Test circuit for wire harness

[goatmeat]

I have an R&D project, where the objective is to test multiple conductors on a wire harness with a voltage source. One particular harness has 16 conductors all pin to pin. There are thousands of these harnesses, and I am looking at building a test fixture to test them in the quickest way possible. I have thought about using a PLC, but I am really getting stumped, thinking there must be an easy way. Any Ideas?

 

[Skogsgurra]

There are about 80 000 hits when you google Wire Harness Testing. Start there - it seems to be a very popular pass-time to invent smart strategies for wire harness testing.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[VEBill]

Ideally the cable's continuity, and lack of shorts, is tested while the harness is being built. The only post-build test might be Hi-Pot testing (if required).

 

[goatmeat]

I have tried google, I have spent at least 8 hours looking for ideas online with no results, so I thought I could try here. Most of them are very sophisticated testing units, and I am looking for something simple.

 

[melone]

What are your requirements? What are you trying to "test"? How many at a time? Are you making it yourself or buying it? How many do you need? What are your space considerations? What temperature range? Fully automated, or will there be user interaction? ....

Without narrowing down the problem for us, we will spend at least 8 hours trying to figure out the requirements so we can help you develop an adequate solution.

 

[waross]

Do you want to check continuity, line to line shorts, correct pin assignments, voltage withstand?
The testing procedures will get more compliccated as you move down the list. 16 outputs from a PLC fed to the pins of one connector and 16 inputs fed from the pins of the other connector will let you check for continuity, cross connections, and line to line shorts. Send a signal to each pin and check that it returns on the proper input line. Then or the signals returning from the other 15 pins. Any output from the or will indicate a cross connecion or line to line short. Repeat 15 times for the other 15 lines.
respectfully

 

[VEBill]

If you really want to be cheap and simple, then make a fixture so that the test path is looped around through all 16 conductors. In each path you have a 1k resistor. Plug the cable in and if the end-to-end resistance is 16k, then it is good. There might be a few unlikely miswires that could slip through, but not many.


PS: Does anyone know why threads occasionally just disappear - usually after I've posted a particularly good response?

 

[waross]

Hi VE1BLL;
I have noticed that threads that may be students or homeowners tend to disappear. If someone red flags a thread as a student or homeowner post, and the moderators agree, it's gone.
respectfully

 

[itsmoked]

VE1BLL; I also notice that the search engine is really not too swift. More of a search, (Briggs & Stratton). I have assumed a post was gone in some cases only to really believe it couldn't have fallen victim to the above malady due to its remembered content. I then use the brute force look-at-all-the-threads back for years and I inevitably find it. (sans search engine)

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[goatmeat]

What are your requirements? What are you trying to "test"? How many at a time? Are you making it yourself or buying it? How many do you need? What are your space considerations? What temperature range? Fully automated, or will there be user interaction? ..

The requirements are just to verify that the wiring is correct. Testing for shorts or opens. I am trying to test various wire harnesses. One in particular is 16 Pins. Each one goes from pin to pin. 1-1 2-2 and so on until 16 to 16. We manufacture 100 per week, and have been having quality issues with incorrect wiring. So I am trying to assemble a test fixture, size is no restriction. Ideal it would be a test fixture that would be put on a technicians bench. Every unit must be tested, one at a time. I will be making it myself. There will be user interaction, with the testing technician plugging each harness in a test fixture, my plan from there would be he hits a test button, and gets either a
pass or fail indication. Temperature is not an issue. I will need about 15 different test fixtures that test a variety of harness, so each one will be different. But for now I am just trying to get through one. And thanks to everyone for all the help.

 

[itsmoked]

Hmm Micro controllers or even a beater PC are often the best for these moving target test requirements.

The first thought is just put two connectors on the board for each end of the cable and source enough power on one end to light LEDS on the other. Just plug in and if all the lights on the other end light up you have continuity! Problem is if a couple are crossed you may still get all the lights on. DOH! So you do the exact same thing - you just use a PC or something to scan the power down the connector endlessly. The person testing the cable jacks it in then watches one or two full cycles to see if the lights all light sequentially. If they do you know there are no crossed connections. No buttons too push! Run the scan fast so there is no sitting around watching each light switch. A human will instantly detect a blink pattern anomaly.

Set things up for the largest number of lines you would expect. I would consider a family of circuit boards that connect to your logic machine (PC or ?) with one large standard connector and then each having a family of of connectors for the cables you want tested. Sit down and plan a board to test all the cables you currently want tested then make new boards as future cables appear.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[OperaHouse]

I was going to post the resistor idea and I see it has been suggested. I would drop the resistors down in value to about 10 ohms so it would give a good resistance test. Use 24V source and a 7805 in constant current mode of .1 amp. Use a comparator on the last resistor for a GO/NOGO LED. Simple cheap and very effective.

 

[waross]

I was going to suggest a lower resistance also, but OperaHouse beat me to it.
You may want to use enough current to "smoke' any bad connections. This may be two tests, one in parallel at high current to test the integrity of the connections followed by the resistor test to check connections.
respectfully

 

[BJC]

Been there done that.
These guys make rotary switches with up to 12 positions. The multiconductor cables we had to test had Cannon plugss on both ends.

http://www.electroswitch.com/electroswitchesandrelays/iSelector.htm

I have seen similar test setups using a telephone stepper relay. Trouble with that is the only stepper relays left in the world may be in museums.

 

[OperaHouse]

Resistance check should test about as well an anything. I actually had a circular high 30's pin AMP connector with a void in the plastic between the pins. The locking fingers poked through and connected.

 

[LionelHutz]

Try this;

http://www.testequipmentdepot.com/b+k precision/cable-harnesstester.htm

We use one at work. It has a PCB with patterns to install connectors to test things such as ribbon cables and DB connectors but it could be used for any type of cable by adapting the right plugs to it.

 

[bsmith123]

Have a look at bite magnetic latching indicators---these can be blk/white displays or pass/fail etc.

a panel with an array of 10 by 18 would be small and test ten cables at a time or if you found a quick way (looped connector with jumpers)to connect two wires together you could test 20 at a time and so on.

this would work for continuity testing and they are available in a variety of voltages---mostly used for aircraft avionics fault indication

They could also be arranged to detect shorts IE two bit balls latch when voltage is applied to one wire.

you could do this with a series of buttons or switches but that would take a long time so maybe a program to apply the voltage would be better

just some food for though---hope it helps

 

[HARNTESTER]

I've been testing automotive harnesses for the last ten years. we sometime test up to 1538 circuits at a time. We use Dynalab testers that send a signal from point to point to check for continuity.

 

[automatic2]

A junker PC is a good platform. Testing is good when you look at the consumers viewpoint. You as a supplier may want to 'certify' your cable to some degree. This may include expressing it with voltage/current/frequency/mechanical to ensure it's minimum performance.

I can tell you that cables are one of my biggest headaches when commissioning equipment. As we 'certify' our product, a failure in any cable requires a recertification. It drives me nuts thinking that any manufacturer will send out a cable that hasn't been internally 'certified' to some minimum value (exceeding customers requirements).

 

[klaxon]

I bought a cable tester for $100 or less that did basic continuity testing on standard computer cables. It tests a bunch of types.

Honestly, if you want a turnkey supplier for a cable tester, you are out of luck. Your connectors are your connectors, but modifying a simple cable tester with an adapater YOU can make is a fast way.

http://sewelldirect.com/Hobbes-PC-C...rce=froogle&utm_source=Froogle&utm_medium=cse

Note.... like all technical questions, the right answer depends on the quality of the question. Your query is limited to the word 'test', which can range from measuring the length to full dynamic testing of each conductor. The difference in cost for the two is enormous.

For simple cable wiring verification, use something like the one I provided a link to and modify it for your purposes. $140. (I have one in my lab.)