Real World Interfaces logo

3.5mm mono sockets for vertical PCB mounting

2014-05-29 Robin Whittle rw@firstpr.com.au

I plan to update this page in the future with photos of some sample sockets and nuts, which arrived in March 2014 from QingPu.  Here is a photo they sent and some news:

For the WQP-PJ3010BM, which is the main subject of this page, the metal shell is now silver-plated brass rather than steel.  The shell material is thinner and has sharper bends.

The spring contact (as noted below, made for some time now with beryllium copper, rather than phosphor bronze) has been redesigned to improve contact with the NC contacts.  They have new machines which ensure the bushing is  now tight - apparently they had some trouble in the past with them not being tight enough.

They have developed a new product, the WQP-PJ301M-12:



which they describe as an alternative to the Kobiconn 16PJ138 .  This has silver plated pins too.  This enables a lower distance from the top of the PCB to the bottom of the panel, since there is no extra height imposed by the metal shell.  The two signal pins are 3.7mm long.  The distance from the back of the shell to the base of the bushing is 10.0mm - this would be the typical PCB to panel distance.  From the base of the shell to the top of the bushing is 14.6mm.  The ground pin is a close fit in a 0.85mm hole while the two signal pins are a close fit in a 1.0mm hole.

While the WQP-PJ3010BM can be used for PCB mounting, its two signal pins have round solder lugs, so they would need to be squished manually for smaller PCB holes, or larger holes used.  The new WQP-PJ301M-12 has straight pins which will work perfectly for PCB mounting.

I tested both these sockets with the most extreme compression of the contact spring which is possible by moving the plug as widely as the bushing allows.  In both cases, the contact spring (now beryllium copper) continued to make good contact with the NC contacts.

Update 2014-05-29: These are available in batches of 50 from Thonk in the UK: http://www.thonk.co.uk/product-category/diy-accessories/jacks/ .

Introduction

When building modular synthesizer modules and many other devices, it is frequently desired to have a front panel parallel to a PCB, with various LEDs, sockets, other kinds of connectors, switches, pots etc. mounted on the PCB and accessible via the front panel.  On the assumption of a horizontal front panel, this means we need "vertical mount" 3.5 mm jack sockets.  (What does "jack" really mean??)

The challenge is to find reliable components which feel and look good - and to choose a front panel to PCB spacing (hereafter "Panel Gap") which works with them all.

Some places where such things are discussed are:
The focus here is on a particular type of vertical mount PCB 3.5mm mono socket - with an NC (Normally Closed) contact, with a threaded bush.  Below that I point to various alternatives which might be of interest.



The QingPu Electronics WQP-PJ3010BM

Short version:

These are the best or perhaps the only type of socket which I think does the job properly.  This style of socket seems to be made by several manufacturers.  Some of these sockets have a primary contact which does not work properly - it can become deformed so it no longer presses against the NC contact.

QingPu Electronics have solved this problem by using a thinner material for the primary contact, made of beryllium-copper.  This works like a charm.

In 2014, US supplier Erthenvar has this kind of socket from QingPu. They sell them at attractive prices in lots of 100, which is far more convenient for many folks than the 5000 MOQ (Minimum Order Quantity) of QingPu themselves.

Update 2014-05-29: These are available in batches of 50 from Thonk in the UK: http://www.thonk.co.uk/product-category/diy-accessories/jacks/ .

The abovementioned wiki page pointed to this socket:

http://www.qingpu-electronics.com/en/products/WQP-PJ3010BM-34.html

made by Wenzhou QingPu Electronics Co. of Wenzhou City, south of Shanghai.  They make a huge range of switches and connectors.

Here are some photos of some QingPu samples sent to me  in mid-November 2012:

3.5mm mono socket with NC contact for vertical mount on PCB, with threaded bushing
1.


3.5mm mono socket with NC contact for vertical mount on PCB, with threaded bushing
2.

3.5mm mono socket with NC contact for vertical mount on PCB, with threaded bushing
3.

3.5mm mono socket with NC contact for vertical mount on PCB, with threaded bushing
4.

3.5mm mono socket with NC contact for vertical mount on PCB, with threaded bushing
5.

3.5mm mono socket with NC contact for vertical mount on PCB, with threaded bushing
6.

In early November 2012 I purchased some similar sockets from Erthenvar in the USA:

http://erthenvar.com/store/eurodiy/jack35mmv

This was a good deal - 100 sockets (without nuts or washers) for USD$25, plus postage.  This is far better pricing than any for other 3.5mm sockets I could find at mainstream electronic component suppliers:

http://au.element14.com  (Australian branch of  element14/Farnell/Newark.)
http://australia.rs-online.com (Australian branch of RS Components.)
http://www.digikey.com (Digikey - big US distributor/retailer.)
http://www.mouser.com (Mouser has  AUD$ prices and free shipping for orders of AUD$200 or more.)
http://www.alliedelec.com  (Specializes in connectors, switches, electromechanical components etc.)
http://www.elfaelectronics.com
http://www.rapidonline.com
http://www.jaycar.com.au
http://www.altronics.com.au

Erthenvar sell knurled (round) nuts, hex nuts and washers separately.

When I bought this set of 100 sockets in early November 2012, the Erthenvar page had a note: "We are currently investigating the reliability of the jack's switch functionality."

Sure enough, the switch function of these sockets would not work if the plug was inserted and moved around with force, bending the primary contact more than would be the case if it was simply inserted in a gentle manner.  The primary contact would become permanently deformed so it would no longer press against the NC contacts.

Then I began searching for alternatives and found the QingPu Electronics page:

http://www.qingpu-electronics.com/en/products/WQP-PJ3010BM-34.html

On a Sunday afternoon (in Australia and China) I wrote to the Sales Manager Edan Wang, via the web-page's contact form, and received a helpful reply at 7:35 that evening.   We exchanged more emails that night and over the next few days.

He assured me that their new production of these sockets involved a beryllium-copper primary contact, and that this had no deformation problems.  I guess the primary contact material of the sockets I bought from Erthenvar was phosphor bronze.

Over the next two weeks I exchanged many emails with Edan.  He always responded helpfully and quickly.  Thanks Edan for your help!

On 27 November I received some samples from QingPu Electronics. One of the sockets is pictured above and below.  There is no problem with deformation - I tried every reasonable and unreasonable motion with a plug in the socket and the primary contact always returned properly to the NC contacts.

The first set of sockets I bought - the ones where the primary contact would deform - had their primary contacts made from 0.250mm thick metal, including plating.  The new sockets with the beryllium-copper primary contacts used 0.205mm thick metal.  Both springs were strong, but the thinner beryllium-copper one could be bent much further without deformation.

In both types of socket, the NC contact was made of 0.250mm metal.

These sockets are the best ones I know of.  Edan quoted me an MOQ of 5000 pieces at USD$0.135 - without nuts.  This was late November 2012.  Many factors determine pricing so please ask for a quote rather than assume this is the price in the future. Edan told me that freight to the US was about USD$175, so I guess the total cost with my choice of nuts and washers would be about USD$900 for 5000 pieces, which is more than I will need in the foreseeable future.  Edan told me that an order of the new beryllium-copper model of these sockets had just been shipped to the Erthenvar people, so by early December 2012 I should be able to get the quantities I need from Erthenvar, without great expense.

The socket in greater detail

The WQP-PJ3010BM has much in common with this kind of non-PCB-mount, 3.5mm socket:

http://www.qingpu-electronics.com/en/products/WQP-PJ301M-59.html

which is pictured in images 7 and 8 below.  Edan sent me some samples of these too, which also had the non-deformable beryllium-copper primary contacts.

The difference is that the WQP-PJ3010BM has an elaborate metal piece for the ground contact, which wraps around three sides of the plastic body and has two strong ground terminals which are ideal for PCB mounting.  This provides a generally robust PCB mounting, but there are some problems.

Firstly, the ground piece is not bent exactly at right-angles from its internal peened-over connection with the threaded bush.  This means that when soldered flush with the PCB, the central axis of the bush will not be at right-angles to the PCB.  I haven't measured this precisely, but in the first photo above you can see the bush pointing up a little, with the base of the ground piece aligned closely with the graph paper.

This would be a consideration in several ways:
  1. If the front panel had multiple sockets like this, it would be best to orient them in one way, so that the error in the bushing position was all the same magnitude and direction.

  2. If the bushings were mounted closely in front panel holes, the entire PCB might be offset a fraction of a mm.
However, once the nut is tightened, this would tend to straighten up the socket's body, so the final angle and positioning error would probably be greatly reduced.

A potential solution might be to use the primary contact terminal (lower pin on the first photo, right pin of the last photo above) to pull that part of the plastic body of the socket closer to the PCB top surface.  However, I think this may fail due to all the tension being taken by a small piece of plastic which clips into the primary contact.  During soldering, I think it is possible that this plastic would be softened by heat flowing through the primary contact.

See below for a better technique for fixing this "not-quite-90-degrees" problem - bending the two larger of the four feet out a little, to reduce their height, and so to correct for the ground piece being not quite on axis with thebody and bush of the socket.

Here are two photos of the WQP-PJ3010BM together with the more conventional, non-PCB-mount, WQP-PJ301M style of socket which I think it is based on.


7.



8.

In image 8, the plastic clip which holds in the primary contact can be seen - a wedge-shaped part of the case inside a rectangular hole in the primary contact.  The two sockets have different pressings for their primary contacts.  The WQP-PJ3010BM's primary contact has a small circular depression in it.  I am not sure what this is for. 

In image 8 a gap can be seen between the metal ground piece and the side of the plastic body, near the top of the WQP-PJ3010BM on the right.  This is due to the bend from the ground piece which is fixed onto the bushing, inside the socket, not being quite 90 degrees.  It is tempting to think this could be corrected by graunching it with some pliers, ViseGrips or similar.  However, this could only be achieved by pushing the two little diagonal tabs on either side into the plastic body, or by allowing those side pieces to spread outwards.

See below for a technique of correcting for this "not-quite-90-degrees" problem.

One notable and welcome feature is that there are two contacts for the NC function.  This is bound to be more reliable than just one contact.  Contacts such as these which are simply pressed against each other are subject to a problem know as "fretting corrosion".  There's no easy fix for it, but having two contacts in parallel will be more reliable, since it would be less likely for both contacts to have fretting corrosion at the same time than just one.

Sidebar on "fretting corrosion":

I first encountered this term when searching for an explanation of why the NC contact in some other 3.5mm sockets (like these: PJ302M-57.html) occasionally went OC (open circuit) even when the contact was pressing normally on it.  This occurred in some Devil Fishes in the CV-In socket, the only one of its 3.5mm sockets which used the NC contact.  The fix was to insert and remove a plug a dozen or two times, especially while rotating the plug to cause the contact surfaces to slide somewhat while opening and closing, rather than just pressing in exactly the same location.  After this discovery, I modified the sockets in all new Devil Fishes to provide greater contact force.  I think this reduced the problem, but it was highly intermittent and so it is difficult to be certain about its occurrence.

A 1974 paper "Fretting Corrosion in Electrical Contacts" may be of interest:

http://www.te.com/documentation/whitepapers/pdf/p154-74.pdf


Although the WQP-PJ3010BM is ostensibly a PCB-mount socket, its primary contact and NC contact terminals are solder-lugs, not PCB pins  I think this is due to the history of this design, being derived from a solder-lug socket as shown in images 7 and 8 above.

The wider part of the solder lug terminals is a little over 3.0mm wide.  If a 3.0mm hole was used, the thinner part of the terminal, which actually transits the hole, would be a loose fit, since it is only about 2.0mm wide, so there would be a large volume of solder in the hole, with considerable gap between the hole and the terminal.  I think the best option would be to squeeze the lug part in pliers to make it 2.0mm as well.  With care, this can be done quickly and without twisting the whole terminal.  (See image 10 below.)

The two ground terminals will be fine for PCB mounting.  One is about 1.0mm wide and the other is 2.5mm.

I found that the drawings provided with 3.5mm sockets were not always accurate.  Here are my measurements of the most important aspects of the WQP-PJ3010BM:

9.

The collar does not extend much above or below the 6.0mm threaded bush, but it does extend to either side, as can be seen in some images above and below.  The diameter of the collar is 7.7mm, so we would need about a 7.8 to 7.9mm hole in the front panel to clear it.  This would typically be a round hole, drilled.  However it could be a laser-cut or punched hole in the truncated-circular shape of collar.

Not shown in image 9 is the distance from the four feet (two are visible at the upper right and lower right) and the base of the plastic case.  This is about 1.3 to 1.5mm.  It is possible, as I illustrate below, to bend these out in order to reduce the total height of most of the socket above the PCB.  This bending is also a means of correcting for the not-quite-90-degrees problem.  With a few second's work with a pair of pointy-nose pliers it is possible to flatten the two larger feet (lower right in the above image) and to bend them out.  Bending them out just a little corrects for the not-quite-90-degree problem.  This will take up more PCB space, but that is unlikely to be a problem, since we can't mount other sockets like this or most other connectors right next to the socket.

The round knurled nuts are 2.0mm thick and 7.9mm diameter.  The hex nuts are 1.5mm thick and 8.0m AF (Across the Flats).  The washers (from Erthenvar - I guess they are either from QingPu or are the same as those from QingPu) are 1.0mm thick and 10.0mm diameter.

This gives rise to a number of Panel Strategies with differing "Panel Gaps" - the term I use to refer to the distance between the bottom (left on the above image) of the front panel and the top (right of the above image) of the PCB. 

Before considering Panel Strategies, there are several Nut Strategies.  I assume that nuts will be used, firstly to ensure most of the mechanical stress of plug insertion and removal goes through the front panel and secondly so the front panel's support of the sockets (I assume several are being used) is part of, or perhaps the main component of, the mechanical connection between the front panel and the PCB.  "FP" is Front Panel.

Nut Strategy
Hole diameter
Notes
nsA: Knurled nut

(2.0mm)
6mm
FP presses against top of collar.  The 2mm thickness of the knurled nut enables FP thicknesses to 2.6mm.  However, how do we ensure that the nut doesn't scratch the front of the FP?  Perhaps there is no visible scratching with these round knurled nuts.  However, without a flat-washer or better still a lock-washer, the nut may be prone to working loose.

Maybe the FP is covered in most places by a screen-printed-from-behind polycarbonate faceplate/decal, as in the Devil Fish.  If so, it is best not to have nut pressure on the polycarbonate.  So the polycarbonate label would be punched with a hole to clear the knurled nut diameter, which is 8.0mm.

I wonder where very fine internal tooth lock-washers could be found for the 5.9mm (actual) diameter threaded bush.
nsB: Knurled nut and flat washer
(2.0 + 1.0 or less mm)
6mm
FP presses against top of collar.

The flat washer is 10mm diameter.  It might be best to get some which are less than 1.0mm thick.  If we could get one about 0.6mm thick, then we could use a 2.0mm FP.  This avoids any visible scratching of the FP.
nsC: Hex nut and flat washer

(1.5 + 1.0 or less mm)
6mm
FP presses against top of collar.

This is more promising, since the hex nut is thinner, even with a 1.0mm thick flat washer, we can accommodate a 2.0mm thick FP without any trouble.
nsD: Knurled nut and flat washer
(2.0 + 1.0 or less mm)
7.8mm
Collar goes through FP hole, so FP presses against the top of the main plastic body.

This reduces the Panel Gap by 1.0mm.

Now we have 5.6mm to play with, the 2.6mm taken up by the  the nut and washer is no problem.  We can use a FP thicker than 2.0mm and/or have a lock-washer or split-spring washer behind the FP to help stabilize the nut.

(We can't have a circular 7.8mm hole without the flat washer, or perhaps an internal-tooth lockwasher, on top of the FP, because both the knurled and hex nuts would fall into the hole.  Even if the truncated-circular collar-shaped hole was punched or laser-cut, there would be too little FP material for the nut to press against.)
nsE: Hex nut and flat washer

(1.5 + 1.0 or less mm)
7.8mm As for nsD, but we have an extra 0.5mm for the FP thickness and/or a lock/spring washer.
nsF: Just a knurled nut somewhat recessed in a judiciously countersunk hole in the FP.
6.0mm
This allows still greater FP thicknesses, or with packing washers between the socket's collar or body and the bottom of the FP, a greater Panel Gap thanis possible with the above approaches.


Here are some Panel Strategies with Panel Gaps assuming a FP thickness of 2.0mm. 

Panel Strategy and Nut Strategy
Panel Gap
Notes
psA with nsA, nsB, or nsC (all 6.0mm holes)
11.4mm
The socket mounts conventionally on the PCB and the FP presses against the collar.
psB with nsF (6.0mm holes)
Somewhat greater than 11.4mm, say 12.0mm.
The socket mounts conventionally on the PCB and the FP is some small distance from the collar, with one or more flat washers, split-spring washers or internal/external lock-washers between the bottom of the FP and the top of the collar.
psC with nsD or nsE (7.8mm holes)
10.4mm
The socket mounts conventionally on the PCB and the FP presses against the body of the socket, with the collar going through the FP hole.
psD with nsD or nsE (7.8mm holes) 10.4mm to as little as 9.0mm
Bend the four feet out to some degree (see photo 10), or as far as bending them at right-angles so they do not lift the socket's body off the PCB at all.  For the lowest Panel Gap, it may be possible to cut these feet off rather then bend them out, but I think bending them out would be easier.

With this strategy, it may suffice to cut the Primary and NC Contact terminals short rather than squeeze the lug part to be 2.0mm wide, since the whole socket is now closer to the PCB and we may not need the lug part of these terminals.  However, see below about stresses on the Primary Contact terminal.

When the plastic body is flush against the PCB, with the four feet bent out of the way at right-angles, the tip of the plug may touch the PCB.  Normally there is 16.0mm between the entrance of the bushing and the top of the PCB, but with the four feet bend out of the way, this is reduced to about 14.6mm.  As far as I know, there's no technical standard for the length 3.5mm plugs.  (From the PDF for the Lih Sheng LJE0352 mentioned below, here is a drawing of a stereo plug, with the length specifed as 14.0mm: 3.5mm-stereo-plug-dimensions-from-Lih-Sheng-Precision-Industrial.png .)  Are any longer than 14.6mm? . . . quick measurements of plugs around the workshop found lengths between 14.1 and 14.8mm.  It probably doesn't matter if the plug is stopped by the front panel from going the last 0.2mm or similar fully into the socket, but we need to ensure that there are no copper tracks in that vicinity.  A hole or countersunk area might be better than the plug hittingthe PCB.
psE
12.0 to 14.0mm?
If there was some way of mounting the socket with its four feet 0.1 to 2.0mm above the PCB, this would increase the above Panel Gaps by this amount.

So it looks like we can use these sockets with 2.0mm thick front panels to get Panel Gaps between about 9.0mm and 12.0mm - or has much as 13 or 14mm with some as-yet unknown method of soldering the socket somewhat above its normal flush with the PCB location.

I want to use C&K Tiny Toggle switches, such as the T101, T201 and T211.  The minimum Panel Gap for these is about 8.8mm.  Their threaded bush length is 5.6mm, so with a lockwasher between the FP and the switch, I still have plenty of bush length for the FP and a lock-washer and nut on top.  A single internal-star lock-washer from C&K is 0.6mm when compressed.  With that between the FP and the switch, I would have a Panel Gap of 9.4mm.  This can be achieved with a few seconds bending of the four feet.  That is enough beyond 9.0mm to have no concern about the tip of the plug hitting the PCB.  So this looks most promising:

10.


Stresses on the Primary Contact terminal

One potential problem with these sockets is that the Primary Contact is subject to considerable forces as the plug is inserted and removed.  This terminal is not firmly anchored in the socket body - it can move up and down a little (left and right in the image 9 above).  With a conventional solder-joint, this means the Primary Contact terminal may be placing considerable stresses on this solder-joint.  Especially with lead-free solder, I am concerned about this joint fracturing.

(I have been fixing electronic things since the late 60s and I am very wary of metal fatigue in solder joints.  I want to manufacture gear which will be used for decades without such problems.)

However, the path for the force is not direct.  At the right of photo 8 above it can be seen that the Primary Contact's base has a rectangular hole in it, with the idea that at least some of the force from plugging a plug into the socket be taken by the wedge-shaped plastic part of the case which holds it in place.

One solution might be to place the terminal through a hole in which it is not soldered and then bend it 90 degrees over to another hole to which it is soldered.  However that would involve flexing of the terminal and changing forces where it meets its solder joints.  Probably the best approach is to use a solid solder joint with a hole which is a close fit for the terminal.  A non-circular hole would be better still, but this may not always be achievable, depending on the PCB manufacturing process.  I will think more about this. Musical instruments can get a hammering, for decades.  While a broken solder joint is easy for a technician to fix - compared to something like a failed SMD microcontroller - it would still be best if the fault never occurred in the first place.


Some other 3.5mm sockets of potential interest


Manufacturer, model, photo and link
Notes
EST - Marushin / Shogyo MJ-1235N



http://www.shogyo.com/...MJ-1235N
http://www.shogyo.com/pdf/mj-1235n.pdf

http://www.ab.auone-net.jp/~est/pg-menu_e.htm
http://www.ab.auone-net.jp/~est/pdf/3.5.pdf

Likewise the QingPu WQP-PJ3010B: http://www.qingpu-electronics.com/en/products/WQP-PJ3010B-37.html
This is a similar to the WQP-PJ3010BM but with a non-threaded bush. 

The Shogyo "earphone jack" PDF is from 1994 and specifies gold plating for the contacts and a phosphor-bronze (I assume this is the meaning of "PBs") main contact spring.
Lumberg / Schurter / Pro Signal MJ-355



These are "Pro Signal" (house brand) sockets from element14, catalogue number 126-7382:
http://au.element14.com/.../mj-355/socket-3-5mm-jack/dp/1267382

These appear to be identical to the Lumberg 1502 03:
http://www.lumberg.com/...produktname=1502_03 , the Schurter 4832.2211: http://www.schurter.com/en/.../4832.2211 and the Kobiconn (Mouser house brand, I think) 161-MJ355W-EX: http://au.mouser.com/.../161-MJ355W-EX/?qs=...  .

This uses a coil spring to bolster the spring nature of the primary contact.  The NC contact is a single point.

The terminals are not for PCB mounting, and it would be difficult to squash them smaller.  However they are long enough that the solder-lug ring can be cut off.

The plug may protrude a fraction of a mm into the PCB if the socket is mounted flush.

The body depth is 9.1mm.  The collar depth is 1.0mm. The bush depth beyond the collar is 4.1mm.  So with the collar inside the FP hole, the Panel Gap would be 9.1mm - or 10.1mm if only the bush was in the hole.
QingPu WQP-PJ304M




http://www.qingpu-electronics.com/en/products/WQP-PJ304M-52.html

Similar sockets are also made by other manufacturers.
This would suit a Panel Gap of 12.5mm to something like 14.0mm.  They have a single point of contact NC arrangement.  I think they are similar to these sockets, which I use in the Devil Fish:
PJ302M-57.html

Instead of the ground terminal being at the bottom of the socket in the drawing, it emerges from the top and goes backwards.
QingPu PJ3410 , Lih Sheng LJE0352





http://www.qingpu-electronics.com/en/products/WQP-PJ3410-54.html

See this Muffwiggler forum thread: http://www.muffwiggler.com/forum/viewtopic.php?t=71955
with photos of a similar socket - the Lih Sheng LJE0352-4R, which is a stereo socket with a single switched contact: http://ec.lihsheng.com.tw/catalog/product_info.php?cPath=38&products_id=88

Here is an image I adapted from one of the photos at the abovementioned thread:



I think all the connector manufacturer sites I have seen could be greatly improved with better photos of their products.

Update 2014-05-29: These are available in batches of 50 from Thonk in the UK: http://www.thonk.co.uk/product-category/diy-accessories/jacks/ .
There are three versions of this - two of them stereo and one of the mono with NC contact style we are interested in.

I know nothing  about the internal construction or robustness.

The drawing shows the terminals get wider near the body, so perhaps the PCB seating plane would not be flush with the body.

This may be suitable for Panel Gaps of 13.0 to 14.0mm.

This has a plastic bush with an 8mm thread, and so has a much larger diameter knurled nut.
Lumberg KLB 13 . . . but this is a Stereo socket, without an NC contact



http://www.lumberg.com/main/common/serie.asp?ser=015&cat=4&lang=eng  Schurter have the same socket as 4831.2300: http://www.schurter.com/en/Components/Connectors/Audio-Plug-Sockets/4831.2300

My guess is that the manufacturer of all these is Kobiconn:  Here is the PDF datasheet from Mouser in August 2013.  The Kobiconn parts supplied by Mouser appear to be identical to those sold as Lumberg by element14.  Kobiconn 161-7400-EX 3.5mm stereo jack.

Kobiconn-161-7400-EX-3.5mm-stereo-jack-datasheetf-from-Mouser.pdf

These might be useful, for a Panel Gap of  to 11.6mm12.0mm.  The body depth is 11.6mm and the bush depth is 4.5mm.

However, the terminals are not really suited to PCB mounting.  They are wide and close together, so care would be needed to get the holes and pads arranged properly. 
There are quite a few other sockets, both stereo and mono, in the 3.5mm section of the QingPu site - more than I saw at any other site.

http://www.qingpu-electronics.com/en/products/DIP---3.5mm-Phone-Jacks-21-1.html




Update history:

© Robin Whittle 2013 - 2014 First Principles and Real World Interfaces
Return to the Parts page.