Signal connector is able to carry 4 Amps constantly?

Hello
Do you think that we can use the Molex picoblade 10 way connector to pass 4 Amps to our LED load? We have paralleled 4 of the "Blades" for the positive, and four for the 4 amps return. Do you think this will be ok,or will it lead to early corrosion of the connector?


Molex picoblade 53398-1071 connector:
https://www.molex.com/molex/products/datasheet.jsp?part=active/0533981071_PCB_HEADERS.xml

It is rated for 1A per contact maybe only under "ideal" conditions so with 4 contacts in parallel your load of 4A will reach its maximum. I would not risk it.
It is rated for mating only 30 times which might decrease its current rating if you need to unplug it a few times.

No it can't carry up to 4A, the max current is 1A.

People on the other website say that only one pin has a 1A maximum rating maybe due to heating. More than one pin with a current of 1A causes more heating which might destroy the connector.

We know redundant contacts help share current and reduce voltage drop and power dissipation.

But when the reliability of the contact resistance is not clearly stated or is suspect such as limited to 30 insertion cycles, then we must suspect the reason may be cost reduced plating. Generally a robust contact may be several layers of different plating materials including nickel for durability. When it is just Phosphor bronze and tin, the tin is very thin sub-micron liquid coating to prevent oxidation and thus erosion may cause wide variation in contact resistance.

You can imagine the contact area as very small even with split or "bi-furcated" contacts so that 1/10W might increase contact temperature and thus increase oxidation rate.

I usually imagine in the absence of specs, the size of an SMT resistor size and compare its power rating e.g. 1/16th W with a pin contact area and realize there is a reason for limiting current.

The main consequence here about Current Sharing, is power dissipation and reliability since voltage drop may not be important in your application with a current source.

In the absence of spec's, I would first contact the factory tech support by email or your buyer ... to ask the distributor to get the factory specs for current sharing. contact tech support from the distributor/supplier to ask "How many 1A pins ( for p/n's x,y) rated for 1A, do I need to carry 4A for Vdc and return for low insertion cycle application?

Forecast demand on p/n's may carry more weight on response time, but let them know how much your value their support.

This chain of communication sometimes ensures better support since that choice of purchase hinges on answers, your organization will know who to contact or you can cc them when you contact the factory direct by email.

I might expect conservative approach of 8 (1A) pins to carry 4A for each PWR & RETURN, ... but only the factory knows for sure.


I suspect they have an "internal only" resistance spec of something like...
initial value <50~100 mOhm
final threshold at MTBF rating .... 100~200mOhm

I might expect conservative approach of 8 (1A) pins to carry 4A for each PWR & RETURN, ... but only the factory knows for sure.

Click to expand...

do you mean 16 pins total for 4 amps? (8 for power and 8 for return)

treez said:

do you mean 16 pins total for 4 amps? (8 for power and 8 for return)

Click to expand...

Yes he means exactly that. You'll need the same number of returns for each Vcc pin you use.

e.g. Vcc1 4A, Vcc2 x 4A: 8-pins Vcc1, 8-pins Vcc2, 16-pins for return (if the board uses a shared ground, which is usually the case)

Sorry but how do we know that SunnySky means that?

Given that SunnySkyguy's American English skills are quite good...

8 (1A) pins to carry 4A for each PWR & RETURN

Click to expand...

He's stating each of the power and ground has 8 pins.

Now I could be wrong, but I don't think so.

Besides you can't expect to supply 4A and then see a return of only 2A. Kirchoff's law would be broken.

ok, I have a reply from molex saying its for 1A, but they don't say whether they mean per blade or per connector.....they did suggest I use a different connector after I explained the spec to them.
The company I work for are also using the picoblade for the microcontreoller signal connections, and want to use the same one for the power..so as to keep the BOM minimised..thus they paralleled the blades, and think that's fine.

Did you ask "How many contacts are need to supply and return 4 Amps in this series of 1A connector pins?

Yes I suggested 16 pins which has a safety margin best case of 50% using 0.5A per 1A rated contact thus 8 per V+ and 8 per return..

Consider a better connector such as the Molex 4A to 25A solution guide

https://rhu103.sma-promail.com/SQLImages/kelmscott/Molex/PDF_Images/987650-8581.pdf

I can understand most Americans but I speak Canadian

I would go with a standard 4 pin Molex Disk drive connector.

Vertical Mount Voltage/Current: 250 VAC (RMS) / VDC @ 10 Amps
Right Angle Mount Voltage/ Current: 250 VAC (RMS) / VDC @ 6.5 Amps

Some questions can be easily answered. When the datasheet says "Current - Maximum per Contact 1A", it means per contact. Without explicite reservations, this also means 1 A for each contact.

It's no so easy to know if the current will be equally shared between multiple paralled contacts in your application. In some cases you can be quite sure, e.g. if each pin is connected to an individual wire of some length so that the wire resistance will balance the currents. That's e.g. the case with ATX power supply cables. If the connectors are used with neglectable cable length, the variation in contact resistance matters more.

In any case, good design practice would suggest a certain margin to maximum current rating. In so far I agree with the 0.5 A per contact suggested by SunnySkyguy.

do you think its easy to say its 1 amp per contact?...because that may just refer to a maximum transient...the description says its a "signal" connector...and the product spec below refers to all the tests with just 10mA flowing in each blade....


https://www.molex.com/pdm_docs/ps/PS-51021-024.pdf

Test conditions have nothing to do with maximum ratings. "Current - Maximum per Contact 1A" is the thermal current rating.

Here are a couple of links about derating, the second one recommends 80% of the max rating per contact.

https://www.componentsengineering.com/wp-content/uploads/pdfs/Component-Derate-Guidelines.pdf

Just like power sharing in resistors, the tolerance of the resistance affects the load sharing. If matched within 20 %, then perhaps derating only 20% is needed. But the thinner the plating ( tin only), then I would suggest the wider the tolerance (50%)

Since this header is designed without nickel or gold plating (cheap) it is limited in aging cycles to 30. The low current test is a worst case for oxidation , which is why Molex did that.


Opinion...
For about a penny per header pin you can get gold plated connectors , which are much more reliable. You can get for less cost from Auk. $500 MOQ This is what your company ought to be using instead for other applications.


IN this case there is no minimum current, just a maximum insertion cycle of 30 to prevent friction damage to plating.

OK thanks, the thing for me is that this connector is called a "signal" connector, and I believe surely that we are mad to use it for 4 A of led current by paralleling 4 for "power" and 4 for "return"?

Signal means signal..surely ..

The Apps guy at Molex got back to me and told me not to use it because it Is "1 amp" rated.....but he didn't say why four blades paralleled could not each carry 1 amp.


ads_ee...thanks, your articles confirm that what we are doing, even though appearing to be adhering to the spec of "1 amp per contact", is totally wrong...and would have put us out of business!..thanks again.

The recommendation I was given when working with a server design (mainframe/super servers) company was to derate the specified current to 60%. This was 10+ years ago but that was the standard the design group worked with at the time.

Thanks, so is the concensus that we should not use a 1 AMP signal connector to carry 1 amp because it will likely corrode sooner due to carrying this high current?
I just asked because it has been said here that the test document refers to low current (10mA) because corrosion is worse at low currents......so is corrosion worse with low or high current? (intuitively I believe high current worsens the chances of connector corrosion?).



Anyway, I believe a point that needs evaluating is the GREAT difficulty that is involved in properly crimping small connectors such as pico blade from Molex.

The following shows the 13 problems of crimping....
Molex: Connector Crimping problems
**broken link removed**


I reckon crimping small connectors is commonly not done properly, but because everyone de-rates their connectors, it doesn't really matter...however, whenever you are operating the connector at its current limit, then the crimp must be perfectly crimped.......and I think all of us would admit that none of us regularly makes crimps perfectly, such a fiddly job this is, -especially with small crimps. For a start, if the wire has a manufacturing tolerance, and doesn't sit in the middle of the insulating jacket, then your strippers are going to cut the braids of the wire itself, and straight away you are going to have an improper connection....so I believe that all connectors must be heavily de-rated. do you agree?

Firstly I read your post as a statement why you would better use a designated 4 A connector.

Secondly it gives reason why one should use correct crimping tools and preferably get qualified service providers making the cables.

I agree with both points.

I don't agree with the viewpoint that small connectors must be derated because it's impossible to crimp them properly. Maybe it's true for hobby projects.