Posted on August 15, 2014


Njay: Is this diagram complete? The turn-on/pre-charge block is more complex, right? And I think there was a connection for the RasPi to control it?… We need up-to-date diagrams to know where we are.


Vnevoa: that turn-on connection was never implemented, but the rest is all there. Basically there’s one transistor missing. You have those schematics, I can’t find anything over here. We spoke about putting it under source control, but that never happened?… As far as I recall, the first thing is to implement the current limitation, otherwise we’ll never get anything tested. I’m a bit tired of re-soldering shit. 🙂 25Amps, if you please. 😀

Njay: Yep, I got the schematics, I wanted to place myself on what we have, you know we made some small changes some times. These will do. Ok, the bridges first. We have to see what we are going to do about the communications.

Vnevoa: The difference between your schematics and reality was just that CPU transistor we never put in, and a protection diode I had to add. About the comms… I can put Amélie [a microcontroller from another project] on it. While we’re at it, we’ll implement an intrinsic safety protocol.

Njay: Amélie? Wasn’t that an electrolysis controller? I don’t know if we shouldn’t build that I/O expander we talked about once, it would give the RasPi 5V outputs, convert voltages, and eventually convert I²C to something else. I need the comms protocol defined so I can finish the bridges, their resources are pretty reduced. We should attack these problems first.

Vnevoa: Ok. Port expander it is. 🙂 I thought Amélie’s microcontroller was good enough.

Njay: Oh, that’s what you meant. Yeah, it’ll do. We’ll need something clever enough for a protocol, besides just “port expanding”.



1 – The comms between the RasPi and the bridges has to have galvanic insulation to avoid the resets we always saw with all the bridges. How about you use your trademark opto-insulator scheme [also from another project]?

2 – I want to create an intrinsic safety mechanism: in case one of the bridges fails to respond, both of them must stop their motors. It can be as simple as a single wire between them. Or 2, because of opto-insulation. The lack of signal (from the RasPi or from the other bridge) causes them to brake and cut the signal to the other bridge. The signal should oscillate, would act like a watchdog. If any wire or micro fails anywhere, both should stop simultaneously.


3 – the emergency stop should be a downwards ramp of negative torque, but I’ll be happy with a simple short-circuit to the motors (they’re not very powerful).

4 – 20 amps is not enough, braking is too weak. I’d like to up that a bit. You told me it couldn’t be done because of the MOSFETs on the bridges; is it just a case of switching out the MOSFETs or would it require more reworking? Please consider it, but if it’s too much entropy forget it, we need to close problems, not open new ones. 🙂

(to be continued…)