Mating connectors from different manufacturers - Common but not recommended?

[tinfever]

The PC hardware industry seems to rely on compatibility of 4.2mm pitch connectors and crimp terminals between many different manufacturers and different ratings (high-current vs normal). However, I suspect if you asked any single manufacturer about such compatibility, they'd tell you not to do that. Is it that common to rely on undocumented compatibility between connectors? How does this actually work without failures everywhere?

[coromonadalix]

you have to search all their specs, pins sizes etc ... before using them,  if choosen carefully  i would not expect problems

we do at my job  no problems  for power stuff

but for sure quality has taken some beating .... mostly stuff coming from china at low pricing ... quality is low, metal thickness is lower  ...

simply put,  dont cut in quality vs known good brands /  vs cheap prices

are you talking about   the "fire connector"   loll    on video cards

[TheUnnamedNewbie]

I can't answer for the specific PC motherboard/GPU connector example, but in general: Standards, standards, and standards. The world is full of examples of standardized connectors from various suppliers. In some markets (automotive) it is even a requirement to have multiple suppliers for a connector, they refuse to deal with a single-sourced component for something that is not a huge differentiator.

Just think of the examples: USB, HDMI, power sockets, LC/SC fiber connectors, the sockets for lights, SMA connectors for antennas, BNC connectors on equipment, RJ45 connectors for your ethernet-over-twisted-pair, FAKRA... the list goes on.

Some of these were really developed as a standard first (HDMI, USB, etc). Some come from a single company (eg, a lot of the coaxial connectors came out of HP/Aglient/Keysight or Anritsu) but then opened up, either as a license (I believe this is the case with some POF connectors, where broadcomm holds the license) or free. And sometimes you just have connectors that are copied and not protected or even officially licensed. Examples include TRS connectors that used to be used in every audio device ever (up to the point that BLE took over).

[tooki]

The PC hardware industry seems to rely on compatibility of 4.2mm pitch connectors and crimp terminals between many different manufacturers and different ratings (high-current vs normal). However, I suspect if you asked any single manufacturer about such compatibility, they'd tell you not to do that. Is it that common to rely on undocumented compatibility between connectors? How does this actually work without failures everywhere?

That’d be an incorrect assumption. In the datasheets/catalogs, you do occasionally find mention of compatibility with competing manufacturers. And many manufacturer websites have competitor cross-reference lookups (though I have rarely found them to be useful.)

[tinfever]

you have to search all their specs, pins sizes etc ... before using them,  if choosen carefully  i would not expect problems

we do at my job  no problems  for power stuff

but for sure quality has taken some beating .... mostly stuff coming from china at low pricing ... quality is low, metal thickness is lower  ...

simply put,  dont cut in quality vs known good brands /  vs cheap prices

are you talking about   the "fire connector"   loll    on video cards

I get choosing parts carefully to stay within their specs, but I'm concerned about subtle incompatibility between connectors are terminals that don't have an official standard, and when there isn't a spec for mating crimp terminals from company A to PCB headers from company B, even if they physically fit.

The 12VHPWR connector I think you're referring to is a decent example, although there is at least a spec for that from PCI-SIG. I don't think they spec all the way down to the exact shape of the terminal contacts though, so they leaves some room for interpretation by the manufacturer to either get it right...or wrong.

I can't answer for the specific PC motherboard/GPU connector example, but in general: Standards, standards, and standards. The world is full of examples of standardized connectors from various suppliers. In some markets (automotive) it is even a requirement to have multiple suppliers for a connector, they refuse to deal with a single-sourced component for something that is not a huge differentiator.

Just think of the examples: USB, HDMI, power sockets, LC/SC fiber connectors, the sockets for lights, SMA connectors for antennas, BNC connectors on equipment, RJ45 connectors for your ethernet-over-twisted-pair, FAKRA... the list goes on.

Some of these were really developed as a standard first (HDMI, USB, etc). Some come from a single company (eg, a lot of the coaxial connectors came out of HP/Aglient/Keysight or Anritsu) but then opened up, either as a license (I believe this is the case with some POF connectors, where broadcomm holds the license) or free. And sometimes you just have connectors that are copied and not protected or even officially licensed. Examples include TRS connectors that used to be used in every audio device ever (up to the point that BLE took over).

That's a good point on standards. I suspect the PC industry mostly falls in to "connectors that are copied and not protected or even officially licensed". I guess there is enough safety margin built-in to the current ratings, and the connector manufacturers aren't stupid, so everything works well enough together.

That’d be an incorrect assumption. In the datasheets/catalogs, you do occasionally find mention of compatibility with competing manufacturers. And many manufacturer websites have competitor cross-reference lookups (though I have rarely found them to be useful.)

I've asked Molex once about mating HCS and non-HCS terminals, all from Molex, and they would only say "We only recommend our parts mate with the listed mates in the mates with/use with section on the product page." and "We can only promote our products mate with the specified Molex mate for quality purposes. I cannot speak to any other products."

I'm thinking, if I want to connect an ATX power supply to something, there's no way I'm going to know if the PSU manufacturer used HCS or non-HCS crimp terminals. So I figure I'd better use HCS headers and hope that mating HCS and non-HCS just means the connection is rated at the non-HCS current level, not 1/2 the non-HCS level or something even worse than just the weakest link.

[tooki]

I've asked Molex once about mating HCS and non-HCS terminals, all from Molex, and they would only say "We only recommend our parts mate with the listed mates in the mates with/use with section on the product page." and "We can only promote our products mate with the specified Molex mate for quality purposes. I cannot speak to any other products."

I'm thinking, if I want to connect an ATX power supply to something, there's no way I'm going to know if the PSU manufacturer used HCS or non-HCS crimp terminals. So I figure I'd better use HCS headers and hope that mating HCS and non-HCS just means the connection is rated at the non-HCS current level, not 1/2 the non-HCS level or something even worse than just the weakest link.

So you asked one manufacturer, got one answer, and from that extrapolated that that is what all manufacturers do, thus invalidating my claim that some datasheets talk about mating with competitors’ products. Got it. 🙄

Here’s an excerpt from a TE Connectivity power connector catalog:

VAL-U-LOK Connector System
• Wire-to-wire and wire-to-board, pin and receptacle connectors
• 4.2 mm [.165] centerline
• 2-24 position dual row and 3-5 single row configurations
• Ratings: 9A, 600 VAC
• Accommodates 26-18 AWG wire
• Easy-to-mate, positive locking housings
• Fully isolated terminals
• Panel-mount or free-hanging versions
• Black, red and blue in addition to the standard white
• PCB headers are available in vertical, right-angle, screw-mount, and blind-mate configurations
• Intermateable with similar connectors from other manufacturers

[langwadt]

I've asked Molex once about mating HCS and non-HCS terminals, all from Molex, and they would only say "We only recommend our parts mate with the listed mates in the mates with/use with section on the product page." and "We can only promote our products mate with the specified Molex mate for quality purposes. I cannot speak to any other products."

I'm thinking, if I want to connect an ATX power supply to something, there's no way I'm going to know if the PSU manufacturer used HCS or non-HCS crimp terminals. So I figure I'd better use HCS headers and hope that mating HCS and non-HCS just means the connection is rated at the non-HCS current level, not 1/2 the non-HCS level or something even worse than just the weakest link.

So you asked one manufacturer, got one answer, and from that extrapolated that that is what all manufacturers do, thus invalidating my claim that some datasheets talk about mating with competitors’ products. Got it. 🙄

and how could Molex possibly speak on behalf of other manufacturers

[tooki]

I'm thinking, if I want to connect an ATX power supply to something, there's no way I'm going to know if the PSU manufacturer used HCS or non-HCS crimp terminals. So I figure I'd better use HCS headers and hope that mating HCS and non-HCS just means the connection is rated at the non-HCS current level, not 1/2 the non-HCS level or something even worse than just the weakest link.

FYI, I checked, and it looks like the HCS headers and contacts differ only in material ("high performance copper alloy", TE calls it) as opposed to the brass or phosphor bronze in the standard parts. They're expressly designed to use the same application tooling, so that means that every dimension is identical. (In particular, even minute differences in materials would require different crimp tooling, so their express claim that existing tooling is used means that it must be identical.) Thus, the only thing that makes the headers capable of higher currents is the copper alloy having lower specific resistance. The crimp contacts, especially the female ones, could in theory use slightly different internal mating surface geometry, but I kinda doubt it, since the male crimp pins are absolutely identical except for material.

I can't see any way in which intermating HCS and standard could result in lower current capacity than standard.

[tinfever]

So you asked one manufacturer, got one answer, and from that extrapolated that that is what all manufacturers do, thus invalidating my claim that some datasheets talk about mating with competitors’ products. Got it. 🙄

Okay, that's fair. I've extrapolated too far from the unhelpful answer of one manufacturer. I didn't mean to imply your claim was invalid.

and how could Molex possibly speak on behalf of other manufacturers

That's kind of my point. It would be a lot to expect each manufacturer to create a big compatibility matrix between their connector and everyone else's similarly shaped connector. And yet, industries seem to rely on these different connectors mating to each other without each combination having been specifically tested. In scenarios where there is no precisely dimensioned standard for the contacts, it seems like an enormous amount of weight is being put on the words "or equivalent"

From the Intel ATX PSU Spec:
5.2.2.1
Main Power Connector – REQUIRED
Connector: Molex* Housing: 24 Pin Molex Mini-Fit Jr. PN# 39-01-2240 or equivalent.
Contact: Molex 44476-1112 (HCS) or equivalent (Mating motherboard connector is
Molex 44206-0007 or equivalent).
18 AWG is suggested for all wires except for the +3.3 V supply and sense return wires
combined into pin 13 (22 AWG).

I expect very few manufacturers of PSUs or computer components are actually using Molex brand parts. And yet it all works! Copies of copies of copies, all equivalent enough I guess.

FYI, I checked, and it looks like the HCS headers and contacts differ only in material ("high performance copper alloy", TE calls it) as opposed to the brass or phosphor bronze in the standard parts. They're expressly designed to use the same application tooling, so that means that every dimension is identical. (In particular, even minute differences in materials would require different crimp tooling, so their express claim that existing tooling is used means that it must be identical.) Thus, the only thing that makes the headers capable of higher currents is the copper alloy having lower specific resistance. The crimp contacts, especially the female ones, could in theory use slightly different internal mating surface geometry, but I kinda doubt it, since the male crimp pins are absolutely identical except for material.

I can't see any way in which intermating HCS and standard could result in lower current capacity than standard.

I hadn't thought about it that way. Thank you.

[tinfever]

Interestingly, it looks like assumed compatibility between manufacturers doesn't always work out:

PV connector mating and intermatability in NEC 2020
https://www.solarpowerworldonline.com/2020/09/pv-connector-compatibility-rules-are-likely-coming-to-the-national-electrical-code/

"As part of a technical assessment of solar project risk, TÜV Rheinland evaluated different PV module failure modes according to severity of effects, likelihood of occurrence and detectability. According to the results of this Failure Modes and Effects Analysis (FMEA), failures associated with dissimilar connectors represent the single greatest risk to PV system performance and safety. Walmart’s now-settled lawsuit against Tesla, which hinged in part on roof fires and incompatible module connectors, illustrates high stakes associated with dissimilar connectors."

[tooki]

Interestingly, it looks like assumed compatibility between manufacturers doesn't always work out:

PV connector mating and intermatability in NEC 2020
https://www.solarpowerworldonline.com/2020/09/pv-connector-compatibility-rules-are-likely-coming-to-the-national-electrical-code/

"As part of a technical assessment of solar project risk, TÜV Rheinland evaluated different PV module failure modes according to severity of effects, likelihood of occurrence and detectability. According to the results of this Failure Modes and Effects Analysis (FMEA), failures associated with dissimilar connectors represent the single greatest risk to PV system performance and safety. Walmart’s now-settled lawsuit against Tesla, which hinged in part on roof fires and incompatible module connectors, illustrates high stakes associated with dissimilar connectors."

This is true, but also a bit of a special case, in that they’re used at quite high currents (up to 95A for the leading connector, the MC4), must withstand high voltages (up to 1500V for MC4), and be weatherproof for decades of outdoor use. That is a lot to ask of a connector.

Unsurprisingly, in addition to compatible connectors from big-name connector manufacturers, there’s a huge number of cheap fake MC4 connectors that simply aren’t made to the same manufacturing tolerances as the originals or big-name clones. They may be off dimensionally, the materials may be different, etc.* And the worst thing is when they’re passed off as the genuine article, causing people to unwittingly mix brands. (Not that the cheap fakes perform as well even when mated with themselves.)

Additionally, when the wrong tooling is used for assembly (be it cheap tools with cheap connectors, cheap tools with genuine connectors, genuine tools with counterfeit connectors, or genuine tools from one good brand with genuine connectors of a different good brand), that can cause problems. The only connectors I’m aware of where tool and contact compatibility are completely guaranteed regardless of tool manufacturer and contact manufacturer are the mil-spec connectors, because the US military designed the specs and any manufacturer that wants to sell to the military (and aerospace, which also uses them commonly) has to follow those specs.**

This article covers a lot of the PV connector issues in more detail. https://www.pvel.com/wp-content/uploads/PVEL-HelioVolta-Ultimate-Safety-Guide-for-Solar-PV-Connectors-Feb-2022.pdf

These pitfalls do apply to small, low voltage connectors like Mini-Fit Jr, but the stakes are far, far lower, since the environment is much less demanding, and they’re not as likely to actually remain in use for 25 years.



*The genuine MC4 connector is from Stäubli, which also makes some of the very best, possibly the best, banana plugs on the market. And unsurprisingly, there are plenty of clones of Stäubli’s signature lantern-style (“lamella”) plugs. Some are good, but some are awful. They look similar at first glance, but for example, on some the plating of the parts is not as conductive, which is a big issue with lantern plugs, where the lantern is free to rotate. On some, the lantern is too loose on the plug shaft, also reducing conductivity. On some, the material of the lantern is bad, so that after a few insertions, it loses compliance (doesn’t spring back as much) and then fits into jacks loosely, reducing conductivity. In contrast, I have encountered well-used genuine Stäubli plugs that are older than I am, and they still perform like new. I’ve literally never seen one wear out.

**I wonder if instead of the NEC requiring installers to ensure they don’t mix connector types, if maybe the NEC shouldn’t simply create a manufacturer-agnostic “NEC spec” that any connector companies that want to sell in USA have to follow, just like the mil-spec connectors. Unified tooling, unified part numbers and specs, etc.