Question about using Cat5e for serial communication 4

[PPettit]

I have a feeling that I'm about to get into a fight with my phone system people.

We've got a phone switch (ESI IVX 72eII) with a DB9 serial port to allow for programming it via a PC. A while back, we had them replace the main board on this unit because I could no longer communicate with it. Something had damaged the serial port on my switch. They replaced the mother board under warranty on that occasion, but stated that they may not do so again because my cable was too long (something like 60 to 65 feet instead of the usual max of 50 feet) and they felt that this was probably the cause.

My cable is made from from CAT5e with modular boots (RJ45 8P8C to DB9) on either end. I've used a low-end (mine) and a high-end (theirs) cable tester and neither one showed a communication problem. If Wikipedia info (

http://en.wikipedia.org/wiki/Serial_cable)

can be trusted, I should be able to use a much longer cable without any trouble.

If for example UTP CAT-5 cable is used with a typical capacitance of 17 pF/ft, the maximum allowed cable length is 147 feet.

I think my cable is actually 14 pF/ft. It's from Home Depot (Cat No. 2133458H) which doesn't provide complete specs. I found another webpage that said it's the same as General Cable item number 5133200e.

What do you guys think? Barring the possibility of a short in the wire somewhere, could my cable be responsible for damaging the serial port on my switch? I'm almost positive it wasn't the PC's I've used because they've never damaged any of the routers and modems I've connected to them.

 

[edfair]

No, you are dealing with a finger pointing technician.

RS232 serial communications deals with shifting signal levels. There is nothing within those signal levels that can hurt anything on the board.

The only thing that is affected by length is the rounding of the signal from end to end. What starts as square wave type signals tends to become rounded on the far end and affects the starting point for counting on the USART. When it starts affecting the signal bad enough you get communications problems, garbling and such, then when even worse you lose communications completely. But this rounding affects the line receiver chip only. When the voltage reaches the flip point, it flips.

The more likely issue is induced voltage from a lightning strike. That is affected by the length of the cable. You get a strike nearby and it fries the linedriver and gets into the USART.

If this becomes an ongoing battle you may want to convert the signalling to 4 wire current flow with adapters at both ends. That resolves the length issue and gives additional spike voltage protection.

Ed Fair
Give the wrong symptoms, get the wrong solutions.

 

[ravnz5]

I have installed many IVX's in the Maryland area and I have never had a problem with a 9 pin serial connection from the KSU to a pc, even at longer lengths. I have used them both for programming and for SMDR purposes. The length you are speaking of is a reasonable, (do-able). The ESI reseller that talked the manufacturer into swapping out the mother board based on the information given here doesn't make sense to me. I have been installing ESI equipment since 1989, and I can tell you that when working under warranty, their tech department is one of the best in the country. I will look at the reseller site and see if there are any known issues associated with the serial connections and re-post.

We all need a little help once in awhile. Tony the Phoneman..

 

[ravnz5]

I have checked all "Known Issues" on Esi's reseller website and I did not see anything related to serial connections. You can go to their site as a customer and request a call from someone about your problem. I feel that is a better route than getting into a quarrell with your vendor. If all else fails, temporarily hook up a laptop at the system with a factory cord to make sure everything is working fine.

We all need a little help once in awhile. Tony the Phoneman..

 

[robertjo24]

Sounds like your phone system vendor is looking for a reason to not have to pay for the support.
They are full of "good 'ol sandwich fixin's" if you ask me.

 

[PPettit]

Thanks for replying, guys.

ravnz, for the sake of brevity, I left out most of the details about the previous service call. It's kind of a long story. What it boils down to is that the tech knew how to program an operational system, but he definitely did not know much about serial communication or proper troubleshooting procedures.

I have a strong suspicion that the first call was due to the lightning conditions that edfair mentioned. The current problem may be the same thing. However, it's hard to tell for sure because even though we've had some storms over the past 5 months, I haven't had to work on the switch since right after the first call was resolved.

I called my vendor about the new issue. They were going to do some research on it and get back to me. I have my doubts that they'll replace the motherboard again, but we'll just have to wait and see what happens.

 

[dagwoodsystems]

This business about serial connections being limited to 50 feet is urban legend.

EIA-232 specs talk about max impedence, capacitance and such...but nothing about cable length. True, a voice grade cable is just within characteristics at 50 feet. But a CAT5e cable should be within the identical characteristics at 1000 feet.

Short story with a similar ending: I installed some serial terminals at a dental office. Weeks later, the equipment manufacturer said that the serial card interfacing those terminals had blown. The cards were replaced and subsequently blew again. I was told that a nearby radio station might be responsible. I replaced my cable with a shielded variety, and grounded the shield on one end only (as you should). The serial cards blew again.

Cause #1: Turns out that the manufactuer of the dumb terminals had Signal Ground and Chassis Ground tied together. Talk about dumb!

Cause #2: The electrical contractor "grounded" the outlets by tying the ground to the junction box itself, rather than running it all the way to the subpanel. The flex running to the junction box acted like an antenna, sending the resulting power back up the Chassis Ground--and therefore the Signal Ground--into the serial card. Readings at the serial cards were nearly 18 volts...nearly three times what was expected! Once the electrician fixed the problem, the serial cards quit burning up.

Bottom line? Use a meter to check the voltage at the serial card connector. That's the surest way to determine if your cable has picked up voltage though it's antenna-like properties.

Tim Alberstein

http://www.dagwoodsystems.com

 

[strmwalker]

PPettit...my two cents worth i agree with all who has
given thier professional advise.one thing they do make
the db9 and db25 connectors for surge protection in case
a foreign voltage spikes it is the vendors responsiblity
to protect their equipment as well as yours i would have
them to do a grounding audit to make sure their equipment
is properly grounded.

no problems only solutions

strmwalker

 

[mforrence]

As others have alluded, a serial connection uses a signal with reference to ground - often causing a ground loop between KSU and your PC. A surge/spike may use the serial port as the best path to ground. One relatively inexpensive solution is to use an optical isolator - essentially a small in-line device that converts signal from electrical to fiber and back to electrical within a 2" space. This isolates one end from the other and prevents such issues. Someone like B&B Electronics has these.
Mike

 

[buitenhek]

I agree. I had a JiffyLube in San Diego that was blowing A-to-D cards regularly when I found an electrical ground loop problem. Isolate and all will be great.

Regards
Peter Buitenhek
ProfitDeveloper.com

 

[wires]

Optical isolators are a great fix for this sort of problem. I used them on a terminal based system with 300' runs over silver satin with great results.

There are two other solutions I can think of that may be better.

1. If the line is only used occasionally for programming simply unplug it at the phone switch end when not in use.

2. Convert the serial connection to TCP/IP with a serial device server. These come in many flavors and prices. For what you want to do a Lantronix UDS1100 would be a good fit. If you need a secure connection then the SDS1100 is available for a bit more. Both devices are less than $200 new and are a more modern way to accomplish serial communications compared to a dedicated wire.

 

[PPettit]

Thanks for the replies, everyone. I'm definitely learning a bit more than I thought I would.

I still haven't heard anything more from my vendor.

An optical isolator seems like the perfect solution in this case. I didn't even know of their existence until now. How do the port-powered/non-port-powered devices compare? Is there any benefit to using one kind instead of the other?

I like the idea of a serial device server, also. It seems a bit more useful/flexible in general. It's what I should have used from the start.

Thanks again for the info and suggestions.

 

[edfair]

It has been a long time since I've used them but if I recall correctly the powered ones were more reliable. They pulled and used more power from the AC than they could get from the serial port. IIRC they used 12v to shove signals down the line and the serial ports were providing 9v to the unpowered ones.

 

[PPettit]

Could someone tell me where I can get my hands on a DB9F to DB9M powered optical isolator? I remember finding a couple of them after edfair's last post but I can't seem to find them again. Either I'm using the wrong search terms or I'm looking in the wrong places.

 

[wires]

http://www.bb-elec.com/product_family.asp?FamilyId=1

 

[PPettit]

Thanks, wires.

 

[edfair]

google for "rs232 current loop converter".

Ed Fair
Give the wrong symptoms, get the wrong solutions.

 

[docjohn52]

Your problem is he system is getting old, two capicitors on eith side of the rs232 chip are drying out and need replaced

RS 232 is defined as 9 to 12 volts dc, and the caps develope that voltage and deliver to the chip... not the wire...

I have done this repair to a million Okidata printers running serial...

 

[edfair]

Sorry, DocJohn, but your understanding of line drivers and line receivers is flawed.

The signal voltage, data out, shifts between the supply voltages provided by the power supply. It never makes it to the voltage because of resistance, but it tries.

The capacitors are filters to keep the supply voltage steady.

And although the specificatons vary between 9 and 15 volts input (depending on manufacturer) the specifications call for 5 to 25 volts. There are companies that implement it with the 5 volts on one or both of the inputs. It doesn't matter as the receiver on the other end flips output when the signal crosses 0v.

Ed Fair
Give the wrong symptoms, get the wrong solutions.

 

[dagwoodsystems]

May I suggest this conversation be continued in a new thread? We're now way off the original topic and poor PPettit continues to receive notifications while you guys sling mud about capacitors. Thanks.

Tim Alberstein

http://www.dagwoodsystems.com