"Bare bones" version for classroom use?

Intros
1

[I’m a SWE/EE with a UI/UX background and a masters in educational technology from MIT]

The newest Jumperless v5 looks great! I particularly like that it can work without a computer attached.

Have you considered a single PCB “bare bones” version for classroom use and hobbyists on a budget?
You would know better than me… where are the “pain points” to making this product family more affordable?

The price point of $350 seems high to me. Even your old version with fewer LEDs costs $300. If it cost half that, I’d buy one immediately. And schools are sensitive to price too. With education in-mind…

Thinking out-loud about paths to cost reduction: True, there a lot of crosspoint chips, but at just over $1/each that’s a BOM of <$14 for the whole matrix. (x2.5=$35 contribution to retail price)

Daughterboards, connectors, and generality, while great for the developer community and prototyping, these cost money, of course.

1 Like
2

I revised (edited) my original post a few too many times instead of replying to my post. This is a conversation about cost reduction. Yes, I know right now is all about getting the flagship Jumperless V5 out the door, but I’d like to understand what knocks this product out of the $150-or-less pricepoint market. What has to stay, what is expensive but can be eliminated for a “little brother” product aimed at schools? Getting rid of daughterboard(s) and connectors seems like a good start, but you have dealt with the factory, and you know where all the costs go (and don’t go). “Penny for your thoughts?”

1 Like
3

Hey Tim,

Yeah, I’ve put a lot of time and thought into getting the costs down on this thing. Working with Crowd Supply’s general recommendations for Total COGs > Sale Price would have put this thing in the ~$425 range. Which puts it into another mental category from “oh that’s expensive” to “no way in hell”, so I just take really tight margin in the hopes of volume making it sustainable. And I just really like people using things I made so it’s worth it for me.

I work with Elecrow and the other manufacturers pretty closely, and I have no reason to believe they’re ripping me off. Sure, if I was 3M and had some weight to throw around, I could put the screws to them and get costs down. But we’re genuinely friends and I want them to succeed as well, they’ve helped me out a lot.

With a ton of volume I might be able to get it down a bit without changing anything, but nowhere close to $150.

The custom spring clips cost 80 x ~$0.16 each (assuming tooling gets amortized to nothing) and have to be placed by hand into the board shells ($7.50) before being soldered.

The LT1054 charge pump for the power supply is straight-up highway robbery at $4.30 plus the low-ESR caps to support it.

Those are the very basic parts that it needs to work. And I do like the idea of eventually making a non-flagship, stripped down version for schools that could possibly hit the $150 mark you set. Maybe using an external display instead of the LEDs ($22.50).

The eventual dream is to put all of this on a piece of custom silicon, which would get unit costs super low (and be much closer to “ideal” wires), assuming volume in the hundreds of thousands. I’ve talked with a few people in that field to confirm is I’m completely out of my mind, and they assured me it is possible to pull off, at least from a technical perspective (not losing a literal million dollars is an entirely separate issue).

But yeah, I’ll take that under consideration and think about what can be reasonably stripped out/optimized when this version is shipped.

(Or I can just find a way to harvest user data and sell the hardware at a loss like most consumer electronics nowadays, ha.)

4

Thanks for sharing your thoughts. This gives me a lot to think about, so I should resist the temptation to “shoot from the hip” reply, and actually think about it some first.

1 Like
5

Here is one bare-bones idea:
Single PCB.
No RPi daughterboard, solder 2040 (or newer Pico) directly on the PCB.
No adjustable power supplies, just power from USB.
No LEDs (-or- maybe only one LED per row)
Only the crosspoint switch
No custom breadboard (this is where I think the big savings will come from)
-Instead- provide a standard (high quality) breadboard, connected to the PCB with little jumpers soldered at one end and inserted into the breadboard at the other.
Require the user to use a computer or tablet to program the crosspoint and see the interconnect topology, take measurements, etc.
You probably want to use a breadboard with no power rails.
Last but not least, have the bare-bones (4-education) Jumperless be compatible with the premium Jumperless–all the same software, share designs, etc.

1 Like
6

Interestingly, something like that was just on Kickstarter. He’s using a CY8C58LP PSoC, which has a sort of built-in crossbar matrix. It’s a cool chip.

Not sure how he’s sticking on the breadboard, for my older prototypes I used Z-tape and screwed a regular breadboard down really tightly.

7

Ah, (rings a bell)… you included Sandwizz in your comparison table.
First thought: 500 ohms is definitely too much resistance. 85 ohms feels tolerable.
Maybe you’re right: a premium functionality product launch > a race-to-the-bottom.
There are enough railroad enthusiasts who will pay $350 for a model locomotive.
Admittedly, I’d rather spend “real money” on something usable (and that will get used), instead of something inexpensive that will sit on the shelf.

One thing I’ve found is that shipping a “hello world” demo–with parts–goes a long way to making sure customers will get started and engaged.

Anyway, I think you could eventually leverage your growing Jumperless community to introduce a software compatible “little brother” product. I do think that all the custom plastic, hand assembly, and daughter-boards are killing your MSRP. But I agree, you don’t want to sacrifice usability.

1 Like
8

For my purposes I would just want a single board with headers. I really just want a matrix switch to use in other projects (although I’ll certainly use a breadboard version for prototyping some things).

1 Like
9

I usually ask the factory not to do final assembly on a couple boards just in case someone wants to solder their own thing on there. They were originally for kits but I stopped doing that when every hardware issue was coming from them.

I think I still have like 2-3 original Jumperlesses without the breadboard soldered on.

But yeah just having a general purpose matrix you could stick into other projects would be really cool. Maybe I’ll do something like that in the future.