Dsub connector reliability question

I just realized 21 wires FWF for thermocouples and red cube are a problem when you want to remove the engine to paint the fuselage or rebuild the engine. My option is to ignore the future possibilities or address it now. A Dsub connector at the firewall pass through would add a lot of failure points, but would address a guaranteed headache some day. Any experience to share ?

Personally I wouldn't take the risk and I can tell you from experience that disconnecting and reconnecting those leads isn't that bad. I'm not sure, but having dsub pins and sockets in the thermocouple leads might be a problem as well.

When building our airplanes it can take many hours to wire and plumb firewall forward. This process would tend to make you favour a connector at the firewall. However, when you actually do remove an engine all the hoses and wires come off quite easily. Easily if you have planned for removal, by that I mean, not wrapping wires and hoses in difficult locations to begin with. Do your engine installation with a future removal in mind.
When you put your engine back on all the wiring, and hoses are ready to just connect again. If you take lots of pics and write a few notes you will be surprised how easily it is to remount an engine.
With that in mind I would not recommend a firewall connector. Most wiring issues are at connections.

Don't do it

I'm with Mark on this one - not seeing a big advantage to doing this. If you need to remove the engine - once every 2000 hours? - just unhook the sensors and bundle them up near the firewall, on an engine mount for example. It will add 30 minutes to the whole process.

As an aside, and sorry for the thread drift, but it would be great if the manufacturers could have an intelligent box that can be placed FWF to collect all these sensor wires, and then a pair going back to the EFIS. I think MGL does this, but none of the others.

rv8ch said:

I'm with Mark on this one - not seeing a big advantage to doing this. If you need to remove the engine - once every 2000 hours? - just unhook the sensors and bundle them up near the firewall, on an engine mount for example. It will add 30 minutes to the whole process.

As an aside, and sorry for the thread drift, but it would be great if the manufacturers could have an intelligent box that can be placed FWF to collect all these sensor wires, and then a pair going back to the EFIS. I think MGL does this, but none of the others.

Click to expand...

Believe you are correct. I was disappointed to learn that my Dynon 220 box was not allowed FWF. IMO this capability should be a given.

Its not a big deal to remove those items and suspend them from the firewall. I did it on my IO-540 one time and it was actually easy. I sat on a bench while I did it. Make sure you label them when they are removed. If your engine is coming off for an OH, they are coming off anyway, and I would rather they stay in my possession than the OH shop. No way I would want any kind of quick disconnect for those items on the FW. That just adds complexity and additional points of failure and weight.

I'm with Mark and Tom on this. Less connections= less potential failure points.

You shouldn't be removing the engine very often.

The old adage- if it isn't there, it can't fail applies.

I used spade connectors for the EGT and CHT probes as I suspected they will need to be replaced from time-to-time.

Other wires were removed at the source the two times in 14 years that I have had to pull the engine.

Time wise, I think it takes me longer to remove a C/S prop than it does to strip all the electrical harnesses & sensors off an engine.
KIS is the key.

Thanks to all !

Strip the engine, no plugs, is the plan. Your input was timely to motivate me to assure I can get the wires off without surgery or embarrassing splices!

I'm going to be the dissenting opinion here, but my viewpoint is not about convenience. What everyone has said so far is true about the ease of disconnecting an engine. But what I object to is the way wires are passed through the firewall. Why use steel fluid connections and steel eyeball cable passthroughs and then just bundle a bunch of wires through the firewall? Most of the passthrough methods I've seen don't impress me. For normal operation they'll be fine but I'm not so sure they'll hold up in an engine compartment fire. I haven't got any data to back this up, like the thread on firewall insulation, but I'd sure love to see someone with the resources do something similar.

My -6A is conventional in this regard, but when I started building the -10, I looked for something better. My main power passthrough is a marine transom passthrough. For the lightspeed ignition, I'm using steel coax bulkhead connectors. And for the various sensors I have metal-shrouded CPCs (they cost $$$ and the bodies are 4" long, making them a bit awkward to locate, since I need that clearance on both sides of the firewall.). I think the CPCs are the weak link (I couldn't source any milspec connectors), but I think they'd hold up better than a grommet-and-shield passthrough.

Understand, I'm not calling the traditional methods unsafe. They've been successfully used for years. But knowing how we builders of experimental aircraft can obsess over some of the details, I'm a bit surprised this hasn't received more attention.

Special TC connectors

N941WR said:

I used spade connectors for the EGT and CHT probes as I suspected they will need to be replaced from time-to-time.

Other wires were removed at the source the two times in 14 years that I have had to pull the engine.

Click to expand...

From what I understand, spade connectors could add a bias in the temperature readings. They make connectors specifically designed for thermocouples that wont introduce errors in the readings. Not sure if this is OCD for homebuilt airplanes or not.

PilotjohnS said:

From what I understand, spade connectors could add a bias in the temperature readings.

Click to expand...

Could you elaborate on this a little please?

McStevens said:

Could you elaborate on this a little please?

Click to expand...

Whenever two different materials are joined, a thermocouple (TC) junction is formed - there?s a small, temperature dependent voltage drop across the junction. So if you crimp the TC wire to copper, another junction is formed, which generates its own small voltage. Fortunately most connectors have a TC wire to copper junction, then, a 1/2? later, a copper to TC wire junction. The two junctions produce equal but opposite voltages, as long as both junctions are at the same temperature, and the net voltage from the two junctions cancel out. There?s only a problem if, by bad luck or poor choice, one end of the connector is significantly warmer or cooler than the other end.

Thermocouple connectors

Here's a somewhat old article that talks about how thermocouples work, and considerations for splicing.

http://www.aeroelectric.com/articles/excerpt.pdf

From what I have read from the same author more recently, modern engine management systems don't have a problem with well-installed connectors. The main thing seems to be keeping things consistent between the two different types of thermocouple wire.

I'm sure that there is the opportunity for some interesting tests using the various engine monitoring solutions, changing the wire, changing the connector, etc.

http://forums.matronics.com/viewtopic.php?p=483166

The cool thing about modern thermocouple
sensor/indicator systems is the very low currents
flowing in the thermocouple 'loop'. Back when
thermocouples were used to POWER an instrument
on the panel, the voltages were small but the
currents significant . . . loop resistance had
to be kept LOW and FIXED. Thermocouple wiring
was rather fat. Long runs to measure CHT, EGT
or oil TEMP on the outboard engines of something
like the B29 might have been 18 or even 16AWG.

This was because thermocouples like Type K

Generated only 7.7 MILLIVOLTS off reference
at 190 degrees-C. Your error budget for voltage
drop in wiring was exceedingly small!

Nowadays, the panel mounted instruments are capable
of resolving the 400 or so microvolt difference
between 180 and 190C with a very high input impedance
to the instrument. No need to maintain
very low resistance pathways in the thermocouple
loops You only need to manage parasitic
couples.

The OLC-1 brings like-metals in the T/C loop direcly
in contact with each other . . . no parasitic couples.
So while the OLC-1 splices would be frowned upon
to join wires in a power feeder with significant
current, potential for upsetting the accuracy
in a modern thermocouple instrumentation loop is nil.

Bob . . .

Click to expand...

My take

So my take on all of this is that connectors are ok in thermocouple wires, but the connection must go back to the same type of thermocouple wire shortly after the connector. Both ends of the splice must be in the same environment.

So dont put a spilce to regular wire in the engine compartment, run the regular wire thru the bulkhead, and then go back to thermocouple wire in the cockpit.

And for the truly anal, use the EI barrel splices as I did: https://www.aircraftspruce.com/catalog/inpages/eioverlapolc-1-10-05470.php. Stagger them inside a piece of heat shrink and you have an excellent and sturdy connection. I hate having loops of thermocouple wire hanging in my engine compartment, so I cut the leads to length and use these. Neatens things up a little.

TC extension junction

I read some input here that is possibly conflicting regarding TC connections. I was advised years ago by Rainer Lamers (MGL CEO ) that extensions with non thermocouple wire are fine as long as the yellow and red splices are the same temperature.
Apparently ALCOR agrees with this because their TC?s have steel ring terminals crimped on the type K leads, staggered about 1? apart. The assertion that mixing metals (type k/steel/copper) is OK at the same temperature for both yellow and red junctions, is supported by the info above , regarding low power demand digital instrument sensitivity compared to a B29 scenario.
If I have misunderstood anything, please set me straight.

Idea

Larry DeCamp said:

I read some input here that is possibly conflicting regarding TC connections. I was advised years ago by Rainer Lamers (MGL CEO ) that extensions with non thermocouple wire are fine as long as the yellow and red splices are the same temperature.
Apparently ALCOR agrees with this because their TC?s have steel ring terminals crimped on the type K leads, staggered about 1? apart. The assertion that mixing metals (type k/steel/copper) is OK at the same temperature for both yellow and red junctions, is supported by the info above , regarding low power demand digital instrument sensitivity compared to a B29 scenario.
If I have misunderstood anything, please set me straight.

Click to expand...

I never had to analyze this type of thermocouple wiring in my day job, but is an interesting idea.
The true test is to calibrate the system once installed by submersing the tc in boiling water and ice water, making sure the temp gauge reads appropriate for the altitude. This method wont calibrate out the mis match of junctions if one is on hot side and one on cold side, but should be good if care is taken to make all junctions in pairs in same environment