Proper way to joint two AWG 28 wire together ?

You really need to think this thing out. if the signals are digital, think the I2C bus.

One thing for sure, 28 AWG is too small. Krone might work for you, but not 110, but only if you can make the wires thicker.

I did a very unique wiring method for a rack for analog I/O, digital I/O and instrumentation. One instrument has digital, analog in and analog out and the cards were analog in, analog out and digital.

i took this enclosure: https://lmbheeger.com/uni-pacuniversalpackagingsystem.aspx without rack handles because they are ugly. I added my own and installed drawer slides. The top cover was never fastened down. I did one magnetically.
I replaced the back with two 1/4" stainless rods with a threaded hole in the ends. I replaced the back panel with the two rods.

I could bring an instrument into one box as cables, strain relieved to the rods. That instrumentmight go from a 9-pin D to a9-pin-D breakout board. Some would go to digital, some analog in and some analog out.

Analog in was a big mux. Analog out came from 3 cards as ribbons. Digital came from multiple cards as large ribbons. It worked out really really nicely and easy to troubleshoot.

There was a larger case that held a backplane of signal conditioners and power and I only needed one isolated conditioner.

Again, everything came from their native instrument as a connector to connector. The connector was terminated into a connector breakout with screw terminals. Wires could exit as cables with connectors if need be. Point is, I didn't have to drill chassis specific holes.

I objected to the "powers that be" in having computer based cards. This initially was a Macintosh NuBus machine with 5 or six slots. Newer macs had 3 and later an entirely different bus. "They" had to replace obsolete technology with obsolete technology. It wasn't my problem. "They" said cards in the computer. I objected. "They" won.

In this design, I also put rack rails on Unistrut which was the framing for the system.
 

Thread Starter

meowsoft

Joined Feb 27, 2021
367
You really need to think this thing out. if the signals are digital, think the I2C bus.

One thing for sure, 28 AWG is too small. Krone might work for you, but not 110, but only if you can make the wires thicker.

I did a very unique wiring method for a rack for analog I/O, digital I/O and instrumentation. One instrument has digital, analog in and analog out and the cards were analog in, analog out and digital.

i took this enclosure: https://lmbheeger.com/uni-pacuniversalpackagingsystem.aspx without rack handles because they are ugly. I added my own and installed drawer slides. The top cover was never fastened down. I did one magnetically.
I replaced the back with two 1/4" stainless rods with a threaded hole in the ends. I replaced the back panel with the two rods.

I could bring an instrument into one box as cables, strain relieved to the rods. That instrumentmight go from a 9-pin D to a9-pin-D breakout board. Some would go to digital, some analog in and some analog out.

Analog in was a big mux. Analog out came from 3 cards as ribbons. Digital came from multiple cards as large ribbons. It worked out really really nicely and easy to troubleshoot.

There was a larger case that held a backplane of signal conditioners and power and I only needed one isolated conditioner.

Again, everything came from their native instrument as a connector to connector. The connector was terminated into a connector breakout with screw terminals. Wires could exit as cables with connectors if need be. Point is, I didn't have to drill chassis specific holes.

I objected to the "powers that be" in having computer based cards. This initially was a Macintosh NuBus machine with 5 or six slots. Newer macs had 3 and later an entirely different bus. "They" had to replace obsolete technology with obsolete technology. It wasn't my problem. "They" said cards in the computer. I objected. "They" won.

In this design, I also put rack rails on Unistrut which was the framing for the system.
My case is that some sensor is outdoor, if some of sensor is get voltage more than limit, it can burnt micro-controller and another things, if 220v wire is insulation broken and current is leak to 5v sensor wire it can burnt almost all things... if there is voltage spike from 220v that go in 5v relay it can also burnt micro controller... lightning... and something like that...
 

SamR

Joined Mar 19, 2019
3,424
Sounds like a standard I/O problem except that you apparently are trying to throw CAT 5 cable into the mix. Look at Intrinsic Safety Barriers. They are used for standard 4-20 mA instrument circuits leaving a general-purpose control room to enter an explosive atmosphere hazardous area to limit the current to a non-hazardous level through Zener diodes and fuses. We often had 19 individually shielded twisted pair No 22 AWG cables coming into control room racks that had to be terminated and distributed and required safety energy level protection. Screwed terminal strips and punchdown blocks were commonly used and we had large numbers of cables coming from the field that had to be marked for identification, terminated, and distributed to the DCS hardware. Do not mix digital communication and analog sensor/output signals.
 
I really didn't have to go that far. My cables were 16 or 18 AWG solid PTFE Fire Alarm Cable. I did have to pass through firewalls and I did have to energy limit to class 2 wiring. Most were 24VAC or 24VDC contact closures or strobes.

The Hydrogen sensors were a 100 mV bridge configuration and were kept totally separate.

The FAP (Fire Alarm Panel) was an input that I used to turn off active toxic gasses at the gas cylinder.

Low level Hydrogen and toxic gas alarms I acted upon. High level alarms, activated the FAP.

During an experiment, we selected the gasses that were to be used and thus they were turned on by air actuated valves at the cylinder. A fire in a gas cabinet, in theory, would sever the air line which would turn off the gas. The gases were extremy toxic or say one spontaneously combusts in air.

I designed the panel in the 80's with relay logic. It really wasn;t a Safety Instrumented System, but it got the job done and it worked unexpected once when the lab exploded. The gas involved was Hydrogen. We had sensors and a panel for that in a box somewhere. Management deemed it unnecessary to be installed.
 

Thread Starter

meowsoft

Joined Feb 27, 2021
367
Sounds like a standard I/O problem except that you apparently are trying to throw CAT 5 cable into the mix. Look at Intrinsic Safety Barriers. They are used for standard 4-20 mA instrument circuits leaving a general-purpose control room to enter an explosive atmosphere hazardous area to limit the current to a non-hazardous level through Zener diodes and fuses. We often had 19 individually shielded twisted pair No 22 AWG cables coming into control room racks that had to be terminated and distributed and required safety energy level protection. Screwed terminal strips and punchdown blocks were commonly used and we had large numbers of cables coming from the field that had to be marked for identification, terminated, and distributed to the DCS hardware. Do not mix digital communication and analog sensor/output signals.
CAT 5 cable isn't digital, is analog signal, like micro-switch, LDR, PIR, ...
 

Thread Starter

meowsoft

Joined Feb 27, 2021
367
CAT 5 cable isn't digital, is analog signal, like micro-switch, LDR, PIR, ...
That wire is for smart office project not for industry control devices... example : smart door control (open/close with fingerprint, RFID, ...), smart lighting control (on/off lamp when human present is detected by PIR sensor, radar sensor, ...), and another such thing...
 

Thread Starter

meowsoft

Joined Feb 27, 2021
367
But that cable have same cable duct with 220v cable even some signal cable is inside same conduit with 2kV HVAC cable, can all micro-controller be burnt if 2kV current leak to 1.5v signal wire ?
 

Yaakov

Joined Jan 27, 2019
3,187
Can't because micro-controller is centralized in one place, by consumer requirement and contract
The independent sensor boards would still communicate with the central MCU just not as a home run for each. They would use a bus, like RS-422 or RS-485 to send their data there. Some might still be home runs if that’s better for a particular application but it would reduce the number of conductors if you used a bus and you can use a loop to wire things making everything easier.
 
Maybe you need to consider designing standalone sensor boards that can send cooked data via a bus instead of home runs for each.

The pieces we got, sounds like it belongs in an addressable fire alarm or it's the control system for a haunted house

I built a model gantry crane with a farmer that had a business. He dripped solder until I showed him resistance soldering.. He could not wrap around his head why I needed the gantry in order to automate it? Like,I needed the motors. He wanted to power the motors like a rail. He gives me a motor, but it;s only of them. The one in the crane is smaller and the one moving the dock is huge. The changes stuff and doesn't tell me. This was about the time the IBM PV hit the market. I got myself an Amiga 1000;
The first was done using 1802 Assembly language by hand. The second modular boards, and they layout was "taped". The computer was a BASIC SBC. The electronics allowed one to move everything manually without the computer. I could disconnect every card from computer control. I probably didn't have to do the edge connector thing to get remote manual. I didn't have a "go to next X" position, it was just move/brake. If the program ran away and it did, the motors would shutdown when they jammed. Jamming got fixed mechanically in the next design. Design 1 used relays. Design 2, motor drivers, One was too small for the gantry motor. Design 1 used a 10-12 conductor coiled cord. When design 2 showed up, there were sliding things. It would have been nice to know that EARLIER because the crane has electronic limits, but not servo limits. Loosing contact on the rail might cause the crane's position to permanently change. Turned out it was OK. It was found out that the dock had to have two positions, not one. The pegs that moved the blocks had to be behind the blocks when loading to the dock. As the dock moved, they would be in the position for removal by the electromagnet. We found out, on the first design, that it needed another sensor and there wasn't any spares. This sensed blocks, so load and unload were determined by a block sitting on a sensor and the "Go" button. Positions of x&y were a multiplexed priority encoder. Positions were fixed. Forward positions were different than reverse. One segment wasn't sensored. The end of the first when going back and the beginning of the 3rd when going forward. He came up with a dual microswitch arrangment for the first design and the totally changed it to a bad design. The dock was sort of autonomous. You bypassed the switch, waited for it to move and unbypased the switch and it would stop by logic. Processor could be dealing with x&y simultaneously and not worry about the dock.
The blocks could get messed up and that was the real value of the manual control. The oval dock had to move up to 3 times.
best option was to load the dock manually, move it 3x and stop in an unload position.

Doing the problem again today would be a really neat exercise. It might have multiple processors communicating. [ Dock control, switches,[ MAIN{y-gantry], [crane and x gantry]

I started to work on other devices but he passed.
 
Last edited:

Thread Starter

meowsoft

Joined Feb 27, 2021
367
No... sensor is very native... only LDR/PIR and such things... and one micro controller can be connecte to dozens of sensor... also 433mhz radio module is prohibited... because that is hospital office..., we use 50x2x0.8 for each micro-controller, and about 12 micro-controller, so.. total is 1200 x 0.8 wire, run about 2-30 meters per pair... also maintenance is nearly impossible... my fault is I was accept that contract
 

SamR

Joined Mar 19, 2019
3,424
Later new installations often used ladder trays for cable distribution throughout the operating areas. We could not mix power and instrument cabling in the same tray as required by the safety restrictions. Power cabling was for 480VAC motors with 120VAC controls all in the same cable. Instrumentation was 4-20mA intrinsically safe analog signals. You mentioned Hospital so you have safety standards that must be adhered to. Wiring must meet safety requirements for plenum and conduit enclosures as specified by those standards. All those requirements should have been laid out during the initial scope of the project and not as an afterthought. When dealing with safety standards that is the framework that all design and construction must adhere to. Those are your restraints, and all your project design must fit within that limited framework. Part of the initial project design was a multidisciplinary safety review of the projects to set the boundaries of the project scope that had to be adhered to.
 
Last edited:

Thread Starter

meowsoft

Joined Feb 27, 2021
367
Later new installations often used ladder trays for cable distribution throughout the operating areas. We could not mix power and instrument cabling in the same tray as required by the safety restrictions. Power cabling was for 480VAC motors with 120VAC controls all in the same cable. Instrumentation was 4-20mA intrinsically safe analog signals. You mentioned Hospital so you have safety standards that must be adhered to. Wiring must meet safety requirements for plenum and conduit enclosures as specified by those standards. All those requirements should have been laid out during the initial scope of the project and not as an afterthought. When dealing with safety standards that is the framework that all design and construction must adhere to. Those are your restraints, and all your project design must fit within that limited framework. Part of the initial project design was a multidisciplinary safety review of the projects to set the boundaries of the project scope that had to be adhered to.
No... it's in China... not really safe... because this also in office... if in operation room safety is maximized (voltage stabilizer, voltage relay aka under/over voltage relay, phase fault relay, MCCB, MCB, also MCB is doubled with NH fuse, 30mA RCCB, also isolation transformer for each device, also online UPS for each medical device with data center grade (built-in stabilizer, also minimized harmonics and power factor problem and 0.001 switching time and add with external battery(normally 12v 50ah we replace it with 12v 400ah)), and generator set also battery bank, fire control system, door lock access with fingerprint, CCTV with face recognizion and more things)
 

Thread Starter

meowsoft

Joined Feb 27, 2021
367
No... it's in China... not really safe... because this also in office... if in operation room safety is maximized (voltage stabilizer, voltage relay aka under/over voltage relay, phase fault relay, MCCB, MCB, also MCB is doubled with NH fuse, 30mA RCCB, also isolation transformer for each device, also online UPS for each medical device with data center grade (built-in stabilizer, also minimized harmonics and power factor problem and 0.001 switching time and add with external battery(normally 12v 50ah we replace it with 12v 400ah)), and generator set also battery bank, fire control system, door lock access with fingerprint, CCTV with face recognizion and more things)
Even that operating room building can survive if earthquake about 8 magnitude squazed that building
 

Thread Starter

meowsoft

Joined Feb 27, 2021
367
Even that operating room building can survive if earthquake about 8 magnitude squazed that building
And we oversized all conductor capacity by 250% from max continuous load, also all protective ground we used is 240mm², also with 2" copper lightning terminal, Chinese Communist Important Person was served here in this building in about 2008 - 2014 before moved to another central hospital, also this is teaching hospital
 

Yaakov

Joined Jan 27, 2019
3,187
If you are constrained to the home runs, I would say you need to connectorize them. Unfortunately, it sounds like there is no particularly good solution.
 
Top