EEVblog Electronics Community Forum
Products => Computers => Vintage Computing => Topic started by: Ampera on December 26, 2017, 01:06:08 am
-
Today, with many landlines being shut down or converted to DSL dry runs, it's nearly impossible in a lot of areas to connect to old BBSes that are still around like the Back To The Future BBS.
So, I had an idea. Obihai sells boxes like the Obi200 that you can link up to a google voice account. The issue with most VOIP services is their use of lossy encoding, and while I can't exactly speak for Google Voice, Obihai supposedly states that their Obi200 boxes support faxing, which is as we all know, just a modem.
So what about results. Originally my test system was my Pentium Pro 200Mhz machine, but due to probably DMA or IRQ issues, I'm not sure what, other devices directly interfered with the signal.
After a quick bout with my 486, I settled on my Pentium 3 450mhz Slot 1 machine.
Now, I am new to BBSes, having never used one in my life, so my experience with terminal programs is rather lacking. I was using Hyperterm as it was easy to get, being already included on Windows. Now right out the gate I do not think this is an amazing option. I had crashes, and I had the program refuse to work or update any settings. Perhaps another terminal would be nicer.
I was using the BTTF BBS at 978-822-2586. First try I think I was using either 4800 or 9600 baud, and it worked! The characters rolled in, and I was asked to login. Here's where the no to unforeseeable issues happened.
Of course the terminal was getting garbage characters in and out every few moments. Now, before you call compression on me, this could be for any number of reasons. My settings could be off, and hyperterminal could just not be all that great. Either way, I am still working on this. I decided to post this as I thought it would interest some people, and if I can get this working reliably, it would make for a very neat and cheap way to hook into a BBS without a landline.
There are potentially other VOIP services that may work better, as the Obi boxes, to my knowledge, are not limited to Google Voice. Any suggestions are welcome, but keep in mind that the question "Why would I even want to do this" is one answered by "because I can".
-
Obi200 boxes support faxing,
...via T.38 (https://en.wikipedia.org/wiki/T.38) protocol.
the job of T.38 is to “fool” the terminal into “thinking” that it’s communicating directly with another T.30 terminal. It will also correct for network delays with so-called spoofing techniques, and missing or delayed packets with fax-aware buffer-management techniques.
Spoofing refers to the logic implemented in the protocol engine of a T.38 relay that modifies the protocol commands and responses on the TDM side to keep network delays on the IP side from causing the transaction to fail. This is done, for example, by padding image lines or deliberately causing a message to be re-transmitted to render network delays transparent to the sending/receiving fax terminals.
-
Traditional PSTN modems hardly ever work over a VoIP channel. If you get a G.711 channel from end to end you've solved one problem (I assume this is what you are thinking of as the converse of having lossy codecs, but the G.711 codecs are also lossy/compressed. However, these are the codecs which the PSTN modems were designed to tolerate). A bigger problem, however, is timing. Very few VoIP channels make any effort to synchronise the clocks at the two ends of the link. If the buffer at the receive end starts to underflow, some fudging is used to fill in the gaps. If the buffer at the receive end starts to overflow, some fudging is used to throw away some samples in the least audible manner. There may also be silence compression, and other manipulations, even when the codec is one of the G.711 ones. Either way, modems cannot tolerate this dynamic manipulation of time. They were designed for a world in which the PSTN has enough rubidium clocks to keep the entire end to end path in lock step.
There is a series of ITU specs - V.150, V.151, V.152 - designed to support modems over VoIP paths, in a similar manner to how T.38 supports FAX over VoIP paths. They put a modem at each end of the VoIP path, and translate between audio and pure data. While T.38 has been widely deployed, the V.15x series specs have seen very limited deployment.
You may have heard of FAX and modem signals being carried reliably over VoIP links run end to end by PSTN operators, using G.711. They are using carefully groomed channels, with rubidium clocks at each end. This is not something a consumer can generally replicate.
-
I wonder if the old OPUS software I used to run is still around somewhere? That was a very powerful yet easy to set up BBS system, quite fast on the message threads, too.
This being back in the 1986-1988 time frame, having a 2400 bps modem was a huge luxury, but we had an on-campus datanet with a serial port in each room, nominal speed was 9600bpbs but someone got the specs from the system manufacturer and found out it could actually do 19200bps, so a bunch of us petitioned the Telecom department and they upgrades use to get the full 19200. One of my friends who helped run the BBS is highly artistic so he drew these amazing ANSI graphics art files, and even animated a few. At 19200, it was very smooth, still usable at 9600, but as off-campus access was through a bank of 1200bps modems, it was unusable for outside callers - but simply hitting a key stopped the animation and went to the next page. I wish I still had copies of this somewhere. Since the name of the software was OPUS (nothing to do with the penguin), we of course went with a Bloom County theme for the BBS. Called it Starchair Enterpoop. The logon animation started with Opus and Milo and couple of other characters standing in Milo's Meadow and then a Star Trek style transporter beam materializing them on the bridge of the ship. Logoff sequence showed the tail end of the Enterprise, the shuttle bay doors would open, and the Basselope would fly out and zoom past.
-
I'm doing things slightly differently, rather than my old computers dialing a PSTN service, I'll be using a device which essentially connects my old machines via a serial port to my Wi-Fi network and connecting to the BBS via IP.
This product might be of interest to you (although it's currently out of stock): http://biosrhythm.com/?page_id=1453 (http://biosrhythm.com/?page_id=1453)
All the old machine sees is a Hayes compatible modem.
-
It would be nice to do this over the telephone. I might try to find someone with a known working modem and try to communicate with them just to see if I can get my settings right.
It just very very nearly works, and I want to be sure that I can't do anything to it through configuring to get it to work.
-
The other thing I thought about doing was running my own PSTN emulator which dialed a modem sitting in my own server. From there, use a proxy to forward IP packets to/from my old gear. Yes, it would be nice to keep the old tech all the way through, but these days, it's becoming less and less viable. I get great joy getting my old gear to work with modern equipment, for example, connecting my C64 to my LAN then out to the internet.
-
I get great joy getting my old gear to work with modern equipment, for example, connecting my C64 to my LAN then out to the internet.
Which is what we would have done back then, if the option was available. My external modem went in the bin as soon as I figured out how to connect my Amigas to the Internet via a PC.
The other thing I thought about doing was running my own PSTN emulator which dialed a modem sitting in my own server.
I have been thinking about this too. I have a number of old PCs with (otherwise useless) modem cards installed. Could install BBS software on one, and plug the modem into my (otherwise useless) internal phone wiring. Then relive the 'good old days' of downloading stuff from a BBS - without having to wait for a free line!
A full PSTN emulator would be neat, but probably not necessary. I am thinking 12~24VDC for loop detection and line current, basic DTMF detection to tell when a number is being dialed, and dial/ringing tones generated at the appropriate times - all controlled by a small MCU. Might also need a ringing generator to trip the PC modem's ring detect. Ringing voltage can be generated by an audio power amp driving a small mains transformer wired backwards - or perhaps just use raw 24VAC at mains frequency (since the PC modem probably won't mind if the ringing frequency is a bit higher than normal).
-
These also just became available a few days ago: https://www.cbmstuff.com/proddetail.php?prod=WiModem232OLED (https://www.cbmstuff.com/proddetail.php?prod=WiModem232OLED)
Available with or without OLED display.
-
HyperTerminal cannot correctly display ANSI codes and often showed control characters when connecting to a DOS BBS. If you can find a copy of Terminator by Bo Bendstien it was one of the better dialling programs with a ton of modes and emulation types, I do have it here on a Nightowl CD somewhere and could probably find it. Going one step further Wildcat 5 came out with a windows dialling that could connect to DOS and display the right codes but it also allowed you to connect to a BBS over the internet. Wildcat 5 was released into the public domain by the company that made it as the internet killed the business. It is a complete BBS system that allows you to have both ANSI screens and windows screens and DOS connectivity via the internet. It runs on Win95 upto Win98SE but I may have had it running on XP as well. As for DOS version I have all of them including manuals and compilers for making custom parts or plugins they called them for the lower Wildcat Versions. I ran a Wildcat BBS with ten lines from about 1993 to 1999 and had tried many other types of BBS software. Nothing really compared to Wildcat and it looked great. I had 40 CD drives in different towers which I still have in my shed and updated the CDs every three months or burned new ones from things I had downloaded from the net. It could multi task like a boss for the hardware it all ran on. Terminator is DOS only if you are going the DOS route. From memory Wildcat cost around $300.00 back then and you had to buy each update or new version at a lower cost. It wasn't cheap to run a BBS with constant hardware upgrades, phone lines which made long distance calls to retrieve its FIDO net NODE packets every night but it was great fun. I,m sure the DOS versions of Wildcat aren't protected anymore so parting with a copy if the floppies are still good wouldn't be a problem either. Cheers
-
I always used Crosstalk Communicator under DOS for BBS's. https://archive.org/details/CrosstalkCommunicatorTerminalEmulator
I've never been able to get a modem to work over a VOIP line but there may be some that will support it. Lots of fond memories of staying up half the night messing around on those, I sure don't miss having someone pick up the extension phone during the last 2 minutes of a 5 hour download though.
-
I get great joy getting my old gear to work with modern equipment, for example, connecting my C64 to my LAN then out to the internet.
Which is what we would have done back then, if the option was available. My external modem went in the bin as soon as I figured out how to connect my Amigas to the Internet via a PC.
The other thing I thought about doing was running my own PSTN emulator which dialed a modem sitting in my own server.
I have been thinking about this too. I have a number of old PCs with (otherwise useless) modem cards installed. Could install BBS software on one, and plug the modem into my (otherwise useless) internal phone wiring. Then relive the 'good old days' of downloading stuff from a BBS - without having to wait for a free line!
A full PSTN emulator would be neat, but probably not necessary. I am thinking 12~24VDC for loop detection and line current, basic DTMF detection to tell when a number is being dialed, and dial/ringing tones generated at the appropriate times - all controlled by a small MCU. Might also need a ringing generator to trip the PC modem's ring detect. Ringing voltage can be generated by an audio power amp driving a small mains transformer wired backwards - or perhaps just use raw 24VAC at mains frequency (since the PC modem probably won't mind if the ringing frequency is a bit higher than normal).
You don’t really need to do any of that. There’s various bits of open source software that can emulate the AT command set and fool software into thinking its talking to a modem, then the data is forwarded over IP to the destination.
So, your setup would be like this:
Old Computer’s Serial Port -> Null Modem Cable -> Modern Computer/Raspberry Pi -> Internet -> BBS via Telnet
One piece of software for emulating the AT command set is called ATduck and works reasonably well. Essentially, in your DOS/Windows terminal software (ProComm Plus, BananaComm, Etc.) you substitute the remote BBS’s IP for the phone number (ATDT 192000001100 to connect to 192.0.1.100).
ATduck can also be combined with a SLIP or PPP daemon, thus emulating a dialup ISP and allowing you to easily connect an old system to the internet (or just your LAN) through WinSock or a DOS PPP packet driver, all without needing a network card in the PC!
I’ve got a BeagleBone Black setup as a Retro Computing router using this method and it works great. I’ve got a Libretto 70CT connected wirelessly via Serial IR and a Compaq Portable III via normal wired serial.
I setup the software on the BBB to directly route data via telnet when it receives an ATDP dial command and to launch the PPP daemon on a ATDT dial command. That makes it easy to configure different applications on the retro machines without changing anything on the BBB.
Anyway, pretty much every BBS still running today provides access over Telnet. So that’s the way to go IMHO. (There are also DOS Telnet clients available that you could use to directly access the BBS via a packet driver, but they’re not quite as advanced as various communications packages from the era, in terms of displaying special ANSI sequences and stuff, hence the reason to emulate a modem with ATduck.)
-
That's not the fun way to go though. Using an Obi box is kinda a stretch enough. The idea is to have it work just like a phone line.
Anyways, I unfortunately have not gotten farther, but I have not attempted to. My Pentium 3 is a bit iffy on all of this with the software, my Pentium Pro has some strange interference issues with the modem and other cards (even on different IRQs) and my 486 refuses to detect the CD drive when I use a 2.5GB HDD instead of a 2.1GB one, a problem I've never had before.
I currently have a laptop with a built-in modem I am trying to get working with it. Hopefully that yields some success.
-
I'm not surprised the 486 doesn't work with a 2.5GB hard drive, many PCs of that era won't support anything larger than 512MB, hence the various overlay programs that were available once drives of that size appeared. IIRC 2GB is another barrier that some older systems couldn't go beyond.
-
I'm not surprised the 486 doesn't work with a 2.5GB hard drive, many PCs of that era won't support anything larger than 512MB, hence the various overlay programs that were available once drives of that size appeared. IIRC 2GB is another barrier that some older systems couldn't go beyond.
That 486 can LBA up to 8GB drives, and can use FAT16 with all that space, it just needs four partitions. The issue is down to the oak CD-ROM drivers not detecting a CD on any controller or connector with just that drive in there even though it worked fine before. Even when I set the drive to limit itself to 2.1GB, making it virtually indistinguishable from my other hard drive by use of a special jumper configuration, it still didn't work at all.
-
That's not the fun way to go though. Using an Obi box is kinda a stretch enough. The idea is to have it work just like a phone line.
Using ATduck or similar software makes it work just as if you have an external serial modem attached to the serial port. If you’re trying to get your old system to talk to a BBS or even a server in your house, I don’t see why you wouldn’t do it this way...
You’re going to have a very hard (if not impossible) time directly connecting two modems together (by that I mean wiring them directly together, Modem <—> Modem). Now, if you have a small analog PBX system in your house, it would work fine (the PBX would act like the Central Office’s switching gear, providing the required on and off hook voltages the modems expect).
So, unless you already have a PBX (or equivalent) system in your house, setting one up just to connect to a computer in the next room seems like a ton more work compared to setting up software on a Raspberry Pi to simulate a serial modem! (And you really only need to do that if the DOS software you want to use only supports modems (I.E., isn’t TCP/IP aware); for Packet Driver aware DOS software and WinSock aware Windows software, you can bypass the modem emulation and simply have the Pi act as a CSLIP or PPP server.
-
That's not the fun way to go though. Using an Obi box is kinda a stretch enough. The idea is to have it work just like a phone line.
Using ATduck or similar software makes it work just as if you have an external serial modem attached to the serial port. If you’re trying to get your old system to talk to a BBS or even a server in your house, I don’t see why you wouldn’t do it this way...
You’re going to have a very hard (if not impossible) time directly connecting two modems together (by that I mean wiring them directly together, Modem <—> Modem). Now, if you have a small analog PBX system in your house, it would work fine (the PBX would act like the Central Office’s switching gear, providing the required on and off hook voltages the modems expect).
So, unless you already have a PBX (or equivalent) system in your house, setting one up just to connect to a computer in the next room seems like a ton more work compared to setting up software on a Raspberry Pi to simulate a serial modem! (And you really only need to do that if the DOS software you want to use only supports modems (I.E., isn’t TCP/IP aware); for Packet Driver aware DOS software and WinSock aware Windows software, you can bypass the modem emulation and simply have the Pi act as a CSLIP or PPP server.
THESE (http://www.natcomm.com.au/au/telephone-line-simulators/) were pretty popular in Australia but bloody expensive for what they are.
I do recall seeing someone building one and providing schematics on the internet, but that was years ago. You'd be able to Google it. Two modems just connected together won't do anything.
-
The idea isn't to connect two modems in the same room. The idea is to simulate an actual POTS line that goes out to the real world on what could be best described as the Bell system using regular telephone numbers, and then to connect to a real BBS that still has a modem to connect to. Even 300 baud would be fast enough to get text in and out for a BBS.
-
You don’t really need to do any of that.
I don't need to do any of that, but I want to recreate the experience of using a dial-up modem - including hearing the tones it produces as it connects. For the most authentic retro-computing experience, I want to use real vintage hardware wherever practicable, including the modem.
The idea of connecting to a BBS via Telnet is interesting, but I don't want to be tied to a power-hogging modern PC (and certainly not a Raspberry Pi, which I don't own and never will). I have a few old PCs with internal modems and network cards that should be suitable. I now also have wireless Internet, so I am thinking an ESP8266 might do it (combined with the telephone line interface and an old 14400 external modem that I have).
Two modems just connected together won't do anything.
True, but they can be made to talk to each other if you just supply some line voltage. The only problem is without ringing voltage you have to tell the receiving modem to answer the call.
-
Some time ago Motorola made a SIP device (the VT1005 I think) that was quite good and also understood the rotary dialing of an old analog phone.
So if you have one you can use an old mechanical dial phone, the kind with the physical bell inside (!) for VOIP.
Why not go for the entire retro experience?
-
gotta use an acoustic coupler for full effect.
-
You don’t really need to do any of that.
I don't need to do any of that, but I want to recreate the experience of using a dial-up modem - including hearing the tones it produces as it connects. For the most authentic retro-computing experience, I want to use real vintage hardware wherever practicable, including the modem.
The idea of connecting to a BBS via Telnet is interesting, but I don't want to be tied to a power-hogging modern PC (and certainly not a Raspberry Pi, which I don't own and never will). I have a few old PCs with internal modems and network cards that should be suitable. I now also have wireless Internet, so I am thinking an ESP8266 might do it (combined with the telephone line interface and an old 14400 external modem that I have).
Two modems just connected together won't do anything.
True, but they can be made to talk to each other if you just supply some line voltage. The only problem is without ringing voltage you have to tell the receiving modem to answer the call.
I don’t understand what you plan to do with the ESP8266 and a external modem. Can you elaborate a bit?
If your vintage PCs already have network cards you can connect to a BBS via Telnet without “power hogging modern PCs” (which are actually a lot more efficient per MHz/Watt than a vintage system, by the way) in DOS or Windows directly.
Under DOS, simply download the appropriate Packet Driver for your network card (or use a NDIS shim) and a suitable Telnet application (there’s one included in the wonderful MTCP package). Under Windows 95/98 you can use HyperTerminal or even PuTTY (v0.61 is the last to run on 9x IIRC) for Telnet access.
I have a beautiful Compaq Portable III (286) system hooked to my network via an ISA based 3Com EtherLink III and it works great! It runs FreeDOS 1.2 now off a 1GB CF card. (Up to last year it was running MS-DOS 6.22 off the original 40MB IDE hard drive from 1987 and had network connectivity via the aforementioned BeagleBone Black Modem Emulator.)
I ended up installing the network card when I replaced the HDD with CF because I needed somewhere to plug in a XTIDE ROM (it’s an open source option ROM designed for PC XT and AT class machines, which gives them access to drives over 500MB; my Portable III was one of the first machines released with an IDE interface and the native BIOS only gives you a choice of about 20 drive types with no custom option). XTIDE also allows you to boot or access floppy and hard disk images hosted from a modern machine over a serial link (they show up as a physical B: or D: drive to DOS) hosted from a modern Windows machine running a console application, which is very handy, especially considering this machine only has a 5.25” drive.
Anyway, I have a NI GPIB card in the second ISA slot and use the system as a bridge to communicate with some older test gear I have. It logs data to a text file and runs an FTP Server TSR in the background, which a modern machine uses to pickup and process the data with. (Yes, I could modernize it with a USB GPIB adapter hooked directly to the modern system, but I love watching waveforms show up and data scroll by on the soothing orange plasma display. It warms the cockles of my heart.)
-
@TwoOfFive-TertiaryAdjunctToUnimatrixZero:
I came across this today while searching for something modem related. (http://goughlui.com/2016/05/03/project-the-definitive-collection-of-v-90v-92-modem-sounds) He was able to establish V.90 (56k) connections over VOIP to a local ISP. He also got V.92 connections going, from Australia to the US (!!!) no less.
He uses a Linksys Analog Telephone Adapter and routes the calls through a local Asterisks Server onto a VOIP provider. Apparently using the Linksys ATA is the key, as he can completely turn off VOIP echo cancellation (modems rely on the echo being there), set jitter and delay to fixed values and force using a specific codec.
(You wouldn’t need a local Asterisks Server; he only used that so he could easily record raw PCM wave files of the modem handshakes, for archival purposes, using WireShark.)
Anyway, I’d guess you need a highly configurable ATA and a VOIP provider that supports a good enough codec, but it looks doable.
-
That was the idea with the Obi200 box, however it has a quite extensive arrangement of settings, so I wonder if I can't disable echo cancellation somewhere in there and get better results.
-
Ive sent fax successfully using a VOIP box.. using an old inline ADSL filter/ RFI noise suppressor slowed down the handshake and proved to be just the ticket...
-
Update: More tests proving better success. At the moment I think it's HyperTerm that's screwing me over. I'm going to try Terminator, and see how far I get with that.
I'm getting full screen transmissions, characters are going back and forth, and it's seems normal for a while, but then it just spits a load of garbage characters and sort of hangs. refusing to send or receive more characters. This does sound awfully like some sort of signal degradation, but I was doing it at 9600 baud, so that could be a factor.
-
Update: More tests proving better success. At the moment I think it's HyperTerm that's screwing me over. I'm going to try Terminator, and see how far I get with that.
I'm getting full screen transmissions, characters are going back and forth, and it's seems normal for a while, but then it just spits a load of garbage characters and sort of hangs. refusing to send or receive more characters. This does sound awfully like some sort of signal degradation, but I was doing it at 9600 baud, so that could be a factor.
If you use G.711 and turn off echo cancellation in the VoIP devices you can often get the modems to sync up. The problem is they won't stay sync'ed up for long unless the clock speeds at the two ends of the VoIP link (e.g. inside the VoIP terminals) match precisely, and you won't achieve that with consumer grade VoIP equipment.
-
The key is getting them to use slower speeds.
-
The key is getting them to use slower speeds.
I've been trying 9600 baud and below.
-
The key is getting them to use slower speeds.
I've been trying 9600 baud and below.
I’d start at 1200bps and work your way up. (It’s a pretty robust, simple encoding scheme designed during an era when the POTS network was entirely analog and the last mile connection was really poor.)
14.4k is getting near the limit of a 1980s era POTS network. (In fact, the only reason 56k even works is because it’s only analog between your house and central office; it exploits the fact that, at that time (late-90’s), most COs were connected to each other with T1s or better, so it’s essentially digital from the CO to the ISP.)
-
The key is getting them to use slower speeds.
I've been trying 9600 baud and below.
I’d start at 1200bps and work your way up. (It’s a pretty robust, simple encoding scheme designed during an era when the POTS network was entirely analog and the last mile connection was really poor.)
14.4k is getting near the limit of a 1980s era POTS network. (In fact, the only reason 56k even works is because it’s only analog between your house and central office; it exploits the fact that, at that time (late-90’s), most COs were connected to each other with T1s or better, so it’s essentially digital from the CO to the ISP.)
Start at 300bps if you want it to work. Anything faster requires training, and the training will be screwed up every time the VoIP buffers need to slip.
-
The key is getting them to use slower speeds.
I've been trying 9600 baud and below.
I’d start at 1200bps and work your way up. (It’s a pretty robust, simple encoding scheme designed during an era when the POTS network was entirely analog and the last mile connection was really poor.)
14.4k is getting near the limit of a 1980s era POTS network. (In fact, the only reason 56k even works is because it’s only analog between your house and central office; it exploits the fact that, at that time (late-90’s), most COs were connected to each other with T1s or better, so it’s essentially digital from the CO to the ISP.)
Start at 300bps if you want it to work. Anything faster requires training, and the training will be screwed up every time the VoIP buffers need to slip.
I do not believe training started until v.32/9600 came around (well, it could have been used for v.27/4800 too), as that’s when echo cancelation started being used.
-
More success! The issue I am facing now is transmitting, but I have a theory as to why that's buggered.
My Pentium Pro is a bit stuffed on cards, and other devices are definitely interfering with the modem, namely the mouse. It's a serial mouse, and when it's connecting I can hear it crackling to the mouse movements in the speaker.
1200 baud worked a lot better, which is strange considering I attempted that earlier to no good success, but I think it was my Obi's settings. Receiving is DEFINITELY reliable, which shows that there is definite merit to this idea. It may not be the best idea, but it's certainly not bad. I was able to get weather, and some other stuff over the BBS before the link crashed after my system sent a load of garbage characters down the line (:/)
I think now it's a matter of getting a more reliable box to do this. I love my Pentium Pro, but it's a bit of a pig in some places. I would LOVE to use my 486, as it's a machine that would have been more likely to see a proper BBS, but I need to do a lot of work to get it in a position where it can use the modem correctly. I namely need to backup my PC-DOS installation and install Windows 95, which will just take some time.
300 baud didn't work for the BTTF BBS. It only received garbage data, and I think it's just an incompatibility with the BBS. 1200 baud worked perfectly fine, however, with, as I said, the only issue being occasional bouts of garbage data being sent because of interference from other devices. All the IRQs are fine, the modem is on IRQ4, that isn't used by anything else, it's just my own setup, I'll figure it out.
Thanks for all the help, I will try to do a video to document my success, and I hope with another machine I can get this into a working state.
-
The key is getting them to use slower speeds.
I've been trying 9600 baud and below.
I’d start at 1200bps and work your way up. (It’s a pretty robust, simple encoding scheme designed during an era when the POTS network was entirely analog and the last mile connection was really poor.)
14.4k is getting near the limit of a 1980s era POTS network. (In fact, the only reason 56k even works is because it’s only analog between your house and central office; it exploits the fact that, at that time (late-90’s), most COs were connected to each other with T1s or better, so it’s essentially digital from the CO to the ISP.)
Start at 300bps if you want it to work. Anything faster requires training, and the training will be screwed up every time the VoIP buffers need to slip.
I do not believe training started until v.32/9600 came around (well, it could have been used for v.27/4800 too), as that’s when echo cancelation started being used.
You can usually fudge V.22/600 and V.22/1200 with a fixed equaliser, although equaliser training is needed for top class results. V.22bis is pretty hopeless without equalizer training. The start up sequence for these modems includes both patterns for timing synchronisation, and patterns for equalizer training. All modems faster than V.22bis, like V.27ter/4800, require equalizer training.
What V.32 first introduced to a PSTN modem was not training. It was echo cancellation. Of course, the echo canceller also needs training, so V.32 has multiple training steps in its startup sequence.
-
The key is getting them to use slower speeds.
I've been trying 9600 baud and below.
I’d start at 1200bps and work your way up. (It’s a pretty robust, simple encoding scheme designed during an era when the POTS network was entirely analog and the last mile connection was really poor.)
14.4k is getting near the limit of a 1980s era POTS network. (In fact, the only reason 56k even works is because it’s only analog between your house and central office; it exploits the fact that, at that time (late-90’s), most COs were connected to each other with T1s or better, so it’s essentially digital from the CO to the ISP.)
Start at 300bps if you want it to work. Anything faster requires training, and the training will be screwed up every time the VoIP buffers need to slip.
I do not believe training started until v.32/9600 came around (well, it could have been used for v.27/4800 too), as that’s when echo cancelation started being used.
You can usually fudge V.22/600 and V.22/1200 with a fixed equaliser, although equaliser training is needed for top class results. V.22bis is pretty hopeless without equalizer training. The start up sequence for these modems includes both patterns for timing synchronisation, and patterns for equalizer training. All modems faster than V.22bis, like V.27ter/4800, require equalizer training.
What V.32 first introduced to a PSTN modem was not training. It was echo cancellation. Of course, the echo canceller also needs training, so V.32 has multiple training steps in its startup sequence.
You’re right! (And yes, I meant the echo cancelation training.) One of the nice things about V.90 (or was it V.92?) was that it could remember line characteristics and thus cut the overall training time practically in half after the first successful connection.
-
One of the nice things about V.90 (or was it V.92?) was that it could remember line characteristics and thus cut the overall training time practically in half after the first successful connection.
With modems other than V.90 and V.92 the link is usually analogue at both ends, and subject to considerable variability. Even if you always call the same number, you probably wouldn't be hitting the same port each time. Nothing remembered from a previous call will help with a new call. With V.90 and V.92 you are always calling a digital port. The ports in a large bank will behave identically, and your own local analogue link probably doesn't change too much unless the temperature is very different between calls. So, there is a high chance that the training from one call to a certain number will be close to the training for the next call to that number. V.90 and V.92 start out with that assumption, and only go into full training if the assumption proves false.
-
@TwoOfFive-TertiaryAdjunctToUnimatrixZero:
I came across this today while searching for something modem related. (http://goughlui.com/2016/05/03/project-the-definitive-collection-of-v-90v-92-modem-sounds) He was able to establish V.90 (56k) connections over VOIP to a local ISP. He also got V.92 connections going, from Australia to the US (!!!) no less.
He uses a Linksys Analog Telephone Adapter and routes the calls through a local Asterisks Server onto a VOIP provider. Apparently using the Linksys ATA is the key, as he can completely turn off VOIP echo cancellation (modems rely on the echo being there), set jitter and delay to fixed values and force using a specific codec.
V.92 (the older V.90 and the very old 56k flex) relies on a digital telephone network, i.e. ISDN. The RAS is connected via BRIs or PRIs to the telephone network and the codec is G.722. There's also a G.722 for VoIP, usually called 'HD voice' by marketing. You have to be very lucky to get a real V.92 connect via VoIP. In most cases the maximum connect is about 24k. Back in the days of dial-in I was involved in designing and building large carrier-grade dial-in platforms, and I'm still running a Fidonet node with a POTS modem via VoIP. BTW, you can make ISDN data calls via VoIP by using RFC4040's clear channel, but only a very few telcos support that.
-
@TwoOfFive-TertiaryAdjunctToUnimatrixZero:
I came across this today while searching for something modem related. (http://goughlui.com/2016/05/03/project-the-definitive-collection-of-v-90v-92-modem-sounds) He was able to establish V.90 (56k) connections over VOIP to a local ISP. He also got V.92 connections going, from Australia to the US (!!!) no less.
He uses a Linksys Analog Telephone Adapter and routes the calls through a local Asterisks Server onto a VOIP provider. Apparently using the Linksys ATA is the key, as he can completely turn off VOIP echo cancellation (modems rely on the echo being there), set jitter and delay to fixed values and force using a specific codec.
V.92 (the older V.90 and the very old 56k flex) relies on a digital telephone network, i.e. ISDN. The RAS is connected via BRIs or PRIs to the telephone network and the codec is G.722. There's also a G.722 for VoIP, usually called 'HD voice' by marketing. You have to be very lucky to get a real V.92 connect via VoIP. In most cases the maximum connect is about 24k. Back in the days of dial-in I was involved in designing and building large carrier-grade dial-in platforms, and I'm still running a Fidonet node with a POTS modem via VoIP. BTW, you can make ISDN data calls via VoIP by using RFC4040's clear channel, but only a very few telcos support that.
In simpler terms, it’s basically an analog connection from the user’s modem to the central office, then it goes digital to the ISP’s modem (over one of the 64k channels of a T1, T3, OC3 or whatever). Right? Basically, it’s ISDN only with the first hop being analog.
Speaking of ISDN, V.92 made bonding two dialup connections at the ISP (for a 115k dialup link) a lot more practical thanks to Modem on Hold; it’s a shame more ISPs didn’t support either feature. Being able to take an incoming call and instantly be able to reconnect was handy and made dialup a lot more like ISDN!
Ah, the good old days of dialup. I don’t even miss it a little. Though, I do kind of miss my old ISDN setup. Cable and DSL wasn’t available in my area yet, so I did ISDN in 2000. The line itself didn’t cost much more than a normal phone line, however all outgoing calls incurred a $0.01/Minute charge, however incoming calls were free. Luckily I knew the owner of a local ISP so we were able to setup the equipment for ISDN dialback. Basically, my router dialed into the ISP, performed a normal PPP login and hung up; the ISP would immediately dial my ISDN modem back, which would answer and go through the login process again. (The outgoing call charge only started after 10 seconds, so I was never charged for the initial calls into the ISP.)
It was a clever system and worked great for 3 years, until I finally got access to cable. That ISDN had fantastic latency (I would get 15ms pings to east coast Quake III servers) and basically never went down.
-
In simpler terms, it’s basically an analog connection from the user’s modem to the central office, then it goes digital to the ISP’s modem (over one of the 64k channels of a T1, T3, OC3 or whatever). Right? Basically, it’s ISDN only with the first hop being analog.
That's simplifying it a little too much, because it makes it appear that beyond the line card in your local exchange, everything was predictable. That's not the case. If your line card connected to a T1 using robbed bit signalling it didn't provide a clear 64k bps channel. Until the very end, the US in particular was still installing a LOT of robbed bit signalling (RBS) on T1s, and relatively little ISDN. Countries using E1s generally had 64k bps channels everywhere. However, even on pure ISDN paths, which were 64k bps end to end, there were often weird things going on with gain adjustment, and other obscure manipulations. ISDN has modes where any manipulation like this could be disabled, but a call from an analogue line into an ISDN line card couldn't get a call into this kind of clear channel mode. So, there was actually quite a bit of funky stuff degrading even the digital sections of the path between you are your ISP's V.90/V.92 digital modem.
The early experimental X2 and K56 modems were relatively simple, but by the time they had made them robust in the face of the funky behaviour of the networks they became quite complex. V.90 was largely the result of making X2 and K56 robust and merging them into a single standard.
-
Yep, but in the US it's a 56k channel typically (US' PDH is based on 7 bits while EU's / Japan's is 8 bits). And the PRI for the US is a T1. T3 or OC3 are used to transport multiple T1s. The bundling you've mentioned has nothing to do with V.92, it's MPPP (Multilink PPP). MPPP allows you to join multiple PPP sessions across different lines. IIRC, the MAX TNT (the box in the telco's data center) supported up to 8 MPPP channels per user across multiple chassis. You could have done the call-back just by signaling your telephone number (a normal call, but the called number rejects the call, free of charge) and used HDLC instead of PPP for a little bit more throughput.
-
Yep, but in the US it's a 56k channel typically (US' PDH is based on 7 bits while EU's / Japan's is 8 bits).
With a robbed bit T1 you get 8 bits for 7 samples, then lose the LSB on the 8th sample. So, you actually get 63kbps, and for voice it sounds considerably better than a simple 7 bit 56k channel would.
Japan (along with Taiwan, HK , and a few others) used plenty of robbed bit T1s in its PDH network, but tried to reduce them a lot after the introduction of ISDN. Most of the world used E1s, and always had 64kbps channels.
-
RBS is a tradeoff, it introduces some new issues. If T1 links aren't synchronized well, more LSBs could be lost. So the S/N decreases and ends up not being much better than for plain 56k. And in some cases, like a 'quiet' line, you can hear a tone created by the RBS.
-
RBS is a tradeoff, it introduces some new issues. If T1 links aren't synchronized well, more LSBs could be lost. So the S/N decreases and ends up not being much better than for plain 56k. And in some cases, like a 'quiet' line, you can hear a tone created by the RBS.
RBS was always a bit of a botch. Most aspects of the original T1 were well thought out. The two key failings were the lack of a rock solid sync scheme, and the lack of transparent channels, equally suitable for both data and voice. A couple of years later, when E1 was being defined, they had realised the failings of RBS, and avoided them. It took until ISDN before both the T1 and E1 people faced up to the sync issue.
-
The bundling you've mentioned has nothing to do with V.92, it's MPPP (Multilink PPP). MPPP allows you to join multiple PPP sessions across different lines. IIRC, the MAX TNT (the box in the telco's data center) supported up to 8 MPPP channels per user across multiple chassis. You could have done the call-back just by signaling your telephone number (a normal call, but the called number rejects the call, free of charge) and used HDLC instead of PPP for a little bit more throughput.
Sorry, part of my post got deleted while editing it. What I meant is that the Modem on Hold feature of V.92 made Multilink PPP (thanks for reminding me of what it was called) a lot more practical. (I.e., you wouldn’t be totally tying up both your landlines.)
As for my ISDN, it may have very well been a rejected call and HDLC, it was 17 years ago so it’s hard to know for sure. I know the ISP guy and I spent a hours and hours getting it to work. It finally took a three way call between me, him and a Cisco technician to get it all going, so it may have been a sub-optimal setup. But it worked and I got speeds consistent with a 128k link.
-
I remember a place my friend worked had a T1, that was the first time I'd ever seen one. He showed me downloading something at 80kB/sec and my jaw hit the floor, being used to topping out around 5kB/sec I had never seen such blazing fast downloading before. Now I yawn as my cheap home broadband hits 3 MB/sec.
-
My Pentium Pro is a bit stuffed on cards, and other devices are definitely interfering with the modem, namely the mouse. It's a serial mouse, and when it's connecting I can hear it crackling to the mouse movements in the speaker.
Check which COM port your mouse is using, it's probably conflicting with the modem. On most machines, COM1/3 and COM2/4 shared the same resources.
-
My Pentium Pro is a bit stuffed on cards, and other devices are definitely interfering with the modem, namely the mouse. It's a serial mouse, and when it's connecting I can hear it crackling to the mouse movements in the speaker.
Check which COM port your mouse is using, it's probably conflicting with the modem. On most machines, COM1/3 and COM2/4 shared the same resources.
I believe I'm gonna use a different modem later on. My tests happen every other day when I have time I want to allocate to it.
-
My Pentium Pro is a bit stuffed on cards, and other devices are definitely interfering with the modem, namely the mouse. It's a serial mouse, and when it's connecting I can hear it crackling to the mouse movements in the speaker.
Check which COM port your mouse is using, it's probably conflicting with the modem. On most machines, COM1/3 and COM2/4 shared the same resources.
I'd forgotten about that, yes COM1/COM3 usually shared the same IRQ, as did COM2/COM4. There were some utilities out there that would show you what was using each IRQ, some hardware could handle sharing an IRQ and some could not. Brings back a lot of memories fooling around with IRQ, DMA and IO range jumpers trying to get various cards to get along with one another. Modern plug n play has really made life a lot easier.
-
Today, with many landlines being shut down or converted to DSL dry runs, it's nearly impossible in a lot of a
I recently got a WiModem232 which I use on my Tandy 1000 EX and Amiga 1200 to access BBS's like BTTF & War Ensemble etc - I just dial them using "ATDT BTTFBBS.COM:23" and voila, the full BBS experience via the internet. :-+
There's a whole heap listed here which can be accessed via their telnet port:
http://telnetbbsguide.com/connection/dial-up/ (http://telnetbbsguide.com/connection/dial-up/)
(https://i.imgur.com/2QFX12vl.jpg)
(https://i.imgur.com/XALwd5al.jpg)
(https://i.imgur.com/so8POrBl.jpg)
(https://i.imgur.com/SnPTat9l.jpg)
(https://i.imgur.com/b2TYftal.jpg)
-
I was a BBS addict. I have been sitting on the source code for PCBoard BBS software the last few years. A little work was done by another person to get it buildable, but it requires a lot of esoteric software that can only be found in junk piles or the dredges of the internet... one upside is one of the libraries it was linked against (name escapes me at the moment), the original manufacture was able to send me copies of... until they suddenly stopped responding. I assume someone finally said "no" internally.
I've been debating putting it up on github as it's been quite a while since Clark development Company went out of business... and it isn't clear to me who the copyright holder is at this point (the bank that had their leins, I imagine).
... it's riddled with assembly for OS/2 and DOS... I've spent a little time over the years updating those functions.
....and it will only compile on specific versions of Windows that aren't available on MSDN anymore....
Nostalgia is a strong addiction as I spent more than a decade trying to find a copy of their source code before succeeding. :)
-j
Sent from my SM-N950U using Tapatalk
-
I too was a BBS addict when I was young (early 90’s), before I had access to the internet. Once, when I was about 11, a friend told me about this cool BBS he found that allowed you to play DOOM with strangers. I tried it at his house and it was pretty awesome, so I copied the connection file from his copy of ProComm Plus onto a floppy and transferred it to my system.
I spent hours and hours playing on it over several weeks. Then I discovered the number I had been dialing was long distance, halfway across the country. Then I had to tell my parents to expect the phone bill. Nearly $500! It took me 6 months to make enough money to pay it off. XD
-
ha ha... This. I learned about LATAs and area codes and that Canada used the same phone numbering scheme as the US... and was expensive to call, all the time, every day, for hours, for one month. (I then spent the next couple of years learning about outdials and tracking down 800# BBS's that provided outdial, so I could hop across the country for free... many US government systems were available via 800# and with a bit of "strategery", one could dial wherever they wanted through multiple hops of outdials.)
-
I did something similar from my grandma's house once, I had multiple dialup ISP numbers on my laptop and used the wrong one by accident for several hours one weekend, wasn't $500 but it was a sizable amount of money. The long distance call is something few people younger than my generation will have any concept of. I can call anywhere in the country for free and even international calling is cheap today, or free with various VOIP options.
-
I've heard of WIFI modems, but that honestly defeats the point for me. I would prefer to use original modem hardware and tap into the regular POTS system with no telnet involved. I know extending the point where it hits the internet to a telephone adapter might seem a bit stupid, but it's just my preference.
-
"regular POTS system" is a dying breed. Many local telcos don't even know how to order a copper line.
-
I've heard of WIFI modems, but that honestly defeats the point for me. I would prefer to use original modem hardware and tap into the regular POTS system with no telnet involved. I know extending the point where it hits the internet to a telephone adapter might seem a bit stupid, but it's just my preference.
Personally I agree, however Im in Aus and via the net I can access BBS's all over the world without paying international call charges, so WiFi modems win it for me.
-
"regular POTS system" is a dying breed. Many local telcos don't even know how to order a copper line.
In Australia, they don't exist anymore. With the conversion over to the National Broadband Network, all legacy services such as ADSL, ISDN and analog phone lines will either need to transition over to the NBN (in the case of POTS lines, all calls will be routed over VoIP) or simply disconnected.
I dare say a similar thing would have been happening in the USA. Even if your direct area isn't impacted, you also need to think about the remote end. If it's still possible, maintaining POTS at either end costs money and while you might find some larger BBSs still have dial-in support, the question remains, "for how long?"
I love the sound and experience of an old modem just as much as the next person (I didn't back then), but for some things, you just need to move on to modern techniques.
-
Eh, it wasn't that expensive of a device, and I definitely have other uses for it. Heck, my modem has very useful voicemail functions, and the number is isolated from the ones I use, so perhaps I could use it as some sort of telephone robot or answering service for future needs. Who knows.
I'm not quite done experimenting, as my successes are quite close, and there is still room to do so, but it's just an attempt to see if I can run this old technology in the modern day.
Btw, I did stumble across a VERY interesting video while just bumbling around on the internet. It's of a very similar sort of thing that I am trying to attempt, but using much much older technology than what I am attempting to use. Very neat indeed, and I suggest you check it out here: https://www.youtube.com/watch?v=X9ctLFYSDfQ (https://www.youtube.com/watch?v=X9ctLFYSDfQ)
My next, or perhaps current? Project is a tad newer. More than a tad actually, it's an AMD Athlon 64 FX based computer on the nForce 4 Socket 939 platform. I put the specifications down here, but safe to say I have had two DOA motherboards in a row, and my third one is taking 10 days to get here (coming within about 12 hours of this post). Hopefully it's interesting, but some people tend to have a varying definition of the word "vintage" even though it's been talked on here to just mean from a certain year (like wine).
-
The video reminded me of something from my high school days. The computer lab was Apple II's, some actual Apples and some of those black Bell and Howell models. We had an ordinary dot matrix printer, an Epson or similar, but there was also one leftover relic, a Teletype Model 43, with an acoustic coupler, and there was a live telephone right there. They also had it hooked up to one of the Apple IIs as an extra printer.
First up, this model still had a physical bell that rang with Control-G. So I wrote a convenient program that mimicked the ringing of a telephone. When we had a substitute teacher, it was basically free lab day, so I set this up on that machine, and then moved over to a different one. I had a time delay built in so it wasn't obvious. Ring ring, ring ring... a few rings then it would stop. Sub came in, picked up the 'ringing' phone - nothing but dialtone. Few minutes later - repeat process.
But, relevant to this thread, we found out the phone line was a dedicated POTS line, didn't go through the main office switchboard, so we could call out or do whatever. So my friend and I would spend a lot of time on some local BBSes. Several boxes of 11x17 greenbar fanfold paper later... For whatever reason I kept much of this as a log of what I was doing, not that any of it was really earth-shattering, but I had all these long printouts for a while before i finally tossed it all. At home I was getting on the same boards with my TRS-80 Model 4P, and in fact had one of the modems shown in the video, direct connect to the phone line but manual answer/originate and no dialing capability. Another friend of mine actually modified one of them so that it could auto-answer, he had TRS-80 Model 1 and write, in all assembly, a purely memory-based BBS system - all message threads were completely stored in RAM, no disk files whatsoever. In those days, 300 bps wasn;t necessarily the limitation on communications speed - except on his board, with everything in RAM and the code being assembled machine code, it could send a steady 300bps with no delays.
Wish I still had all my code - I wrote several systems in BASIC on my 4P, including one that did matchmaking based on answers to a range of questions each new user had filled out - long before Match.com or any of that stuff. If I still had the disks and code, I coulda sued them LOL.
-
Good times. Computers weren't nearly as useful back then but they sure were a lot more fun. Talk of the Apple II's and dot matrix printers reminds me of using Print Shop on the ones in my elementary school classroom to print out huge long banners on fanfold paper. That's something that seems to have virtually disappeared, despite there being no real practical reason that a modern inkjet printer or even a laser printer couldn't print from a continuous roll of paper.
-
Good times. Computers weren't nearly as useful back then but they sure were a lot more fun. Talk of the Apple II's and dot matrix printers reminds me of using Print Shop on the ones in my elementary school classroom to print out huge long banners on fanfold paper. That's something that seems to have virtually disappeared, despite there being no real practical reason that a modern inkjet printer or even a laser printer couldn't print from a continuous roll of paper.
I actually have a dot matrix printer and could do that if I wanted.
Would look like crap by today's standards, but the point is I could make a greyscale banner with it.
I was actually wondering how practical it would be to hack the printer so it could use colour ribbons. Might be an interesting idea for me to try out at a future date.
-
The first web browser was a command line program called "www" which ran on Unix. It was written by Tim Berners-Lee (timbl) at CERN for NeXt and Unix.
It didn't support inline graphics, it was 100% text. There were several text based browsers before inline graphics came along. (Lynx was another, which is still around)
At the beginning they were rarely used over modems because at first the Web was confined to the research community who tended to be in academic, government labs, or corporate settings, The first browser to support inline graphics was probably XMosaic, which was written by Marc Andreesen (marca) at the National Center for Supercomputing Applications, or NCSA. It ran on Unix workstations. The NCSA web server which later on became "a patchy server" (Apache) was written by Rob McCool.
At the beginning it was difficult to get a TCP-IP connection to the net. People who had them were very lucky.
-
Good times. Computers weren't nearly as useful back then but they sure were a lot more fun. Talk of the Apple II's and dot matrix printers reminds me of using Print Shop on the ones in my elementary school classroom to print out huge long banners on fanfold paper. That's something that seems to have virtually disappeared, despite there being no real practical reason that a modern inkjet printer or even a laser printer couldn't print from a continuous roll of paper.
I actually have a dot matrix printer and could do that if I wanted.
Would look like crap by today's standards, but the point is I could make a greyscale banner with it.
I was actually wondering how practical it would be to hack the printer so it could use colour ribbons. Might be an interesting idea for me to try out at a future date.
I've still got 2, one is even colour! I even designed and built an adapter so I could use them again (and LPT sound cards) on my Tandy 1000 EX. :-DD
(https://i.imgur.com/UuukCsqm.jpg)
https://youtu.be/rYrEzZlc1iI
Btw, fanfold tractor feed paper is horrendously expensive now. :'(
-
Good times. Computers weren't nearly as useful back then but they sure were a lot more fun. Talk of the Apple II's and dot matrix printers reminds me of using Print Shop on the ones in my elementary school classroom to print out huge long banners on fanfold paper. That's something that seems to have virtually disappeared, despite there being no real practical reason that a modern inkjet printer or even a laser printer couldn't print from a continuous roll of paper.
I actually have a dot matrix printer and could do that if I wanted.
Would look like crap by today's standards, but the point is I could make a greyscale banner with it.
I was actually wondering how practical it would be to hack the printer so it could use colour ribbons. Might be an interesting idea for me to try out at a future date.
I’ve still got an old my old, really, really, really nice Panasonic Dot Matrix Printer in storage somewhere. I need to pull it out and get it working. It’s 24-pin with a color conversion kit. The kit was an official Panasonic add-on that basically consisted of a 4-way solenoid that installed under the front of the ribbon cartridge holder (on this model, the ribbon and print head move together). The color ribbon had red, green, blue and black strips one on top of another and the solenoid would lift the ribbon so the appropriate color would be struck by the pins each pass. (Color printing took four times as long, because it would have to do each line up to four times!)
It was crude, but colored text and line art came out pretty darn nice! This was at a time when color bubble jets still cost thousands of dollars, so bringing in color printed homework (with images from Encarta or Brittanica on CD-ROM) was impressive as hell to my teachers.
Man, I loved that printer. It was soooo much quieter than the old Epson 9-pin it replaced. That thing was like a freaking machine gun!
-
I have an NEC something or other AND a box full of paper sitting here under my desk - it goes with the PC-8801 I have set up on top of my desk.
Perhaps some who owned a TRS-80 of some sort (the Z80 based ones, not the Color Computer or later MSDOS machines) may remember a program they had called DOTPLOT. Since the graphics resolution of the basic TRS-80 (unless you installed the hi-res board they had for the Model III or Model 4 (one thing I wish I had but never did), this program allowed you to use a dot matrix printer as a high resolution 'display' to plot out equations and systems of equations. It made some pretty awesome printouts, but talk about noise - and since it always used bold, it ran through ribbons rather quickly.
-
I have a Commodore dot matrix printer but the ribbon needs replacing. I've been meaning to do it. I'm guessing new "original" ribbons would be hard to find. I've heard of "refilling" them yourself, has anyone tried this?
-
My 6th grade teacher had a little gadget for re-inking printer ribbons. I don't recall how it worked exactly but I think it just ran the ribbon through a little puddle of ink as you cranked it.
-
I remember doing that by hand, the kit came with a plastic 'glove' so you wouldn't get ink all over your hands.