Update
Ok, it may be related to the adhesive I'm using to glue the toroids on the 3D prints.. I'll wait till tomorrow for it to cure and test again..
I'm trying to build a portable/battery powered core rope ROM after successfully having built a "desktop" one.
This is the device I built:
I'm having trouble however with the waveforms of the transformers when the 2 AA batteries start to drain, with just one or two transformers it works fine until 2.5V and even less, but in the case of the following set of transformers at 3V already the one with the blue line does not trigger the latch (ch 4 on the scope).
I've been empirically swapping toroids with more or less windings to see how well I can get it to work at lower voltages but now I'm just plainly confused because the three transformers shown on the scope shot are made from the same toroids(datasheet), with the same amount of windings. I'd understand if there was a bit of variation (i.e. maybe I counted the windings wrong or there is some fabrication tolerance) but this waveform is just way too different. I wound a new one and it had the same lower waveform, however if I swap a "good" toroid with a "bad" one in position, the waveform follows the toroid. So I guess it is something related to the winding or material of the toroid.
- Is this something design related or fabrication related? (i.e. the permeability is way lower on this one and I have to test a lot of them?)
- Is there something that maybe I'm not taking into consideration? Do the order of the toroids matter? Does the distribution of the windings around the toroids matter? (although they are identically wound as far as I can tell)
These are the 3D prints that the toroids are glued on:
Here follows a block diagram of the system that produced these waveforms:
simulate this circuit – Schematic created using CircuitLab
(Sorry I could not find a better way of drawing transformers with multiple primaries)
p.s.: this just wasn't a problem on the desktop version because I could drive them a lot harder since parts count and power weren't issues