Close

Pre-Biased xPDT relay configuration

A project log for AMBAP: A Modest Bitslice Architecture Proposal

Trying to unify and simplify a minimal architecture for various implementation technologies...

yann-guidon-ygdesYann Guidon / YGDES 11/21/2016 at 02:506 Comments

The 3rd drawing from the last log (Dual-voltage power supply) asked for a verification.

The CC-DPDT circuit works well, from 2.5V to 3.9V :-) [yes I bias the capacitor in reverse, booooh]

If it works well under 3V, there is no reason to tune the 12V power supply and I can string 8 relays ! Of course I should account for about 10% resistance/temperature drift due to Joule heating in the copper.

The other thing to consider is power-up: half of the relays are "on" while the others are "off" so at least one on/off cycle is required before the system is in a known state.


A 4-relays version is tested successfully between 5V and 7V. The capacitor is connected to either 0V or Vcc, I haven't checked the voltage influence but it makes sense to keep it that way.

6V seems to be the center voltage, though that might vary with relays. For example, two relays can't hold a value for a long period when close to the end of the range. OTOH, they are stable at 6V (not too sensitive to vibrations for example). Out-of-median specs should be paired : a higher coil voltage can be paired with a lower-voltage to create a median-voltage pair...

Discussions

Eric Hertz wrote 11/21/2016 at 04:36 point

I'm lagging on the log-updates, and y'all capacitor-abusers probably already know, but I just came across Electrolytics that aren't polarized, apparently they're commonly-used in speakers. ...now who was that other cap-abuser that's been in my feed recently?

  Are you sure? yes | no

Yann Guidon / YGDES wrote 11/21/2016 at 04:41 point

NP Caps are larger and more expensive...

I got thousands of polarised caps about to come in the mail so I must adjust to the constraints ;-)

  Are you sure? yes | no

Eric Hertz wrote 11/21/2016 at 06:24 point

I figured... just thought I'd throw it out there.

  Are you sure? yes | no

Yann Guidon / YGDES wrote 11/21/2016 at 04:44 point

The other solution is : 2 caps and 2 diodes

   +               +
--------||-------||---------
   |---|<|---|---|>|---|

(sorry about the formatting)

but it's a bit cumbersome...

  Are you sure? yes | no

Yann Guidon / YGDES wrote 11/21/2016 at 05:10 point

Now, I have seen a few instances where no diode was used. It's curious.

If the 1N4148 are too cumbersome, I might as well use a transistor or even a couple of resistors ?...

Having 2 caps in series is OK but more parts would be a bit overkill. I'll have to wire "quite a few" so I'm mindful :-D

  Are you sure? yes | no

Yann Guidon / YGDES wrote 11/21/2016 at 06:13 point

Test result : yes the diodes are necessary to reduce the backwards polarity... But a backwards voltage remains, like 0.6V in half the cases.

A different voltage rail is simpler and safer to handle.

  Are you sure? yes | no