9
\$\begingroup\$

Below is a picture of PCB that has a bunch of copper(?) points for connections.

What are these connections called and how would someone solder wires to these points.

When trying to solder 30 AWG wires it would seem to work if it was a one-off but when there were multiple next to each other it seemed too tight.

Also what does TB stand for (each point seems to be named TBxx)

enter image description here

\$\endgroup\$
  • 2
    \$\begingroup\$ Are you sure your question is "How to solder connections to this board"? It doesn't look like, considering the answer you've accepted. \$\endgroup\$ – Dmitry Grigoryev Feb 14 '18 at 8:54
  • 3
    \$\begingroup\$ TB could stand for "Test Bed" connection, you sometimes also see TP for "Test Point" \$\endgroup\$ – Florian Castellane Feb 14 '18 at 10:39
26
\$\begingroup\$

Those are test-points, in your example gold plated ones, and are not intended to be used to solder wires onto them.

Normally they are used with a bed-of-nails using spring loaded pogo-pins to make contact to the circuit using testing equipment. Pogo-pins come with a variety of tips for contacting with various features on the board.

enter image description here enter image description here

Note the use of the large pointed pins in the above image that go through the mounting holes in the board that act to align the board with the pins. The big fat ones are keyed to the edge of the board to get it aligned close to the smaller ones that go through the holes.

The alignment pins are longer than the pogo-pins so the board is in the correct position before they make contact.

Sometimes double sided, "suitcase", fixtures are used...

enter image description here

\$\endgroup\$
  • \$\begingroup\$ Trevor -- I've wondered about building fixtures like this before. Who exactly builds them (like your PCB fabricator?)? Any idea how much these cost? Our boards are so simple, we have never needed a "bed of nails" but I've wondered. \$\endgroup\$ – Leroy105 Feb 13 '18 at 21:41
  • \$\begingroup\$ @Leroy105 everywhere I worked we built the boards in house and had a department that designed and built those "tools" along with the test equipment software . A good PCB fab shop will have that technology in house too. Of course, at a price. \$\endgroup\$ – Trevor_G Feb 13 '18 at 21:45
  • 5
    \$\begingroup\$ Just to be pedantic, the sharp-pointed pogo pins pictured typically are used with through-hole test points on a board. There are also flat-headed pogo pins better suited to boards like the OP's. \$\endgroup\$ – Lee Daniel Crocker Feb 14 '18 at 1:32
  • \$\begingroup\$ @LeeDanielCrocker good add. \$\endgroup\$ – Trevor_G Feb 14 '18 at 2:06
  • 3
    \$\begingroup\$ @Leroy105 - The company I work for does this as their primary business. The fixtures vary wildly in cost depending on what they do. Someone put together a white paper showing a lot of the basics: circuitcheck.com/images/pdf/… Many of the pogo-pins are provided by companies like QA (see their website for lots of info about the styles of probes used - QAtech.com). \$\endgroup\$ – JDB Feb 14 '18 at 3:53
11
\$\begingroup\$

While it's true that these would be contacted by pogo pins rather soldered in production, for a personally owned unit (or indeed, for the original firmware developers') soldering would be quite reasonable and likely.

Fine gauge silicone insulated wire is best for this, since you avoid the whole complication of potentially melting the insulation while soldering with only a couple of millimeters of stripped length, and you can with a little care put a crimp connector on the other end to connect to your programmer, USB/serial adapter, logic analyzer, or whatever. But you can also use wire-wrap wire with a bit more care. Others prefer magnet-type wire insulated with a solder-through insulation.

While the pads are close together, compared to what they could be, they are really not that tight at all - be glad you aren't trying to pick up the signals from the 0402 resistors or the pads of that little QFN chip.

Pre-tin the wire (I tend to start overly long and trim to about a millimeter exposed after tinning), possibly pre-tin the pad, and you just need to hold the wire in place while you touch it for a second with the iron.

\$\endgroup\$
  • \$\begingroup\$ @Crhis_Stratton +1 on the 0402 resistors reference :) guess gotta count my blessings \$\endgroup\$ – pzirkind Feb 13 '18 at 17:22
  • \$\begingroup\$ ... and then make sure not to pull or twist the cable by any means, as it will come off very easily, or, worse, take the bad with it. If possible, somehow glue the cable to the board before the final soldering. \$\endgroup\$ – AnoE Feb 13 '18 at 20:10
  • 1
    \$\begingroup\$ You should call out a specific gauge that is "fine gauge" -- I remember trying to tackle this scenario once and had to order a few different spools to get a good one that would not rip off a pad. I have found 30 gauge + hot glue on the wire soldered to the pad, to prevent ripping off the pad. \$\endgroup\$ – Leroy105 Feb 13 '18 at 21:44
4
\$\begingroup\$

those are gold plated test points

you would not normally solder wires to those

but if you really, really, really have to, apply solder to the pad, then apply solder to end of wire, then join the two

\$\endgroup\$
  • \$\begingroup\$ what is the best way to temporarily connect to these test points \$\endgroup\$ – pzirkind Feb 13 '18 at 17:08
  • 1
    \$\begingroup\$ search for "bed of nails" \$\endgroup\$ – Vladimir Cravero Feb 13 '18 at 17:08
  • 1
    \$\begingroup\$ did you not read @Trevor_G's answer? \$\endgroup\$ – jsotola Feb 13 '18 at 17:09
  • 1
    \$\begingroup\$ reading it now, the page auto scrolled to the bottom and did not see it at first \$\endgroup\$ – pzirkind Feb 13 '18 at 17:09
4
\$\begingroup\$

I had a design, built by someone else that used extremely small test pads like this. I needed to have the pad constantly hooked up to an oscilloscope, so I needed a permanent wire attached to the pad.

I found that it was VERY EASY to rip the pad off and kill a board when soldering a wire to it.

What I found is using a 30 gauge magnet wire (yes, you have to tin both sides) you can solder on to the pad, and it isn't so heavy to rip away at the pad.

To prevent ripping off the pad as your wire bends around, if you have it connected to a scope or something else, was to put a dab of hot glue on top of the wire. The hot glue isolates the wire from bending and flexing, so you aren't pulling at the pad.

I learned this the hard way killing many boards, while testing this design.

IMO: if you think you need a test point, consider a thru-hole first.

\$\endgroup\$
2
\$\begingroup\$

For long rows of evenly spaced test points, you might try to solder pin headers to them - as long as the test points are laid out in one of the standard grids (1/10", 1/5", 3/20", IIRC). SMD pin headers may or may not work - the geometry of a test point is different from that of a SMD pin header solder pad. A (regular) wire-through pin header might work in a pinch, but will have a less mechanically stable solder connection to the PCB.

In any case, test points are meant for temporary connections, not for permanent connection with useful life expectancy. It might help to additionally glue the wires to the board for added stability.

Note that test points might not only be designed to be used for "live tests" of the operating circuit, but also be meant for "off-line" tests of traces or single components. Adding a wire to such "off-line" test points may cause the circuit to malfunction (because of the added capacitance).

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.