Compatible wire-to-board STLINK-V3MINIE connector for CN2 (BTB, AVX 009159010061911)

[Spaghetto]

Hello forum,
I'm designing for my company a very-low-cost parallel (10 tap) programming tool for STM32F4 target.

The solution identified is an SBC with an isolated usb hub where 10 STLINK-V3MINIE are installed. Actually we've successfully tested this setup in lab, where 10 instances of STM32_Programmer_CLI.exe (pointing each one to one of the 10 different programmers' SNs) are launched via cmd.exe.

Every target board is fixtured to the assembly via a bed-of-nails and my actual difficulty is to connect each nails to STLINK-V3MINIE SWD lines: I can't solder directly the nails to the STDC14 cable since I should use the flat from IDC as pigtail and I don't like the ridiculous AWG and, for various reasons, I don't want to customize the STLINK-V3MINIE soldering, e.g., a female header to card-edge pads connector (as I did for my lab tool).

We're considering in using the board-to-board connector with is designed to use with the AVX 009159010061911 which is a board-to-board connector but I haven't find in the AVX catalog a compatible 2.00mm top+bottom wire-to-board (wire to card-edge) alternative that would enable the bed-of-nails cabling to reach a connector which I could easily remove from the programmer for maintenance.

Anyone have already find a solution for that?

Thanks in advance,
Luca

[Pavel A.]

For the CN2 (on the edge) use this connector, specially for low-cost edge connection. No pins, no holes, no cable.

[Spaghetto]

Thank you Pavel for the photos,

I've always been curious about that b2b connector and never saw one, but I'm afraid you completely missed the point of my post.

Thanks anyway,

Luca

[S.Ma]

Beyond the Board to Board Kyocera adapter, the footprint is also compatible with DF11-10DP which enables, board to board, board to cable by inserting and soldering it the same way. Edge connection is better for signal integrity.

[Spaghetto]

Hi @S.Ma , thanks for the hint as far I see the DF11 series is just a 2.0mm "Pitch, Double-Row Connector" with no card edge options, If I've understood correctly your message you're suggesting to solder a 2mm pitch connector to the STLINK-V3MINIE programmer in order to make it compatible with DF11 mating parts. Please correct me if I am wrong.

I already do this in my lab even with 2.54mm pitch headers (which fit with an acceptable tolerance), pic shown below, but as I said in my original post I don't want to customize the STLINK-V3MINIE boards since this solution will be installed on the other side of the globe respect to us and we want the end customer make maintenance locally without sending them spare parts (modified stlinks) but ordering themselves online (original stlinks).

This is why I was asking for a compatible "wire-to-card edge" connector. I hope I made myself clear, sorry non native speaker here.

Luca

[Spaghetto]

Long story short, given the lack of commercial solutions with the B2B CN2, we went with a COTS solution of this type (the STDC14 cable is already included with the programmer):

https://www.adafruit.com/product/2094

To which I would pair a Hirose HIF3B connector (a female "idc style" connector that accepts wires).

HIF3BA-20D-2.54C(63) + crimp AWG22-26 HIF3-2226SC

Luca

[Ozone]

I just soldered two 1x5 pin connectors to each of the edge connector pads of the V3MINIE, and used patch wires to hook it up to the debugee.

Of course one can make proper cable to connect it to the test pins of the bed of nails.

On a related note, my company, which prepares the test equipment for our electronics subcontractors, uses a Segger programmer for this purpose.

[Spaghetto]

Hi @Ozone, we share the same professional background!, Gang Programmers (MSP Gang, as Segger Flasher ATE),  were categorically excluded due to their costs compared to the V3MINIE (which hits 12€ per channel) which in the lab proved to have performances compatible with the application. Since we have no experience in using them in production lines (where LLN & Murphy apply), I am always afraid of robustness of the I/O lines this is why I'm giving the no customization spec in the sys (as I have extensively said above) to contain the MTTR.

[Ozone]

> Gang Programmers (MSP Gang, as Segger Flasher ATE),  were categorically excluded due to their costs compared to the V3MINIE (which hits 12€ per channel) which in the lab proved to have performances compatible with the application.

And here you named the reason why we don't (and can't) resort to such a solution - compatibility.
A lot of those products use Cortex M devices from different manufacturers, and ST not amongst them.
And of course the costs of maintainance, replacement, and possible loss of production have to be considered.
In my case, we speak about a total throughput of a few thousand per year.

About two decades ago (and with another employer), I developed and installed inline test equipment for large-scale electronics manufacturing customers, i.e. a few thousand items per day. Initial costs of components like gang programmers are secondary in this environment.

[Spaghetto]

During the last 10 years, this customer compulsively moved manufacturing plants from country to country in order to cut labor cost, getting to the point where nowaday the cost of a manual load in\load out is less than an automated line even for thousand of units a day he produces (maybe the answer fits in the fact that after the test&programming the board is manually installed, by the same operator, into its chassis).

At the same time the same customer ask for a tool (which cost is the very only issue) that targets only a specific board (which mount a specific STM32F4), in short this will be really far from a general purpose equipment.

To be honest, over the years I’ve lost track of the reasons behind the choices that customers make. All I’ve realized is that KPIs are really subjective, depending on who is asking.