1
\$\begingroup\$

I'm new to this realm of purchasing and I'm not exactly sure which of the different package names (SSOP, VQFN, etc.) are more friendly than others for basic through-hole soldering.

I would also be willing to use a SMT breakout board, but it seems most of the USB 3.x superspeed muxes I find come in strange dimensions that I can't find any breakouts for (or I don't realize they are in fact compatible).

Alternatively, a more USB-specific question -- is it absolutely necessary I use a mux "designed for USB 3.x" (per product branding) ? Or can I use traditional differential multiplexers? what attributes of these multiplexers make them better for USB 3.x SuperSpeed mux anyway?

Any info would be greatly appreciated.

Thanks

\$\endgroup\$
4
  • \$\begingroup\$ USB 3.x interface on general purpose PCB, what is the purpose of this testing ? \$\endgroup\$
    – user19579
    Commented Dec 25, 2019 at 8:45
  • \$\begingroup\$ @user19579 i've successfully built a USB PD protocol analyzer and sink controller with nothing but PCB, soldered jumper wires, an STUSB4500 breakout, and a Teensy. I would like to make use of the SS data lines on the USB-C receptacle breakout so that my build can supply data in addition to power (for certain voltage ranges), and it would be convenient if data was terminated with a USB Type-A jack, which only has 1 instead of 2 SS lanes like the Type-C pinout, hence the mux \$\endgroup\$
    – ardnew
    Commented Dec 25, 2019 at 10:04
  • \$\begingroup\$ Part shopping questions are off topic, though perhaps this can survive as the main response is that the thing you are seeking cannot exist. If you want to hand build gear and feel you need older package technologies in order to do that you will be limited to legacy USB versions. \$\endgroup\$
    – Chris Stratton
    Commented Dec 25, 2019 at 13:19
  • \$\begingroup\$ Sell on Amazon, and make the money you need to get the tools that you need. There's no substitute for needed tools, just different ways of getting them. I have heard of people doing surface mount on a hot plate, and a frying pan would probably work. Make a friend who can afford things and ask them to make a breakout board for your needed part. Then sell a couple when you get it working, and pay your friend back, with interest. Make a unique flashlight, resistor (or cap) substitution box, inductor assortment, 9-volt replacement, and sell, sell, sell until you get what you need. Good Luck! \$\endgroup\$
    – MicroservicesOnDDD
    Commented Apr 8, 2021 at 2:10

2 Answers 2

Reset to default
4
\$\begingroup\$

I have never seen any USB chips that ever supported DIP packaged parts; not even in the original USB 1.0 era. Generally, support for new DIP parts industry-wide pretty much ended around the late 1990's. I assume the reason you're asking about through-hole is for ease of prototyping, but the world has moved on. All the good stuff is surface-mount now.

The USB Implementers Forum usb.org has all the specifications that go along with the "Designed for USB 3.1" branding. Basically if the device says designed for USB x.x, that means the designers read the USB spec and met all of the requirements.

The USB interface is very particular about maintaining signal integrity, trace impedance, and route matching between the matched pairs. Any USB design is going to require a specific PCB just for that design.

For hand-assembled prototyping, the key difference between surface-mount packages is whether or not the contacts are exposed.

"Friendly" packages such as SO (Small Outline), SSOP (Shrink Small Outline Package), TQFP (Thin Quad Flat Pack) connect to the PCB through small metal protrusions. These can be soldered in the usual way using a soldering iron to contact the pins and the PCB pads. You will need a stand magnifier or +10x glasses, steady hands, fine solder, good quality electronic soldering iron, and soldering flux.

"Unfriendly" packages such as TQFN (Thin Quad Flat Nightmare?), BGA (Ball Grid Array), CSP (Chip Scale Package), or WSP (Wafer Scale Package) connect to the PCB through metal contacts on the underside of the package. In the case of BGA/CSP/WSP these connections are completely hidden from view. Professional electronics technicians use flux and a temperature-controlled hot plate or temperature-controlled hot air wand (not a "heat gun" paint stripper, those will destroy the PCB). These packages are popular with OEM customers such as cellphone manufacturers, due to their extremely high circuit density. But they are really a pain to work with.

These related QA are a good resource:

What tools, equipment, and techniques do you use to solder fine-pitch SMT parts?

How to hold SMD parts in place while soldering?

\$\endgroup\$
3
  • \$\begingroup\$ You hinted at the difficulty of meeting signal integrity requirements, but I must say from my very naive tests, this hasn't appeared to be the source of my usual problems. Granted, I've only been working with USB Power Delivery; the most frequent roadblock was comm speed, keeping up with all the PD handshaking (using Arduino I2C to relay messages to a USB PD controller IC); e.g. USB PD v3.0 requires compliant sink device to read, store, and ACK source capabilities in under ~3ms from when the source message is SENT. Doesn't leave much room for overhead. \$\endgroup\$
    – ardnew
    Commented Dec 25, 2019 at 10:28
  • \$\begingroup\$ For maintaining SuperSpeed integrity though, surely that is just a matter of reducing crosstalk between the lines? There isn't really much a device must do to host the data lines, right? Keep em separate and silent, all of the complex protocol business happens on the endpoints. Would you classify the description of my device as a passive device even if it uses a mux? All I'm wanting to do is select 1 of 2 SuperSpeed pairs to passthrough and connect to a USB Type-A port \$\endgroup\$
    – ardnew
    Commented Dec 25, 2019 at 10:33
  • 2
    \$\begingroup\$ @ardnew - signal integrity is not just about reducing crosstalk; it is also about impedance and length matching. Relatively long pins such as are found on DIP packages would be a very challenging situation for that. \$\endgroup\$
    – Peter Smith
    Commented Dec 25, 2019 at 14:22
4
\$\begingroup\$

a more USB-specific question -- is it absolutely necessary I use a mux "designed for USB 3.x" (per product branding) ? Or can I use traditional differential multiplexers? what attributes of these multiplexers make them better for USB 3.x SuperSpeed mux anyway?

The Power Delivery uses transmission rate of 300 kbps. At this low rate almost any MUX will do the job of muxing the slow CC lines.

The USB 3.x SS transmission happens at 5,000,000 kbps or 10,000,000 kbps, or 10,000 times faster. This kind of frequencies cannot be delivered over through-hole DIP-style packages. That's why all USB SS chips are packaged into small-pitch VQFN or BGA. It is not possible to handle SuperSpeed USB 3.0 data with jumper wires, simply forget it.

To handle PD and SS, you need to develop your own PCB following electrical rules specified for USB 3.0 tracing, and use corresponding quality multiplexer chips.

\$\endgroup\$
1
  • \$\begingroup\$ You say It is not possible to handle SuperSpeed USB 3.0 data with jumper wires, simply forget it. Should I interpret that literally? Or are you referring to hacking together a mechanical DPST switch with jumper wires? The reason I ask is because i have most certainly bypassed muxing altogether and just ran straight jumper wires connecting 2 SS interfaces (of course the SS speeds would only function for a single cable orientation...) \$\endgroup\$
    – ardnew
    Commented Jan 26, 2020 at 6:07

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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