Can anyone suggest a single chip 2400 baud modem solution please?
in [Basics]
|
From: David Eather on 17 May 2010 08:27 Can anyone suggest a single chip 2400 baud modem solution please? (sorry to repeat myself)
From: David Eather on 17 May 2010 08:30 On 17/05/2010 10:27 PM, David Eather wrote: > Can anyone suggest a single chip 2400 baud modem solution please? > > (sorry to repeat myself) I should also add in small quantity, PDIP, SOIC or TQFP preferable
From: Tim Wescott on 17 May 2010 12:25 David Eather wrote: > On 17/05/2010 10:27 PM, David Eather wrote: >> Can anyone suggest a single chip 2400 baud modem solution please? >> >> (sorry to repeat myself) > > I should also add in small quantity, PDIP, SOIC or TQFP preferable These days? Either a really old chip that's still in production, or a fairly new microprocessor with an ADC and some appropriate code. Have you looked at the usual suspects? National Semi, ON, Fairchild? At one point there were a number of these chips to be had, all (IIRC) either 24- or 40-pin PDIP parts. But I suspect the market for them has become a victim of advancing technology: this is something that you could easily do in one corner of an ASIC that's also capable of 38.4k, 56k, and auto-negotiated speeds. So I don't know if anyone still makes them outside the context of one-chip Hayes compatible telephone modems. If all you want to do is just a cruddy V.22bis 2400 baud modem, and you don't want to haunt back-ally gray-market used semiconductor brokers, consider that this is an easy task to accomplish with either a ridiculously small DSP chip or fairly accomplished microprocessor. There's a lot of details to get the procedural signaling correct, but the actual signal processing would be a snap for the right guy*. In fact, if you can't find a new modem to do this for you, this is probably the reason. I did a quick search to remind myself just what the standard was, and found at least one vendor that'll sell you modem software -- although by the time you've gotten your system twisted around to match their assumptions, you may have wasted more time than it'd take to roll your own, and spent licensing money to boot. * Like me. Since you have given me this outstanding opportunity for a shameless plug I'll take it: contact me if you want to roll your own, but don't feel up to writing the signal processing code. -- Tim Wescott Control system and signal processing consulting www.wescottdesign.com
From: David Eather on 17 May 2010 20:49 On 18/05/2010 2:25 AM, Tim Wescott wrote: > David Eather wrote: >> On 17/05/2010 10:27 PM, David Eather wrote: >>> Can anyone suggest a single chip 2400 baud modem solution please? >>> >>> (sorry to repeat myself) >> >> I should also add in small quantity, PDIP, SOIC or TQFP preferable > > These days? Either a really old chip that's still in production, What can I say? I'm a man of the 20th century - it's just a pity were in the 21st. I was hoping the sales pitch / blurbs about smart vending machines etc had translated into easy to access parts. or a > fairly new microprocessor with an ADC and some appropriate code. > > Have you looked at the usual suspects? National Semi, ON, Fairchild? At > one point there were a number of these chips to be had, all (IIRC) > either 24- or 40-pin PDIP parts. Yep, tried them. No real luck. I came to the same conclusion of microprocessor and code. Microchip's dsPIC30 series have a soft modem library with a free sub library or 5$ per copy for evaluation of the higher performance library (up to 14400). But I suspect the market for them has > become a victim of advancing technology: this is something that you > could easily do in one corner of an ASIC that's also capable of 38.4k, > 56k, and auto-negotiated speeds. So I don't know if anyone still makes > them outside the context of one-chip Hayes compatible telephone modems. After more searching and re-examining the project assumptions I found these guys www.radi.com who seem to be happy to sell small quantities and samples (under 50$ for the 1/2" 14400 and 33600 modems). So I have a grin from ear to ear. I don't really need the speed but over time project bloat might make it useful. > > If all you want to do is just a cruddy V.22bis 2400 baud modem, and you > don't want to haunt back-ally gray-market used semiconductor brokers, ummm..... no. > consider that this is an easy task to accomplish with either a > ridiculously small DSP chip or fairly accomplished microprocessor. > There's a lot of details to get the procedural signaling correct, but > the actual signal processing would be a snap for the right guy*. In > fact, if you can't find a new modem to do this for you, this is probably > the reason. I had come down to thinking my options were just that, or to go and buy dial-up modems (big, lots of cables, extra power-supply requirements, looks horrible etc) > > I did a quick search to remind myself just what the standard was, and > found at least one vendor that'll sell you modem software -- although by > the time you've gotten your system twisted around to match their > assumptions, you may have wasted more time than it'd take to roll your > own, and spent licensing money to boot. I have some experience with PIC's but was also coming to that conclusion. > > * Like me. Since you have given me this outstanding opportunity for a > shameless plug I'll take it: contact me if you want to roll your own, > but don't feel up to writing the signal processing code. Shameless plug welcomed. Thanks for your answers, experience and considerations. >
From: whit3rd on 19 May 2010 14:51
On May 17, 5:30 am, David Eather <eat...(a)tpg.com.au> wrote: > > Can anyone suggest a single chip 2400 baud modem solution please? > I should also add in small quantity, PDIP, SOIC or TQFP preferable Pull open a set-top box and see what it uses. A Samsung STB I have handy has Si2400 and Si3015 on the telephone port. |