IO expander alternative

As far as I can see in the photo of the board in Amazon...
The DC- trace extending to the optocoupler side (maybe @koraks saying it), it connected to pull-down the gate output side of the optocoupler.

Perhaps this board can provide insulation.

In that case, isolation should be possible.
It's a bit confusing there are two separate Amazon links we seem to be talking about. A green board and a red board. On the red board (IRF540), it seems that on the outputs, all minuses are separate from each other and separate from GND on the input side. On the green board (60N03 mosfets), the minuses on the output side all connect to each other and I think also to the GND on the input side. On both board, all GND's on the input side connect to each other.
Neither of the boards uses LR7843 as a switching element, so apparently you've got a third board which I cannot see the PCB of and I have no schematic of it obviously.

So the above is all second-guessing based on indirect information; your measurements can only tell the story the way it really is.

@koraks No the green board definitely uses LR7843 it even is so in the picture. i have not check continuity on the red board IRF540's however the green board, the grounds on both the output and input sides are not connected.

If I zoom in on it I see FDD60N03, which is a common choice for boards like these.
https://m.media-amazon.com/images/I/61gmapsTMML.AC_SL1001.jpg

\Edit: looks like they posted pics of different versions of the same board in the ad, since another pic does show the LR7843 mosfet:
https://m.media-amazon.com/images/I/71396clHKYL.AC_SL1500.jpg

That makes isolated drive possible.

1 Like

wow your right, too bad they don't just post the schematic.

Yeah, that would make things so much more easy. It's not like it's super secret intellectual property anyway. As you now realize quite well I think, it really helps to have that kind of information when deciding what you need and how to implement it!

PS: now you probably also realize why we keep asking for schematics on this forum, no matter how simple they may be - there are often these little details that seem insignificant that turn out to be quite essential :wink:

1 Like

This is because why we build our own PCBs and that's why you buy discrete components. :wink:
Causes a lot of confusion when using a board that doesn't provide the correct schematics from seller and isn't guaranteed to consist of certain components.
I wouldn't be surprised if the board was sent in a modified to don't provide isolation without your knowledge. :expressionless:

until i get a machine to mill pcbs i dont have much choice in the matter. i seen mosfets,diodes,optocouplers and said ill take my chance. i bought the red ones to destroy and run torture tests on. i think the green ones are much better quality. if this was a single mosfet project then i would absolutely build it. i actually have mosfets to build one i was just feeling lazy. I wont even have to reprint my enclosures with the green ones, the drop in right where the relay modules sit. with @chrisknightley solution, i dont even have to change the HIGH,LOW logic in the code for to work properly anymore.

Mainly I outsource PCBs to Chinese manufacturers.
They make two-layer PCB up to 100mm x 100mm for a few dollars... :yum:
But, well... For a simple prototype, I use the perf board still now.

The problem have been solved, so I'm getting out.
Have a good time! :+1:

1 Like

I'm an old-fashioned guy, I just etch them. The equipment is far cheaper, but it requires a bit more skill. Still, it's not very difficult. I tend to stick to double-sided boards and trace widths >0.25mm (10 mil), which is perfectly feasible with low-cost equipment and materials.

Outsourcing like @chrisknightley does is certainly viable as well, but I personally don't like to wait around for some boards to get here from China. A simple board I can layout, manufacture and populate in <24 hours here at home without breaking out a sweat.