Modifying a MIL SPEC connector?

[jkrienert]

I have a few 97-series MIL-SPEC connectors. Their 20-29 pin mapping includes 17-cnt 16awg contacts. The specs on these plugs state within the connector, each pin has a 500vdc 13a rating. The pins individually outside the connector have a rating of 500vdc 21a.

I was considering modifying these 20-29 connectors for a charger and controller in a 1.5kw 3-phase hub driven by a 72v 40a max controller powered by a 73.6v nominal 3kw traction pack (LTO, 32S2P). This will be an environmental field research vehicle, so it will encounter splashes, mud, and dust, hence seeking some connectors with some durability (these are IP54).

For the chargers x2 10awg +/- wires and 20awg relay activation wire, I would parallel solder 16awg solid core copper from each pin to a copper plate with bolt mount for charge wires, twice for 7pins (+) and 7pins (-), that leaves 2pins for a stand-alone male plug interlock wire, and 1pin for the 20awg wire to activate charger relay. The max this charger is programed to put out is (quiq-1500) is 84.3v 12a. I would make the charger the female side so there are no exposed pins, and the male side connected to the charge contactor on the traction pack. Drew a crude sketch with copper pin extensions and bolt plate, that would all be solder connected.

For the motor phase x3 12awg wires and x5 22awg hall and temp wires, I would parallel solder (same method as above) 4pins x for each of the 3-phase connections, leaving the other 5pins for the hall-sensor and temp-sensor wires. I would make the hub connection female, and the controller male, so no pin shorts could occur from idle spinning of hub when disconnected.

On the back of the bulkhead and plug for each of these connections I planned to isolate all appropriate parallel pins with G10-FR4 sheet (0.5mm), and heat shrink all pin connections. The photos below show the basic pin maps mentioned above, and a loose mock up of some g10 isolation sheets for the charger layout.

I'd like to find out if this is a reasonable way to use these connectors. However, I'm concerned that IF I use these plugs with these modifications, there will be too much risk. Even more than posed by their olive drab cadmium plating *cringe*.

Example of a possible risk, should any pin(s) fry inside the phenolic insert of the plug due to arcing or faulty contact, the total current will unload on all other remaining pins. There are enough pins to loose a couple on the chargers connector, and maaaybe one on the phase connections. However, I am also planning that my interlock and safety systems would shutoff the main contactor for the pack to prevent any secondary damages from this example scenario.

I've not seen any other examples of parallel wiring these MIL SPEC plugs, however I know that parallel wiring is common in a lot of other types of connectors and wire management methods.

So, I ask y'all ... Do you see any red flags, or have recommendations on how this could work safely? I'm not an electrical or mechanical engineer, so if this is unsafe I'm ready to scrap and find another affordable plug/socket that will serve the needs. I could still use one of these plugs nominally for its rated design, to connect the vehicles 12v accessories, so not a total loss.

Thanks.

 

[jkrienert]

Are these the correct ways to estimate the total pin cross section for effective AWG?
> each pin cross section = 1.5mm diameter and 0.25mm wall thickness is [((0.75mm)^2 * pi)-((0.5mm)^2 * pi)] = cross section per 0.98mm^2
> total pin area per all charge poles = 0.98 * 7pins = 6.9mm^2 = equivalent to 9-8 AWG
> total pin area per all phase poles = 0.98 * 4pins = 3.9mm^2 = equivalent to 12-11 AWG

If so, then 4-pins wont be enough for the motor phases. However what about 5 pins and the hall/temp sensor pins (5) are used for the motor phases, and I incorporate another connector elsewhere...
> total per phase poles = 0.98 * 5pins = 4.9mm^2 = equivalent to 10-11 AWG

Not sure that will be enough to keep things cool. I could be doing the pin contact cross section for current capacity wrong... For what its worth, the surface area of each pin is:
> each pins ideal full contact surface area = 1.5mm diameter * pi * 14mm length = each pins outside surface area 65.9mm^2

 

[Gruesome]

You said the individual pin rating was 13A, for pins sized for 16 gauge wire. This seems to agree with this Amphenol document . Why are you then second guessing it with the cross section and surface area calculation?
Also, why are you saying you are modifying the connectors, when from what I can tell you are not exchanging pins, and simply wiring them?
If the correctly sized wires are correctly crimped or soldered to the pins, then the voltage rating for the connector should be as specified in the connector document, and your G10/FR4 sheets are superfluous.

Added: 72V doesn't seem all that high. If you (or your safety guys) are worried about high voltage on exposed pins when the cables are not connected, you could wire the power supply with an interlock so that it only turns on when a resistor or short at the far end connects two pins. Assuming you have two spare pins. It also seems overkill for these voltages.

 

[thepronghorn]

Have you read ebikes.ca's connector guide? They go over the most common ebike connectors. I would recommend sticking to something more standard than trying to parallel a bunch of pins like this. If you really want something more IP rated, I could maybe make some suggestions.
Connectors - Learn

 

[Gruesome]

I don't know. When I first encountered them, these cylindrical mil-spec Burndy/Amphenol/whoever screw-lock connectors seemed unnecessary cumbersome, but they sure are bombproof with their metal housing. You also get a pretty good current per pin, and the density is not all that bad for a connector you can still hand-solder.

 

[jkrienert]

...Why are you then second guessing it with the cross section and surface area calculation?
...why are you saying you are modifying the connectors...?

@Gruesome Since electrical and mechanical engineering is marginal to my working knowledge, I was putting some added effort into trying to confirm specifications and my inner confidence about what they mean - sort of revealing what's inside the 'black box'. I suppose that all stems from the possibility that others might ride this vehicle than myself, and I'd like to minimize any user errors if possible. To the same purpose, I took a few photos that confirm agreeable pin to conductor cross sections. I just don't yet have a strong technical notion on the best way to wire the pins in parallel. I'm considering scraping the bolt on plate idea since space is limited, and instead direct wiring the conductors (see crude sketch).

You're right... what I'm suggesting probably isn't really modification of the connector themselves. It is more a modification of the standard methods used to put them into service. I've not resolved a good document on this, nor have I seen any evidence that others have parallel wired pins on a MIL SPEC connector. So, the FR4/G10 sheet was a way of implementing any preventative aspects I can, given I don't have a whole lot of evidence on the functionality of what I'm proposing.

Have you read ebikes.ca's connector guide?

@thepronghorn , I've read the connector guide and several others provided by Grin. They are excellent. My lean to use these milspec connectors is that I had them on hand, and they 'should' be mechanically capable for this task. Though, as you note, their IP rating leaves me wanting to seek an alternative. If you have some recommendation in that regard, I'm stoked to hear. I can give you some more of the builds technical details to help that if needed.

These connectors are somewhat serviceable (I can order NOS inserts with different pin outs ... which might be an alternative I take if the proposed doesn't work), and as @Gruesome mentioned, bombproof. Their connections on this build will not be in need of many mating cycles, as the charger will be onboard, and the motor should generally always be plugged in other than when servicing, which will hopefully be sparingly if at all.

 

[Gruesome]

I would wire them as they are intended to, one wire per pin, use the proper strain relieve, and then connect the parallel wires at the feeding and the receiving end via soldering to a PC board or clamping. (I.e. I would NOT run a big cable into the connector, and parallelize there.)
A quick google found this: Is it bad practice to parallel power connector contacts?

 

[jkrienert]

The charge connector to the contactor bolt is about a 7 inch long run. For the controller board to motor phase connector it's a 24 inch length run.

From a wire management stance, I'd favor a but-splice of the multiple 16awg wires to a single 8awg close to the connector. That would also prevent having to remove the existing contactor charge wires and controller phase wires... however, this notion might be my lack of electrical engineering background speaking...

If thats not the best way, then I will run each pins 16awg conductor to their respective contacts. This will require quite a few cables. Once I fix proper strain relief at the connector, can I sheath each group of pins as a single wire for management sake? I'm not sure there is enough real estate on the controller board for the phase bundles. For the charge bundles I'd suppose strip, solder, and crimp into a ring terminal for the contactor bolts?

 

[Gruesome]

Well, the fewer connections, the better. Connections are what breaks. If you already need a splice to connect to 8AWG wire on both ends, then maybe you could use individual bullet style connectors instead of the 16AWG multi-pin circular mil-spec connector?

 

[jkrienert]

Well, the fewer connections, the better. Connections are what breaks. If you already need a splice to connect to 8AWG wire on both ends, then maybe you could use individual bullet style connectors instead of the 16AWG multi-pin circular mil-spec connector?

Bullet style connectors would be ideal, but i don't have any on hand that have a bulkhead style fitting for the watertight vented case all this will be housed in. Would the bundles of 16awg be reasonable given the lengths mentioned? That would be less connections overall I suppose.

EDIT:
All the helpful recommendations so far, leaves me searching for a rated work around... There are 97-20-15S/P inserts that will fit these Size 20 fittings. So, if I can find a few that are affordable, I can go strait to the pins with 8awg. The last comment about bullet connectors made me think of this. A lot of common sense there thank you.

 

[jkrienert]

... maybe you could use individual bullet style connectors ...?

You were right, relatively speaking. Utilizing 'bullet' style connecters was the ticket. These 97-20-19 inserts are properly sized for the existing cable runs and I can still reap the benefits of the durable housings. Thanks for the helpful advice @Gruesome.

 

[jkrienert]

Finally moving forward with these connectors.