog positive and negative supply voltages (AVDD/AVSS), the voltage level on the VREF+ and VREF- pins, or the internal voltage reference if available (VRH, VRL)." Does this mean that only the AVDD, and AVSS pins are necessary, and VREF+ and - are not needed, assuming I set the reference voltage to the analog positive and negative supply? If someone would care to point out the mistakes I made above, point me in the right direction etc.. that would be great! Nathan -- http://www.piclist.com hint: To leave the PICList spam_OUTpiclist-unsubscribe-requestTakeThisOuTmitvma.mit.edu

8Kbit/s sec sampling will be fine. An 8 bit (256 level) ADC will also be fine for telephone quality - it's what the phone companies use after all! You don't need 12 bits and 10 bits is probably more resolution than justified. For degraded speech quality I think you can drop to ~6 bit resolution and ~6k sampling but it sounds pretty rough! I'll leave the exact ADC connections to someone else but you need to use a 5V regulator to drop the 9V supply to 5V for the ADC; a resistor will not be sufficient. The phone line will also need to be isolated from the "outside world". If the kit is in a plastic box along with the battery, you may be able to get away with it (at least for private use !) but for any external connection (eg the ethernet link) or sellable product you will need to provide an isolation barrier (and some sort of compiance/approval) depending on your market. Don't forget the 100Vac or so you get back down the line when the phone rings - you will need to provide protection for this also! Richard P Hello, I have a few questions in regards to ADC capable PICs (in regards to the project mentioned below). Unfortunately, my knowledge is lacking although I am trying to correct that), so bare with me. I am wanting to sample input from a telephone, encode that input into mu-law audio format, and then send that data over a LAN to another machine. I found a company that sells boards that use the CS8900 ethernet chips (preassembled and kits), as well as a tutorial on how to send UDP packets over an ethernet link using it. Right now however, I am more concerned with sampling the signal from the telephone, and to make sure my basic design for that is correct. Since human speech usually is below 4000 hz, the minimum sample rate I would need is 8,000 samples per second. Since the signal itself's highest frequency should be around 4000 hz, a 12 bit resolution (4096 possible values) should work. Is this correct? I have been looking at the PIC16C77X MCUs.. I was also looking at the PDF datasheet at: http://www.microchip.com/download/lit/pline/picmicro/families/16c77x/30275a.pdf So here are some more somewhat basic questions: The pinout diagram shows pins 11 as AVdd, pin 12 as AVss, pin 4 as VRef- and pin 5 as VRef+. These are all of the pins I will need to use specifically for the ADC part of things, correct? I wish to have a telephone connected via a standard telephone cord to an RJ11 jack on a standalone unit (with among other things the PIC in it). A 9V power supply will be used to power the phone line. Since 9v is way too much to connect to the AD on the PIC (correct?), a resistor will need to be used to get it below the 4.096V level (which is the max voltage level for AVDD,AVSS right?). So, basically I was thinking the negative terminal on the battery can be connected to the RING pin on the RJ11 jack, and the TIP pin on the jack can be connected to the Avdd pin on the PIC, and have the AVss pin connected to the positive terminal on the battery... I would need a resistor inbetween the RJ11 TIP pin and the AVDD pin (correct?).. I have the bad feeling that I have something (lots of things?) backward however (hence the reason for this post).. I also don't know what to use for VREF + or VREF -... In the datasheet it mentions "The analog reference voltages are software selectable to either the device's analog positive and negative supply voltages (AVDD/AVSS), the voltage level on the VREF+ and VREF- pins, or the internal voltage reference if available (VRH, VRL)." Does this mean that only the AVDD, and AVSS pins are necessary, and VREF+ and - are not needed, assuming I set the reference voltage to the analog positive and negative supply? If someone would care to point out the mistakes I made above, point me in the right direction etc.. that would be great! Nathan -- http://www.piclist.com hint: To leave the PICList .....piclist-unsubscribe-requestKILLspam@spam@mitvma.mit.edu -- http://www.piclist.com hint: To leave the PICList piclist-unsubscribe-requestKILLspammitvma.mit.edu

On Mon, 23 Sep 2002, Nathan wrote: > > I am wanting to sample input from a telephone, encode that input into > mu-law audio format, and then send that data over a LAN to another > machine. I found a company that sells boards that use the CS8900 > ethernet chips (preassembled and kits), as well as a tutorial on how > to send UDP packets over an ethernet link using it. Right now however, > I am more concerned with sampling the signal from the telephone, and > to make sure my basic design for that is correct. As I recall mu-law is 8 bit coding of 12 bit dynamic range. So if you wish to provide full dynamic range you should convert at 12 bits and have a lookdown table. If full dynamic range is important to you. Personally I'd probably use a 10 bit PIC since I have them already. > > Since human speech usually is below 4000 hz, the minimum sample rate I > would need is 8,000 samples per second. Since the signal itself's > highest frequency should be around 4000 hz, a 12 bit resolution (4096 > possible values) should work. Is this correct? First part yes, second part no. The lowest sample rate for 4000 Hz is 8000 samples per second. It has nothing to do with resolution of the converter. > I have been looking at the PIC16C77X MCUs.. I was also looking at the How about the 18F252 or 18F452? Cheaper and better. {Quote hidden}> So here are some more somewhat basic questions: > > The pinout diagram shows pins 11 as AVdd, pin 12 as AVss, pin 4 as > VRef- and pin 5 as VRef+. These are all of the pins I will > need to use specifically for the ADC part of things, correct? I wish to > have a telephone connected via a standard telephone cord to an RJ11 jack > on a standalone unit (with among other things the PIC in it). A 9V power > supply will be used to power the phone line. Since 9v is way too much to > connect to the AD on the PIC (correct?), a resistor will need to be used > to get it below the 4.096V level (which is the max voltage level for > AVDD,AVSS right?). > > So, basically I was thinking the negative terminal on the battery can > be connected to the RING pin on the RJ11 jack, and the TIP pin on the > jack can be connected to the Avdd pin on the PIC, and have the AVss > pin connected to the positive terminal on the battery... I would need > a resistor inbetween the RJ11 TIP pin and the AVDD pin (correct?).. Interfacing to the telephone line requires a bit more understanding of telephone lines in general. First off, it's balanced line audio. You are familiar with unbalanced lines. Secondly, if it is telephone lines outside of the central office, it has DC on it as well. So you need to block the DC. You can probably get away with a less-than-perfect interface with the phone line to prove your concept. Just don't hook an oscilloscope or any other grounded piece of equipment to the phone line in a way that grounds either tip or ring. > what to use for VREF + or VREF -... In the datasheet it mentions "The > analog reference voltages are software selectable to either the > device's analog positive and negative supply voltages (AVDD/AVSS), > the voltage level on the VREF+ and VREF- pins, or the internal > voltage reference if available (VRH, VRL)." Does this mean that only > the AVDD, and AVSS pins are necessary, and VREF+ and - are not needed, > assuming I set the reference voltage to the analog positive and negative > supply? Yes, that's what you'd want to do. Amplify the audio so it has 2.5 volts of DC on it and enough signal to go rail to rail at maximum level. Keep asking questions! Cheerful regards, Bob -- http://www.piclist.com hint: To leave the PICList .....piclist-unsubscribe-requestKILLspam.....mitvma.mit.edu

> so bare with me. Um, that could get me arrested, at least if we do it in public. > Since human speech usually is below 4000 hz, the minimum sample rate I > would need is 8,000 samples per second. Since the signal itself's > highest frequency should be around 4000 hz, a 12 bit resolution (4096 > possible values) should work. Is this correct? The signal frequency has nothing to do with the A/D resolution. Yes, you need to sample at 8KHz minimum. And this assumes a very sharp external filter that cuts off at 4KHz. It may be easier to sample a bit faster, like 16KHz for example, then do software low pass filtering to get the 8KHz samples. In either case, you don't need 12 bit resolution if you just want to get voice from here to there. Voice will be quite understandable at 10 bit resolution. > I have been looking at the PIC16C77X MCUs.. I was also looking at the > PDF datasheet at: I would stay away from those PICs. They useful when you really need 12 bit A/D resolution and therefore have to put up with all the other stuff to get it. 10 bit A/D is a lot easier and gives you a much wider choice of PICs. ***************************************************************** Embed Inc, embedded system specialists in Littleton Massachusetts (978) 742-9014, http://www.embedinc.com -- http://www.piclist.com hint: To leave the PICList EraseMEpiclist-unsubscribe-requestspam_OUTTakeThisOuTmitvma.mit.edu

On Mon, 23 Sep 2002, Olin Lathrop wrote: > > so bare with me. > > Um, that could get me arrested, at least if we do it in public. > > > Since human speech usually is below 4000 hz, the minimum sample rate I > > would need is 8,000 samples per second. Since the signal itself's > > highest frequency should be around 4000 hz, a 12 bit resolution (4096 > > possible values) should work. Is this correct? > > The signal frequency has nothing to do with the A/D resolution. Yes, you > need to sample at 8KHz minimum. And this assumes a very sharp external > filter that cuts off at 4KHz. It may be easier to sample a bit faster, like > 16KHz for example, then do software low pass filtering to get the 8KHz > samples. > > In either case, you don't need 12 bit resolution if you just want to get > voice from here to there. Voice will be quite understandable at 10 bit > resolution. I probably should have phrased that better.. That was actually 2 separate questions.. One was to confirm that the the sample rate of 8,000 samples per second (which is the Nyquist rate for a 4,000 hertz signal) is sufficient, and the other is if a 12 bit resolution is needed to accurately reproduce the signal for 'telephone quality' audio. I don't know how much the difference in the quantization error rate for 12 or 10 bit resolution sample vs. 8 bit resolution sample would affect the audio quality (given that its human speech). As I understand it most telephone calls are eventually converted into mu-law format.. Also, as I understand it, and as Bob Blick pointed out in an earlier response, mu-law encoding takes a sample with a 12 bit resolution, and produces 8 bit code.. If the difference in the way a 10 bit sample vs. a 12 bit sample encoded in mu-law format sound is negligible, then I would much prefer to take the 10 bit resolution route. Mainly I was just wanting a sanity check on that portion of things. I was mainly concerned with the part of my post that describes how I understand things would be connected to the PIC, as that is the area I am most 'fuzzy' on. Feel free to bat me over the head with a clue bat if any of that doesn't make sense, or I am wrong, etc.. I'm still very new to this field, and am trying to learn :) Nathan -- http://www.piclist.com hint: To leave the PICList piclist-unsubscribe-requestspam_OUTmitvma.mit.edu