Short cut interface

[APL]

I just added a 65c51 to my 65c02 project board. As a bit of a shortcut I considered one of these for the interfacing circuits -

http://www.ebay.com/itm/MAX232-RS232-To ... 1022603873

Then I came across one of these;

http://www.ebay.co.uk/itm/like/40068696 ... 108&ff19=0

which seems to be easier to interfact to the modern PC. Since RS232 ports seem to becoming increasingly rare.

[BigDumbDinosaur]

APL wrote:

Since RS232 ports seem to becoming increasingly rare.

...but are readily added to PC hardware if needed.

Something to consider is that 15 feet is the outer limit for USB cable length. TIA-232, on the other hand, can easily span ten times that distance if a suitable line driver (e.g., the MAX-232 or MAX-238) is used, and TIA-232's big brother, TIA-485, can go out to 4000 feet. Both can be extended to almost any distance with a modem.

Also, USB is a "consumer interface" designed primarily with cameras, phones, etc., in mind, whereas TIA-232 interfaces with an endless variety of devices, such as the magnetic stripe card reader I have here for testing credit card software. You might want to think about it some more before giving up the flexibility of TIA-232.

[cbscpe]

I use a bunch of those

http://www.aliexpress.com/item/1m-PL230 ... 68451.html

cheap and you know in advance they are crap. Why that? They come with a fake prolific chip and above 57600bps they do not always work ( I have reproducable transmission errors at 115200bps).

Most of my project don't use that speed (limited mostly by the 65x51) and use much less than the 500mA that USB can deliver. So they are power-supply and RS-232 in one cable (which true RS-232 cannot provide) for most of my small projects (and even most of my SBC use less than 500mA). The limitation in length is not an issue for me, especially during the build and test phase. And as I have only Mac's, RS-232 can only be added via USB-RS232 adapters, so at the end it would connect to a USB port. The prolific driver on the Mac recognizes them, so they work like a charm. That's why I once bought a couple of them. As it goes for bang for the buck they do an excellent job in my tinker den.

[APL]

BigDumbDinosaur wrote:

...but are readily added to PC hardware if needed.

Agreed, my mini atx has one which I use for the programmer.

For my purposes, it's the 'consumer interface' perspective. I think the usb interface would be fine.

I'd really like to be able to just download new code directly into the project, I'm using an Amtel 28c64 for the firmware and with a little more circuitry ( including write enable ) I think that would be feasable.

Your other points about the distance of RS232 probably don't apply to me as it's unlikely that this project will get off my desktop, so 'extreme' distance isn't a consideration at least at the moment.

BigDumbDinosaur wrote:

You might want to think about it some more before giving up the flexibility of TIA-232.

Given the price of either device and considering that both devices interface on the project side TTL RS232 signals, It'd be neat to just make the connector identical - carrying the main RS232 signals, then I'd be able to swap in either interface. I'm thinking a ten way IDC interface between my project board and these interface devices, should do it..

cbscpe wrote:

cheap and you know in advance they are crap.

Not sure if that's a recomendation

But it's certanly a consideration at the price. I recently got into a dispute with an Ebay vendor that shipped a card with solder spatter all over it, that wasn't so bad except it didn't work either.

[BigDumbDinosaur]

cbscpe wrote:

And as I have only Mac's, RS-232 can only be added via USB-RS232 adapters, so at the end it would connect to a USB port. The prolific driver on the Mac recognizes them, so they work like a charm. That's why I once bought a couple of them. As it goes for bang for the buck they do an excellent job in my tinker den.

It's ironic that the Mac no longer supports TIA-232 or TIA-485, as the Apple-Talk network used TIA-422. Equally ironic is the "universal" monicker in USB's name, as it is nowhere as universal as TIA-232.

———————————————————
Edit: Apple-Talk ran on TIA-422, not 485.

[HansO]

APL wrote:

I just added a 65c51 to my 65c02 project board. As a bit of a shortcut I considered one of these for the interfacing circuits -

http://www.ebay.com/itm/MAX232-RS232-To ... 1022603873

Then I came across one of these;

http://www.ebay.co.uk/itm/like/40068696 ... 108&ff19=0

which seems to be easier to interfact to the modern PC. Since RS232 ports seem to becoming increasingly rare.

Since serial ports are extinct on all latest computers and USB is everywhere, the conversion from 0-5V to +/- 12-25V and back using DB25 or DB9 bulky connectors makes no sense anymore. Saves a lot in between on cables, interfaces with IC's with charge pumps and the effect is identical.

Go for the USB TTL level serial interfaces if you plan to connect to a PC or Mac.
The PLxxxx types are very cheap ($2.50) and the older ones have no Windows 8 support and most are counterfeit low quality but adequate.
The FTDI types are more expensive ($5), but the better FTDI interfaces bring out all RS232 signals such as RTS/CTS, DTR etc) which make them very versatile.

USB cabling is indeed limited in length to several meters. So is RS232 to be honest, the distances BDD is quoting are way over the specifications, 10 to 15 meters at 9600 baud are (I should scan the original standard EIA document one of these days!) .

[GARTHWILSON]

HansO wrote:

USB cabling is indeed limited in length to several meters. So is RS232 to be honest, the distances BDD is quoting are way over the specifications, 10 to 15 meters at 9600 baud are (I should scan the original standard EIA document one of these days!) .

From http://aplawrence.com/Unixart/serial.art.html:

Quote:

Before we start, let's dispense with some common myths. The most prevalent is the old "RS232 runs have to be less than fifty feet". If this was ever true, it related to equipment in use long before I started playing with this stuff (1983). I have not seen a UART (Universal Asynchronous Receiver-Transmitter) [I assume he meant a line driver & receiver-- GW] that couldn't support 9600 Baud or better for at least several hundred feet over ordinary telephone wire. Most good quality multiport cards can drive 9600 Baud for 20 times 50 feet and never garble a byte. If you think about it from an electrical viewpoint, it's a simple a matter of how much degradation in signal strength there will be over X feet of Y size wire, and just how much of said degradation the devices at each end of the wire can put up with. So forget about 50 foot limitations. Start worrying when you get close to 1,000 feet, but don't be surprised if it works at that distance, too!

And, if necessary, you can slow the data down and go much farther.

[Tor]

Back before my workplace changed to Sun workstations, Unix, TCP/IP, we ran RS232 cabling all through the building from the central computer room (where the minis were located) to every office. I had four coming into my office so that I had a set of computers to connect to (and that worked even better when I got a terminal with four serial ports..). It was probably 9600 bps and not 19200, but my memory isn't certain about that.
No problems with the (significant) distances involved in any case. Earlier than that we used current loop, but RS232 turned out to work just fine so when we moved to a new building it was RS232 only.

However, USB TTL, FTDI and all that _is_ very convenient these days..

-Tor

[Aslak3]

Quote:

And, if necessary, you can slow the data down and go much farther.

On the negative side of the coin when distances are stretched: random dataloss, no standards for checksuming or retransmission of lost data... There's something to be said for protocol stacking, and I'm not aware of any standard for character-wise reliable '232. Might be wrong though. File transfer of course have many standards for reliable transfers.

Of course as a nearly 40 year old computer user I lament the passing of '232 on desktop computers. I can't see it going away any time soon though; it remains (thankfully) engrained in configuring everything from network routers to televisions. Not so for the parallel port, which I won't shed any tears over...

[BigDumbDinosaur]

GARTHWILSON wrote:

From http://aplawrence.com/Unixart/serial.art.html...

Funny, I was going to link the same article. 

In 1997, we did a major system conversion at a large paint manufactory in Chicago.  At the time, everything was host-based on UNIX, using TIA-232 to connect to terminals, printers and several modems.  The company had a separate warehouse that was several blocks away from the plant.  We made a connection to a serial printer in that warehouse over a pair leased from the phone company, using a short-haul modem operating at 14.4Kbps at each end of the line.  The effective length of this run was about 3000 feet and the printer worked great.

The longest hard-wired TIA-232 run was 272 feet, which drove a terminal at 38.4Kbps without any problem.  All of our servers at the time used Equinox's SuperSerial (SST) expandable hardware, which could support up to 512 ports on a single machine.  At no time did we ever experience any TIA-232 anomalies due to cable length.

HansO wrote:

USB cabling is indeed limited in length to several meters. So is RS232 to be honest, the distances BDD is quoting are way over the specifications, 10 to 15 meters at 9600 baud are (I should scan the original standard EIA document one of these days!) .

The original EIA document (which I first read in 1968) has been revised many times and doesn't stipulate cabling characteristics.  When I started regularly working with this stuff, it was already accepted that a reliable connection could be assured at 75 feet using shielded cable.  It was subsequently discovered that much greater distances could be achieved on ordinary CAT3 UTP cable, as the shielding actually degraded signal quality.  The switch to CAT3 UTP cable coincided with the introduction of the 8P8C (aka RJ-45) connector, which eliminated the dependence on the DB-25 connector.

Toward the tail end of the era where we regularly installed serial hardware (up to about 2001), runs of several hundred feet using CAT5 UTP cable were routine.  I know from application experience that 115.2Kbps is readily achievable on CAT5 UTP cable over a distance of 325 feet.  Coincidentally, this is very close to the 100 meter (328 feet) limit of Ethernet.  During the 1990s, when we did a lot of installations, we wired in such a way that a cable run could be used for either serial or Ethernet traffic by simply unplugging one type of hardware and plugging in the other.  I have the serial ports on my POC unit wired so that they can work through T-568(B) wiring.

Correction: in an earlier post I mentioned that Apple-Talk was based on TIA-485.  Actually, it was based on TIA-422, which is very similar to, but not as flexible as, TIA-485.  The latter is widely used in industrial data links, as it is very resistant to noise, plus can operate at several megabits per second.

[BigDumbDinosaur]

Aslak3 wrote:

On the negative side of the coin when distances are stretched: random dataloss, no standards for checksuming or retransmission of lost data... There's something to be said for protocol stacking, and I'm not aware of any standard for character-wise reliable '232. Might be wrong though. File transfer of course have many standards for reliable transfers.

Actually, there are standards for checking for single-character corruption and doing something about it. They're just not used much for terminal and printer applications. In the few cases where we had to extend a serial link more than about 300 feet we used short-haul modems, which use standard error correcting methods like those in telephone modems. This assured error-free transmission over long distances (miles, if necessary).

Quote:

Of course as a nearly 40 year old computer user I lament the passing of '232 on desktop computers. I can't see it going away any time soon though; it remains (thankfully) engrained in configuring everything from network routers to televisions.

You'd be amazed at how much stuff continues to support TIA-232. In my office alone, I have (not counting my POC unit) seven devices that are TIA-232, one of them being a magnetic strip reader (MSR) that is used for testing software that works with credit/debit cards. Another is my Okidata ML395 printer, which I use for printing accounting reports on greenbar paper (old habits die hard...). One of our two servers has an Equinox SST16 serial interface on it. One of the SST's ports is connected to the auxiliary port on POC so I can transfer code to the latter for testing.

Quote:

Not so for the parallel port, which I won't shed any tears over...

What do you have against parallel ports?

[GARTHWILSON]

Quote:

It was subsequently discovered that much greater distances could be achieved on ordinary CAT3 UTP cable, as the shielding actually degraded signal quality.

Was it because of the greater capacitance? If greater drive strength is needed to overcome it, multiple drivers could be parallelled. I know that violates the measures put in place for if someone connects two outputs together, but we're no longer talking about consumer and commercial applications, and we techies have enough usderstanding of the situation to avoid such problems, and to fix any damage in the unlikely case that we do mess it up.

Quote:

On the negative side of the coin when distances are stretched: random dataloss, no standards for checksuming or retransmission of lost data... There's something to be said for protocol stacking, and I'm not aware of any standard for character-wise reliable '232. Might be wrong though.

The commercial world no longer has the support for RS-232 that the industrial world has, so we might be on our own as far as doing things like error recovery. I don't really see that as a problem though, as you can just use an oscilloscope to see what the integrity of a signal is at the end of a long cable. There's no need to work in the dark. A nice thing about it is that you can make up your own cables (unlike USB whose connectors are not hobbyist-friendly), with whatever type of cabling suits the need, even doing things like putting large ferrite beads over twisted pairs if desired. Cat-5 cable is used for long runs at 100MHz. Its characteristic impedance is too low to do that with RS-232, but the principle applies, if one wanted to use a different cable.

Edit: Since I just realized I have nearly a thousand feet of cat-5 cable here and nearly another thousand of RG-59, plus spools of 4-conductor cable intended for audio, alarm, remote control, etc., my interest is getting piqued to do some experiments. It won't be today though. Too much work to do.

Quote:

Not so for the parallel port, which I won't shed any tears over...

The only thing I use the parallel port for is my dot-matrix impact printers using fanfold paper which I prefer for program listings. Even this week I'm printing out strips several feet long as I'm developing code, and I don't want page breaks in the middle of program structures. Other times, I want to print just a line or two at a time and see it, without wasting the rest of a page every time. To my knowledge, there is no laser or inkjet printer that can do these things. For printing stuff I get off the web though, the color laser printer I use at the other end of the house is connected by Ethernet.

[BigDumbDinosaur]

GARTHWILSON wrote:

Quote:

It was subsequently discovered that much greater distances could be achieved on ordinary CAT3 UTP cable, as the shielding actually degraded signal quality.

Was it because of the greater capacitance?

I believe that was the case, although I never undertook any analysis (too much to do and, as usual, not enough time in which to do it). Sometimes shielding was a liability when hookups were made by someone who didn't understand the phenomenon of ground plane potential imbalance. Current would circulate in the shield and inject common mode noise into the signal lines. GPPI could also cause current to circulate on the ground connection between host and terminal/printer/modem, causing the effective signal voltages to fluctuate at twice the line frequency.

Quote:

Quote:

Not so for the parallel port, which I won't shed any tears over...

The only thing I use the parallel port for is my dot-matrix impact printers using fanfold paper which I prefer for program listings. Even this week I'm printing out strips several feet long as I'm developing code, and I don't want page breaks in the middle of program structures. Other times, I want to print just a line or two at a time and see it, without wasting the rest of a page every time. To my knowledge, there is no laser or inkjet printer that can do these things. For printing stuff I get off the web though, the color laser printer I use at the other end of the house is connected by Ethernet.

The programming rig I have for Atmel CPLDs is driven from a parallel port. Any upgrades I might do to my PC will have to keep that in mind.

[HansO]

What as brilliant case of topic drift again!

The original question is about connecting via TTL level or via RS232 level. It ended up via discussions on how long RS232 cabling can become to the extinction of parallel ports!

Only two replies were an attempt to answer the orginal question, the rest was not! Sorry APL!

[BigDumbDinosaur]

HansO wrote:

What as brilliant case of topic drift again!