|
|
|
*************************************************************************
How to upgrade your 24oo modem into a 96oo baud. (US Robotics compatible
of course)
Written By
Death Bringer (CPT/Info-Net/IRA)
06/7/91
Be sure to turn off your modem before you begin.
First of all, you need to go to your nearby electronics store, don't try
to go to Radio Shack, their parts have their own model numbers and you'll
have a hell of a time getting the parts you need. This upgrading of your
modem will not work on old hayes modems, but they will work on any clone
maker of modems (everex, practical peripherals, anchor, ati, intel, etc.)
It will only work with modems with the Intel Chipset.
This technique widens the band width by using a different chip that is
similiar to the chip already in your modem.
Parts Needed:
16550 UART
L4313545 IC
SC11020CN IC
L8730183 IC
First of all, check to see if your biggest chips are soldered or socketed.
if they are soldered, you'll have to de-solder the chips before you begin.
After you get that done, replace the 8250 UART with the 16550 UART. (16550
UARTs are required for 9600+ speeds). The other 1 or 2 chip(s) need(s)
to replaced with INTEL L4313545 instead of the chip that is already there.
If your modem has another socket or more, than you are in luck, you can
purchase a chip from US Robotics, and make your modem faster than 9600.
If you don't have the socket, than you can only go 9600. I'll get the
chip number to you in the next edition and explain how and where to
install it. I am working on a way for the people (like me) that do not have
the extra socket in their modem to speed up the rates. I am working on a way
to piggy back the chip. I'll get back to you. There should be a 22 pin
chip on your modem also with a model number of SC11005CN or some where close
to it. Replace it with the SC11020CN. Also, replace the L8630173 with the
L8730183 chip. If you have any suggestions or questions, you can locate me
through Compuserve USER ID 74702,97524
*****OPTIONAL, LAP-M Error correction*****
Adding error correction is rather simple. For this these parts are
necessary:
Potentiometer. This is a 5k audio taper variable resistor.
Capacitor. Any non-polarized 1.0 to 1.5 uf cap should do.
100 ohm resistor - quarter or half watt.
Wire
Solder, soldering iron, etc.
Solder one end of the capacitor to PIN 1 of the phone line input jack
|-------------------|
| \ line in ::
| \ phone in ::
|-----------||||||||| pin 1^
If your modem is external, the diagram looks like this.
|------------------|
| /|
| ||RS-232 port
| \| ::
| |-|line in ::
| |-| pin 1^
| |-|phone in
| |-|
|------------------|
This should be on the back of the modem.
Pin one should the farthest on the bottom left.
Solder the other end of the capacitor to the center lug of the potentiometer
(there are three lugs on this critter). Solder one end of the resistor to the
PIN 4 of the line input jack. Solder the other end of the resistor
to either one of the remaining outside lugs of the potentiometer. Doesn't
matter which one.
-------------------------------------------------------------------------
ADDITION TO ORIGNAL FILE - 6/15/91
Bill McCauley & Dark Spyre/Dr. Brains/Ryan Schwartz/Lamer/Loser
First, a personal recomendation. _THIS WORKS!!!_ I have been plagued with
2400 baud for some time. I hate waiting for the transfer to finish so that
i can use the phone. Threw the gismo together in about 10 or 15 minutes,
took another five to adjust the pot for best results on my worst
conection, and guess what? No more slow connections!
-------------------------------------------------------------------------
Have Fun!
I am working on cheap ways to install V.32/V.42bis on your modified modem
*************************Updated, 6/16/91********************************
Installing V.42bis:
1) You will need to purchase a 27C512 ROM chip.
2) You will need to purchase a Sony CXK58257P-12L (or equivalent) Ram chip.
3) Install the Rom in the Supervisor (SUP) socket.
4) Install the Ram in the Ram socket.
5) Make up two 3 pin jumpers for P9 and P10 with pins 1 and 2 jumpered
on one and pins 2 and 3 jumpered on the other. Connect the vacant pin 1
on the one to the vacant pin 3 on the other with a jumper wire.
6) Remove the jumpers on P9 and P10.
7) Install the jumper with pins 1 and 2 connected together on P9.
8) Connect the jumper with pins 2 and 3 connected together on P10.
DeathBringer '91
40 responses total.
Don't -- I repeat -- don't do this. It will not work, and you will ruin a perfectly good modem in the process.
It sounds perfectly bizarre, to me. Using *which* Intel chip set? Using *which* revision of that chip? The most absurd thing in the above item is the implicit assumption that any modem using the Intel chips will have its PC board identical to this! Now, I have no doubt this works on a specific modem with a specific rev of the modem chip, but to generalize like this is either stupid, or perhaps malicious?
(by the original author, we assume, not Mike.) Did you give this a shot, Mike?
Let's have a show of hands. How many of you would trust their 2400-baud modems to a guy with a pseudo like Death Bringer?
Hah! I forgot about the name--thats a good point...
Let's see here... I got 1,2,3 2400 baud modems. Humm. One of em's got to hahe the right Intel chippies in it...
Well, it won't work on ANY modem. Let's take, as an example, the fact that it tells you to buy an EEPROM and replace your current EEPROM with it. What should you program it with? Any random sequence of bytes? This was making the rounds on Usenet a few months ago, and several knowledgable people -- including Toby Nixon, chief engineer at Hayes -- have stated that there's no way, no how, that this will work. Combine that with the fact that this guy's CompuServe address is bogus, and I don't think I'd trust these instructions farther than I can throw them.
If the CIS address is bogus, that's a pointer. I passed this info to one of the internet lists I am on, and was very embarresed when I saw the responses here, of course I sent immediately another mail saying don't do it. If intel has a modem chipset as described, it is possible that you can buy a different ROM from intel to implement different compression algorithms. The person may not know the electronic terminology well enough to distinguish EEPROM and ROM. Who knows? Alltogether it looks pretty unreliable. I am realy embarressed for being that stupid.
Don't be--its a wonderful idea in theory, no? Being able to upgrade like that, so easily...
I like the 27C512 part. "See how well it runs with all 1's."
The capacitor/pot setup, BTW, isn't V.42, but it *is* a way of supressing some noise on the line. It works by absorbing spikes and such in the telephone line, though, not by doing actual error-detection and -correction between the two modems.
Gee, let me dig up my instructions for the 33MHz Vic-20...
(:)
You might be interested to know that the supposed "this works" response in that file was pulled from a file for a line noise filter and modified slightly.
Just go buy a new modem, it's a heck of a lot easier.
Yeah, I made the little thing to supress line noise. Helped a bunch at 2400 back home where we have lot's of wire in the house. I think the Cap compensates for inductive loading in the non-twisted pairs. I find the little box to be a *problem* beyond 2400 however. It clamps the voltage swings too much to talk fast. But, a couple passive devices on the wire sure are not going to send 10 or 11 bits in a trellis code for every 8 bits you want to send. :)
A tad old, eh? Im suprized this wasnt archived my slowest modem is a 14.4 right now, and DSL and Cable make this whole topic absurb
So, how much does DSL and Cable modem service cost these days? Last I heard just basic cable TV runs $360 a year. How much more for cable mode? How about DSL?
Since this is kinda a modem item...are PCMCIA modem cables interchangable? That is, will such a modem cable for one make of modem card work with another make?
Re #18: I've seen DSL advertised for around $50/month, lately. Cable modem service seems to run $10-$20 cheaper, but you have to buy cable on top of that. A lot no doubt depends on your local cable and telephone monopolies, though.
Some PCMCIA modem cables are interchangeable, but most are not. One of the earlier cards (intel) came with a strange curvey-Y shape plug, and there were at least 3 makers using that connector. I don't think anyone uses it anymore though. Some modern cars take regular modular jack cables directly using a flip-out connector, but those tend to be rather flimsy. So far as I can tell, the PCMCIA makers would rather you bought a new modem than a replacement cable.
I have a modem with a pop-out connector. (A Megahertz XJACK 28.8...which I've upgraded to 33.6 with a firmware upgrade.) It's great for occasional use, which is what I wanted it for, because there's no cable to lose. I agree it'd be a bit flimsy if you were using it constantly, every day, though -- it looks like tripping over the phone cord and tugging it *just* right would snap the whole thing right off. The other potential problem is that, since the phone cord plugs in vertically, and the plug end protrudes below the card, it interferes with any other card that has a cable connection. I can't easily have both it and my ethernet card connected at the same time. Again, for me this isn't a problem, though. Recently I've started seeing 10baseT ethernet cards with the same kind of connector. Given the stiffer and heavier nature of 10baseT cable, this strikes me as a terrible idea.
Then to be more specific, I have an Optima 144 PMCIA modem card with no cord. There is a 3COM cord available at a web site. Match?..or Miss?
Given that only two of the wires in the whole connector matter, and that you know how to solder, I'd say that any cable which has the right connector ought to be sufficient.
There's the rub....the right connector. If I could see it, or even try to match it to the card, fine, but I can't, so I was hoping one of you would know. Maybe you could tell me the pinout on the Optima card, so I can just poke wires into the right ones?
The connectors I've seen generally have 5-6 wires. I think at least some of them are designed to work with cellular phones as well (exactly how this works, I've never understood.)
Thinking about it further, it seems to me odd that the phone line would plug directly into a PCMCIA card, because the on-hook voltage on the line is ca. 50 volts, and the ringing signal is ca 90 volts AC, 30 Hz. I wonder whether the connector cables have some passive elements in them to adapt the card to these kinds of signals. Of course, only two (2) wires go to the phone hookup, yet the connectors have either four (4) or fifteen (15) pins.
Seems like it ought to be cheaper to put that stuff on the card where it can be part of a chipset or something.
From a signal-level standpoint, but for 90 V AC? But, you may be right. So, anyone have a pcmcia modem, who can test their modem cable to see what the pinout is? Not that it would be what I need (unless you have a 2x 15 pin Hayes pcmcia card), but the question of whether it is straight through or not could be tested with an ohmmeter.
Well, I'm sure my card deals with the 90V AC on board, because it has one of those little XJACK things. There isn't anywhere else for it to go. ;>
Good point - for more recent cards with XJACK - but maybe they were introduced after cards were beefed up. Why wouldn't they have started with simple RJ-11 (etc) jacks right from the beginning? Why a special cord?
No one had thought of the XJACK idea, yet. And as I mentioned earlier, the XJACKs are probably flimsier.
Modular phone plugs (RJ-11, etc) have been around for very much longer than PCMCIA cards. They are really trivial devices. I seem absolutely no reason why they would not have been the first connector thought of when a modem was reduced in size to a card. All you have to is bring out a short wire with a RJ jack on the end - a pigtail. Instead, they designed (another) specialty connector and cable. Of course, one must always be suspicious that that was done because, even though the card stays in the computer, cables get lost, so users would have to buy them regularly...a planned "obsolescence" (?).
PCMCIA cards have a specified form factor - a permamently attached cord might have violated that spec? Certainly it would make it harder to pack the card in those plastic cases they usually supply with the card.
Also, cords have a tendancy to get bent and flexed at attachment points and suffer opens and shorts as a result. Maybe they were actually trying to be helpful -- letting you replace a $5 cord instead of a $70 modem? I've already replaced the cord on my PCMCIA ethernet card, not because I lost the old one, but because the old one broke. Most of these modems are also meant to connect not just to a wall jack, but also to a cell phone. That requires a different cable and connector -- in fact, a different one for *each brand of phone*, since they were never standardized. On my XJACK modem, they actually went to the trouble of providing the XJACK for an RJ11 connection, and a seperate socket for a cell phone cable. On your modem they probably saved money by making one socket do double-duty.
(On the high voltage chip thing... sure! I used to have some Burr-Brown catalogs with single chip isolation amplifiers and DC-DC converter which could withstand hundreds of volts between isolation sections.
Well, that's good: before PCMCIA cards came out? Replacement cords cost $ 25, not $ 5.
Well, the point still stands...that's only half the price of a new modem. ;> (I think the cable for my ethernet card was $9.)
They've had opto-isolators for years; the only thing impressive about squishing it onto a pcmcia card is the form factor, and even that is dwarfed by any lsi logic present. With phones, the high voltage is really only present and used for ring, and even then the current is relatively modest. For the useful signal, the voltage & amplitude (and hence the power) is negligible.
|
|
|
- Backtalk version 1.3.30 - Copyright 1996-2006, Jan Wolter and Steve Weiss