Re: "Barrel" connectors

Jun 16, 2025 Last reply: 1 year ago 9 Replies


> Is there a better way?


>
> Cancel that.  A colleague has something in the mail to me, already! >

Let us know if it is any good. It is a problem that I have too.


I somewhat distrust that type of connector in general, unless both parts come from the same manufacturer - some of them have a slit inner pin that is slightly springy/compliant and would be appropriate to fit in a solid tube in the other part, whereas others have a solid pin and some of the sockets have a bifurcated springy inner contact. Where there is a solid pin fitting in a solid tube, it seems they only work because there is a sideways force on the whole thing from the asymmetrical outer contact. Seems dodgy.


I've also come across the belief amongst the general public that it's ok to plug in any of these barrel connectors as long as they fit, presumably because that's mostly true of USB and so on. So they fry their modem that wants 12VDC by plugging in a 15VAC adapter or whatever.

Power supplies for POE+ switches often use the same connector type but deliver about 52V at a few amps.

John

The barrel connector situation is a nightmare. It doesn't help that products come with unmarked warts with random voltages and powers and connectors.

There should be two power sources, USBc and PoE.

We need more cheap surface-mount modules that pick off and isolate power for us. We use a Silvertel module in our PoE gadgets. It's not cheap.

People are starting to put magnetic isolators on monolithic ICs, which is progress in the right direction, even if most are not very good yet.

They're not infinitely variable.

Some info on the web wikimedia barrel connectors.

Here's a spreadsheet with ID, OD and center pin diameters, along with mfr ident and typical use (user beware):

formatting link
Center-pin equipped parts expect resistively programmed info for correct function a la USB-C. Some expect to be output-voltage-adjusted by the same means.

Note that current ratings assume correct mating and may not be guaranteed by various mfrs or vendors.

Labelling the installed connector can be useful if your memory doesn't serve you well.

RL

Some Dell notebook power supplies have used that approach.

The barrel consists of two concentric cyclinders with the outer being the negative connection and the inner one being the positive supply. The pin directly connects to a OneWire Dallas Semi/Maxim memory that has the power supply ID.

The computer can then interrogate the power supply to determine current capability and that it is a genuine Dell supply.

My first exposure to one resulted in me destroying the supply!

There was no protection for the memory device and I accidentally shorted the centre pin to the inner barrel when attempting to measure the voltage output.

The 19V from the supply destroyed the OneWire memory. Whenever the PSU was plugged into the computer it would not identify it as a Dell supply and would not charge the battery and would only run at reduced speed.

It is not obvious that the outer barrel and inner barrel are not the same connection.

Be careful with Dell supplies. Modern ones may be different this is from ~15 years ago

I'm reminded of a pair of lists that was making the rounds a decade or two ago; "Best and worst gifts for a Guy". Number one on the "Best" list was a label maker, and the reason given was the fact that absolutely everything on the premises would be labeled. Number one on the "Worst" list was a chainsaw, and the reason given was "see label maker."

They still do this.

Interestingly, Dell and HP use compatible "Big Barrel" and "Small barrel" power connectors. The voltage and polarity are the same. A Dell adapter will work just fine on an HP laptop, because the HP does not use the third pin to identify the adapter. However a Dell laptop will complain about an HP adapter because it cannot read any identifier on the third pin.

It should be possible to make an adapter using the chip from a cannibalised Dell PSU. I intend to try this. However, newer Dell notebooks use USB-C for power, so the problem has mostly gone away.

John

The trouble with 'just try it' is that the 'trial' sockets can wear over time. eg if you insist on putting a 6mm OD plug into a 5.5mm OD socket, it might go but bend the spring contact. Eventually the 5.5mm OD plugs will feel loose because the contact is now bent out of place.

Also you might have a plug that securely fits the OD but it has too large a hole in the inside so it's a poor fit. I've previously killed equipment like that (2.5mm plug on a 2.1mm socket -> poor contact -> voltage drop ->

switching converter has to pull more current to keep up -> fried something).

I think a series of drill bits to measure the ID is a good plan, then maybe there's some way to measure the OD while getting 'acceptable' tension on the spring?

Theo

A number of years ago, while working for a company that repaired laptops, I built a jig using a 190W Dell power adapter, and a switch to select which ID chip to connect to the third pin. So we could test the ability of the laptops to ID specific adapters. The one-wire protocol is not difficult and I considered using a microcontroller to spoof the various messages, but we had lots of adapters with bad cables so it was easier to rob them for the part.

I did build a gizmo that would read the ID, so we could test adapters before shipping them back to the customer.

Join the Discussion

Have something to add? Share your thoughts — no account required.

Didn't find your answer?

Ask the community — no account required