Pogo pins and I2S speed


Well-known member
I have the following problem:
I wanted to test some audioboards (using CS5361 , own design and pcb) with teensy 3.2.
So, I use spring load pins (from Mill max) (Digikey ED90345-ND) to test the board without soldering the board to an T3.2 adapter board. Obviously final configuration is soldered.

Doing so, I run into problems as frequently the most significant bit was not transfered correctly (e.g. missing for negative number). (so for 33kHz sampling frequency: MCLK, BCLK, FS were 17.0667 2.1333 0.0333 MHz)
only reducing the sampling frequency to 25 kHz (MCLK, BCLK, FS = 12.8000 1.6000 0.0250 MHz) I got error-free data.

I assume that the Pogo pins could limit the transfer rate, in particular, they may have a LP characteristics that would result in the incorrect transfer of the first Bit (MSB).

Unfortunately, I have no oscilloscope that could show me the waveform of the I2S signals.

- Anyone out there who has experience with Pogo pins and high speed digital data transfer?

- should pogo pins be used tight (full pressed) or with little load?
- should pogo pins be used tight (full pressed) or with little load?

Yes, the spring should be compressed so the pin applies enough force for a low resistance connection.

For testing Teensy, we use the pogo pins with a fairly sharp point. If you inspect a brand new Teensy under a microscope, you'll probably see the tiny indents those points leave in the bottom-side SMT pads. But I usually keep every signal under a few MHz.
Pogo Pins come in many different head geomitries and sizes. The ones you linked have a round top, which is really only good if you have very delicate surfaces that cannot be scratched by the probe. for a good contact, you want sharp edges or points to break through oxide layers or other surface contamination that may have built up on the contact surface. If you are probing through-holes i recommend a pin head that looks similar to a metal drill. For example here: http://image.dhgate.com/albu_644660573_00/1.0x0.jpg look at the P75-LM2 and -LM3 heads, these are ideally suited for plated holes and vias. The shape also helps centering the device under test before applying pressure.

Edit: You should not use pogo pins fully pressed, let the pogo-spring define the pressure, don't push to the pogos mechanical end. Excessive force will degrade the life of those sharp edges and the springs are tailored to give good lifespan and pressure.
Last edited:
@Paul, Ben,
thank you for advice and link.
In fact, I wanted to 'replace' a standard 0.1" header by pogo pins, so I can test multiple add-on boards without soldering, or better desoldering after testing.
I will see if I can get somehow an analog waveform to deduce max frequency through pogo pins.

I may also play with I2S protocol,
CS5361 is set to have a leading empty clock cycle, So I will see if disabling on Teensy I2S the "Frame Sync Early" will provide a workaround.
once connections are soldered I expect a more robust operation.
I use "fixtures" for test and programming with pogo pins. Fixture is Plexiglass pieces screwed together to form a fixed platform and a moving platform. PC board under test (UIT) gets sandwhiched between the plexiglas halves.
My fixture is slightly similar to this

But with 4 guide posts and anti-static plexiglass. The red-handled clamp thing is the same in mine. I can't show a photo of mine for proprietary reasons.
The UIT board is either a normal PCB or a flexible PCB.

Don't try to mate pogopins with soldered pads on the PCB.. need dedicated gold plated dots. Solder will bend/contaminate the pins.

Very accurate positioning is needed ... a few mills, to be flat plane and keep the pins centered on tiny copper dots on the board made to interface to the pogo pins. The pins solder into a PCB themselves and this PCB is hard mounted to the fixture. Precision needed here too, for the depth to be consistent.

The two halves of the fixture (planes) move up and down on stainless steel posts that pass through precision round bearings so there's no slop.
The plexiglass is anti-static but opaque. Wish it was transparent.

These fixtures are used both for development and production test to config and program the MCU code and constants in flash. Using SWD to program with automation software - due to the volume. This board has a 1MB cortex M4 and GB of flash and many other chips/sensors.

My concern is the pogo pins will wear out too quickly. But that's a good problem.
Last edited:
if someone is curious, here is a picture of my setup
the audioboard (to the right) is hold in place by 4 plastic bolts.
The pogos were soldered after placing the audioboard and adjusting the pins to fit to the through holes.

The final version has two audiocard stacked one on top of the other, so I only wanted to assemble the final product, if the audioboards work correctly (assembled by myself, so errors are possible).
Do you have decent ground-return path alongside all the I2S signals? - without that you may have bad ringing on edges.
Do you have decent ground-return path alongside all the I2S signals? - without that you may have bad ringing on edges.

FYI - if you look at the earler message, it is from about 7 years ago. Not sure if the messages from CNOMAX should be condiered spam or not. Shows currently they have 3 posts, I thought earlier it was closer to 6, so probably several have aready been reported.
Indeed, this conversation was back when Teensy 3.5 and 3.6 were still in development and Teensy LC & 3.2 where the newest products!