As owner of some DC test equipment and as a big fan of automated tests, I've needed more than once to be able to route the instrument channels to multiple paths. Doing this by hand is of course not convenient, so I looked for better ways to do it. One way would be to use a switching systems such as the Keithley DAQ6510, but this would be costly. On the inexpensive end I could have bought cheap relay board, but it would not integrate seamlessly to be controlled through the Keithley instruments. In-between these 2 price ends I could buy a Keithley 2000-SCAN board to plug it to the Keithley DMM6500 bay, but this would limit me only to the DMM.
This led me to design a board that I could control directly from my Keithley DMM6500 and 2450 SMU digital I/O ports. Later I thought that it would be even more flexible if I also supported other instruments, so I added other control interfaces such as USB, WIFI, 5V and 3V3 logic.
The board can be used in 2- and 4-wire modes and uses the following topology.
I used reed relays mainly because of their longer lifespan than electromechanical ones and better behavior than solid state ones. I made the switching tracks 0.5 mm wide with a clearance of of 1.4 mm so as to keep resistance low and enough clearance for higher voltage. I added an extra layer of isolation just to be safe that current would no take the "wrong" path. And for flexibility I designed it so that I could powered through USB, JTAG, or directly from the Keithley.
Here is a video that shows how the 2450 is powering it and controlling it to test different digits of a Nixie.
I've been thinking on releasing this to Crowdsupply. So I would like to get some feedback from the E14 community. Any comment is more than welcome.
I'll address some of the comments here to keep everything more organized (and also keep editability).
As Andrew J suggested, use cases are probably a good start, because it will allow me to guide the effort on implementing what actually would be useful to the final user.
Target use cases:
A) PCB testing (e.g., Proper operation of the PDN, power safety, detection of shorts or unsoldered joints, etc.)
B) Analog sensor measurements (e.g, temperature sensors, pressure sensors, vibration sensors, or any low frequency analog sensor)
C) Component characterization (e.g, electronic components in ambient conditions or in special conditions such as high temperature, where the same board be could used to measure the component properties and the conditions that affect the component)
Control use case:
A) The board is controlled through the USB virtual serial port using a Python program that controls the board through an arbitrary protocol.
B) The board is controlled through SCPI driver from LabVIEW, Python or any other software that supports SCPI. This could be done through the WIFI or virtual serial port.
C) A TSP program that runs on a Keithley controls the board through the Digital I/O port.
D) The board is manually controlled through a simple webpage.
E) An app connects to the board through the boards Bluetooth or WIFI and manually controls it (this may take a lot of effort as I don't have experience with mobile phone programming).
F) The board controls the instrument(s), and makes the measurements itself through its WIFI.
G) The board is controlled through an ethernet connection (So far it does not support ethernet, but it could be added, but I may run short on pins)
H) Multiple boards are "stacked" to reduce the bench space and used either as a virtual large board or as separated boards.
On some decisions, selection of components and some unsolved issues:
- Price target, as much as I would like to add more features to it, I'm not sure how much would customers be willing to pay for it. Whatever price target I set, the rule of thumb says that final price should be around ~2.5 the cost of the board.
- Of the 3 type of relays, they all have their pros and cons, so its hard to stick to just one type. The 2 ways to support more than one are to either build boards for every type of relay, or to make it possible to solder different relays into the same board.
- I don't like the terminal block that I used. Pluggable block connectors look much more convenient because cables are connected to a block instead of to the board, and also because it is possible to have multiple semipermanent connected blocks which can be quickly plugged into the board.
- In some way this would have to be encased. Laser-cut acrylics could work, but there might be better alternatives.