RS-485 splicing

[mikefamig]

A RS-485 interface uses differential signalling on the data lines.  As such, the voltage difference between the A and B signals is usually on the order of a few volts.  The main concern would be any common mode voltage between each data line and ground, which RS-485 specs in the range of -7V to +12V.   So a clamping voltage of 22V is a reasonable choice. 

RAL
 
 Is there a way for a novice like me to know what is too high a clamping voltage with any certainty? Does a 22 volt clamp guarantee that the Elk is protected from damage? I'm sure that there is no magic number but there must be a threshold that we know will certainly damage the M1 control. This is the gray area that I am trying to clear up.
 
Mike.
 
BTW - Happy New Year!

 

[Del91574]

@ Work, as you stated, application. In the case of a tower, no gel, but in the case of UG, gel. Usually it gets transitioned somewhere if it goes from UG to riser. I'm pretty sure it's the 50' rule also for buildings, but would need to look it up. Usually it's a combination of the gel being possibly flammable and the cabling not being rated the same as indoor or GP cable.
 
@ Mike, think of it this way, Elk may be aware of how the product gets installed and/or their components, but in the specifics, they can't put the manual or documents together to be a "how to" for anyone installing a system. It's like purchasing a desktop computer; the basics of how to connect all the components together are there (even color coded on some) but if you want to get it on your network, DNS it and the like, there are some other steps involved not covered in the basic windows configuration. They may state you need to connect to your ISP, but say you have multiple computers and a single modem? Does your ISP or computer vendor tell you that you must install and configure a router behind the modem or install a dumb switch?
 
Some manuals and docs spell out certain items, such as cable ratings that must be used, types, etc. and others use boilerplate terms (meet applicable code, etc.) but a install manual can never be used or classified as a system design guide. Elk did, however, change in the years and started supplying a surge with every M1 panel that goes out the door for telephone connection, but not one for AC power.....think about that.

 

[mikefamig]

With the info collected here I have a plan that I want to pass by you guys for your approval.
 
I like the scotchcast kit. I plan to splice, solder and heat shrink the flooded cable leads to cmr rated cable leads and encase the splices in a scotchcast enclosure. This splice will still be in the L-box outside the house. Then run the clean indoor cmr cable into the house where it will be attached to the Ditek 4LVLPx surge suppressor just inside the same wall. Then attach the home-run to the M1 to the output ot the Ditek.
 
One question....would you go to the trouble of adding another box to house the Ditek? I'm thinking of attaching it to the exterior of a nearby load center or to the wall up high near the ceiling. It is an unfinished basement. Also any criticism is always welcome.
 
Mike.

 

[drvnbysound]

Are you doing the same on the other side of the cable?
 
Overall, it sounds like a lot of possible failure points to me. How many connection points will be between this cable's end points? At a minimum I foresee the connection point in the garage, the soldered connection, the Ditek (x2), and the connection to the Elk databus. So, at least 5 places this cable could go bad...

 

[mikefamig]

Are you doing the same on the other side of the cable?
 
Overall, it sounds like a lot of possible failure points to me. How many connection points will be between this cable's end points? At a minimum I foresee the connection point in the garage, the soldered connection, the Ditek (x2), and the connection to the Elk databus. So, at least 5 places this cable could go bad...

Yes I plan surge protection at both ends. What do you suggest?

 

[mikefamig]

Are you doing the same on the other side of the cable?
 
Overall, it sounds like a lot of possible failure points to me. How many connection points will be between this cable's end points? At a minimum I foresee the connection point in the garage, the soldered connection, the Ditek (x2), and the connection to the Elk databus. So, at least 5 places this cable could go bad...

I should add that I don't consider a properly spliced  and soldered union as a point of failure. If you make a good mechanical connection with your splice and then solder and heat shrink you will have no more point of failure than the wire itself.
 
Mike.

 

[mikefamig]

BTW guys - I've had a squeaky clean log file and no abnormal behavior since I replaced the rs485 crimp connectors with soldered splices. For a while I thought hat I'd never pinpoint that problem but it looks like  "I" ........."we" finally got it. I've been adding UPB lighting and adding rules and I'm really enjoying the system.
 
Now when the overhead doors go up the entire bay lights up instead of the little bulb in the door opener, the overhead doors are automated, the entire garage lights up on alarm, exterior lights follow the sunlight,  the garage arms itself based on time of day and occupancy detection, I have a single F4 keystroke at the house keypad to arm and disarm the garage, and I have a sump pump call me on the cell phone on a failure.
 
The system still emails and texts me on alarm but I plan to connect with a monitoring system before long, I needed to get the system stable before doing so.
 
I couldn't have done it without this group's help. Ya gotta love the internet!
 
Mike.

 

[Del91574]

Would not put the Ditek on the load center. I'd probably put it either surface just inside or inside/on a box. I doubt highly, assuming the cable and Ditek are inside, the connection at the Ditek developing an issue, physical issues notwithstanding (loosening of connections, etc). If that was the case, then the connections on the M1 would develop the same issues.
 
Assuming a good solder joint, wiped and not a cold joint, the cable will develop issues before the solder joint. The splice is above ground, so not really a huge issue.
 
In the case of the installation of surge protection, splitting the cable is a necessary evil, where it is done is the art.

 

[mikefamig]

In the case of the installation of surge protection, splitting the cable is a necessary evil, where it is done is the art.

About locating the Ditek, I have a question
The install instructions say
 
1 - to allow the surge protector reaction time by providing 3 - 5 feet of wire between the Ditek and the equipment to be protected.
 
2 - to allow for the wire distance from the surge protector to the protected equipment to be longer than the wire distance from the surge protector to ground.
 
Does that mean that the wire from the surge protector to ground can be no longer than 5 feet? Or does it mean "at least" 3 - 5 feet of wire between the Ditek and the equipment to be protected.
 
Mike.

 

[RAL]

In general, the shorter the wire is to ground, the better, but they are not saying that the wire to ground must be shorter than 3-5 feet.
 
MOVs that are typically used in surge protectors respond in about 1ns or less.  Electrons travel through wire at a speed of about 1ns per foot.   So having 3 to 5 feet between the Ditek and the protected equipment would allow enough time for the MOV to conduct and dissipate the surge to ground before anything bad could reach the protected equipment.

 

[mikefamig]

In general, the shorter the wire is to ground, the better, but they are not saying that the wire to ground must be shorter than 3-5 feet.
 
MOVs that are typically used in surge protectors respond in about 1ns or less.  Electrons travel through wire at a speed of about 1ns per foot.   So having 3 to 5 feet between the Ditek and the protected equipment would allow enough time for the MOV to conduct and dissipate the surge to ground before anything bad could reach the protected equipment.

 
 
So the distance between the surge protector and the protected equipment is not limited to 3 - 5 feet but should be at least 3 - 5 feet long and the ground should be shorter than that, correct?. I have about 30 feet between the surge and the M1 and it would be easy to put a ground stake outside for a short ground lead. It makes sense to me that the longer the distance between the Ditek and the M1 the better.
 
Mike.

 

[RAL]

RAL
 
 Is there a way for a novice like me to know what is too high a clamping voltage with any certainty? Does a 22 volt clamp guarantee that the Elk is protected from damage? I'm sure that there is no magic number but there must be a threshold that we know will certainly damage the M1 control. This is the gray area that I am trying to clear up.
 
Mike.
 
BTW - Happy New Year!

 
I wouldn't say there is an easy way to pick the right clamping voltage.  You need to know two things.  One is the maximum voltage that would normally be on the wire you are trying to protect.  The other is the maximum voltage that the device connected to the wire can withstand before it is damaged.
 
In the case of RS-485, the maximum normal operating voltage on the A/B wires can be up to 12V.   The voltage that would damage the device can vary depending on who makes the part and the semiconductor technology that is being used.  That number is tougher to ferret out.
 
TI specs an absolute maximum voltage of 14V for a RS-485 chip that they make.  Now, that doesn't mean that a voltage of 15V will kill the chip, just that TI doesn't guarantee that it won't.   Usually, manufacturers are conservative and spec things quite a bit lower than the level that they know will do damage.  Would this chip survive 22V?  Maybe, but TI won't promise that it will.
 
I'm not sure what part Elk chose for the M1.  It might be something other than the TI part I checked the specs on, so the absolute maximum numbers could be different.
 
So, what's the right voltage to choose?  My preference would be to use something as close to the RS-485 absolute maximum voltage as possible.   The Ditek doesn't look like a perfect match, but comes pretty close.  I think it is an ok choice.
 
Glad to hear your system is working better!
I hope you have a Happy New Year!

 

[mikefamig]

RAL
 
If you choose a clamping voltage that is too close to the operating voltage and the operating voltage rises to equal the clamping voltage, exactly what happens? 
 
I think that with what I learned from you about reaction time and with the relatively long run between my M1 and the surge protector I'm beginning to feel comfortable with the 22 volt clamp.
 
Mike.

 

[RAL]

So the distance between the surge protector and the protected equipment is not limited to 3 - 5 feet but should be at least 3 - 5 feet long and the ground should be shorter than that, correct?. I have about 30 feet between the surge and the M1 and it would be easy to put a ground stake outside for a short ground lead. It makes sense to me that the longer the distance between the Ditek and the M1 the better.
 
Mike.

 
 
Yes, you understand it correctly.   The 3-5 feet minimum length would allow enough time for the surge to dissipate if you had a very short ground wire.  But they are also saying that the wire to the protected equipment should be longer than the wire to ground, so that the time for the surge to travel through the ground wire is less than the time to travel to the protected equipment.
 
If you can, the best thing to do would be to ground the Ditek that's in the house to the same ground point as your electrical panel, rather than to its own ground rod.   When there are multiple earth grounds, a lightning strike can create a large voltage difference between the grounds (because of the resistance of the earth) and that can cause its own set of problems.

 

[RAL]

If you choose a clamping voltage that is too close to the operating voltage and the operating voltage rises to equal the clamping voltage, exactly what happens? 
 
I think that with what I learned from you about reaction time and with the relatively long run between my M1 and the surge protector I'm beginning to feel comfortable with the 22 volt clamp.

 
 
If the clamping voltage is too close to the operating voltage, the MOV may conduct under normal operation, which would do two things.  On a data bus, it would short the data signal to ground, corrupting the data and causing errors. Each time the MOV conducts, it will use up some of its protection ability and service life, and eventually will fail.  And if you do get a real surge before the MOV totally fails, the MOV may no longer be able to dissipate as much as it is supposed to.