Alternative SeaTalk wiring

[Sailoog]

Update

Dylan has already sent some pictures of the signals coming into the transistor and we can see a difference between the real world measurements and our simulations or our working setup in the lab.

Duncan, Boatingbaileys, any reports are still very valuable. If you get to the same behavior as Dylan that would be good news:

- Data flowing on the Seatalk1 bus without errors in any device.
- Seatalk1 LED blinking.
- No data in Signal K server.

[Duncan]

Hi
Having checked my SeaTalk bus, I did have a wiring issue (sorry). Having corrected it, I'm now getting the same test results as Dylan.

- Seatalk 1 bus isn't reporting any errors
- SeaTalk 1 LED is now blinking
- I'm still not seeing any data in Signal K

Duncan

[Sailoog]

Hi
Having checked my SeaTalk bus, I did have a wiring issue (sorry). Having corrected it, I'm now getting the same test results as Dylan.

- Seatalk 1 bus isn't reporting any errors
- SeaTalk 1 LED is now blinking
- I'm still not seeing any data in Signal K

Duncan

Thanks!

It doesn't look like it but it's good news

Still working on the solution. Stay tuned. Any other issues with the rest of the features are also welcome in a new thread.

[Boatingbaileys]

Little late the to this one but I’m getting the same results. Flashing seatalk light with some off periods which is good to see but I don’t get any data in SignalK. The Seatalk instruments all seem to be working correctly.

[Sailoog]

Thanks Boatingbaileys, that is good.

We are waiting for Dylan's board to travel back to Adrian to try to apply a mod. I will update you ASAP.

[itmustbeacamel]

Hey all! This is Dylan reporting in… just got the board reinstalled after Adrian’s mod. We have first light! (or, electron…?)

   

[Adrian]

For those playing along at home:

I ditched the opto-coupler and took some inspiration from a supposedly "official" Seatalk1 circuit:
https://www.electro-tech-online.com/thre...ce.130122/

The Seatalk1 circuit on the MacArthur HAT v1.1 looks like this:

   

That worked fine on my and Sailoog's bench, but failed again in the field. Some probing by Dylan revealed, that the issue was transistor Q5 not fully turning off when SEATALK_DATA goes low. This is fixed by replacing resistor R21 with 20K.

This lowers the (theoretical) voltage level on the base of Q5 from 0.5V to 0.3V, giving more headroom to the transistor's Vbe(on) which is specified as 0.58-0.7V.

[Boatingbaileys]

Sounds good thanks for the updates. I’ve not got the board here at home, but is it accessible enough to modify ourselves? And do you want us to try and give further feedback?

[Sailoog]

We are almost there

I feel very optimistic, but we need to confirm that everything works in the 3 open cases that we still have because they are really representative of what we will find in the field.

Dylan: Everything seems to work electrically after latest mods in v1.1. Analyzing data. This is a very important case because it is one of the oldest network and devices we will find.

Sailoog: My setup was already working with v1.0 and v1.1 but we need to confirm that mods in v1.1 still work for really small networks of just 1 device.

Boatingbaileys: v1.0 working, v1.1 not working. This is a complex network containing multiple devices and converters working with all protocols NMEA 2000, 0183 and Seatalk1. Also very important because it is the most common scenario.

Duncan: This is the most important case to confirm because both v1.0 and v1.1 were failing and some devices in the network stop working after connecting the HAT. This is a behavior we have often observed using the old way (optos and yellow and black wires), actually it was the reason to switch to the new way (transistors and yellow and red wires). Failing devices were already fixed with v1.1 and all should work after applying mods.

R21 is an SMD component that must be replaced by a new 20K resistor. The problem is that R21 is very small and very close to other components. It is necessary to have some experience working with SMD components and some specific equipment (hot-air rework station...).

   

If you feel comfortable with that, you can give it a try. If you need help you can send the boards back to Barcelona. I have the equipment but little practice, so I will probably end up sending all 3 boards to Washington because Adrian has both equipment and experience.

[holgerw]

i would modify R22 to half or put another 3K3 on top of it.
The lower Impedance of the circuit is better against Peaks.

ym2cents