School me on relay latching, good articles welcome "how to make a relay latch"

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
latching relays that use magnets to hold the contacts after energization ( is that a word?  :eek:  )  can sometimes be defeated by a sharp blow to the relay case,

so if you have a really important situation that you want to monitor, like a nuclear reactor meltdown shut off switch, you might want to use a latching relay that needs to be reset mechanically, so that it does not revert to the untripped state by itself, either by vibration  or a voltage being applied when you do not want it to be applied,

other types if systems have relays that revert to a tripped state if power is lost, so that if a tsunami knocks out power to your reactor, the relays revert to a tripped state automatically, this is known as a fail-safe system,

what are you using these relays for?  could you use solid state instead? there are circuits that can mimic latching relays, like the flip flop,

if you are going to have the relay go through a lot of cycles, then you may want to use a longer lasting  solution (solid state) as mechanical relays are only good for so many energizations,

why yes it is a word,

energization
English
Noun

energization (plural energizations)

    The act (or state) of being energized




 
Good thing about English is that there is no official authority to call what is a word or not, so if it's not, you call it a word and it is from now on, depends on people start to using it or not. In Spanish we have an oficial authority, nobody gives a s**t about, of course, so people start to use a new word and some time after they have to call it a word or they will be regulating a different language.

About those mechanical reset relays, some old meters had as protection something like that, I think there was no other way around to properly reset those circuits since the circuit can't know when the protection is acting if there is a problem or not, since the protection leaves no signal through, with DACs is another story and over range is not as destructive as it was with most mechanical needles, so here we are, already forgot about those protection circuits...

JS
 
great discourse.

That word has a strange hidden implication to it. Read a philosophy book and you'll find more made up words than you want. Big words often have big explanations.

too late to switch to SS relays. I'be gone too far with messing with these panosonic latching relay. Great to know they work for us though!

After bread boarding  a regulator circuit they seem to be non latching even though they are listed as latching on mouser. Maybe I have the definitions criss crossed? I have a link to these on page one of this thread. I'm on a phone or I would link you. you guys are right about them not being picky. They are 24VDC rated but as I watch they caps discharge they release at about 6VDC (and when simply disconnected). Regardless I understand how to make these work now so viola.

In another thread kingston mentioned that they are best used with regulator circuits. I'm assuming thats to keep the voltage constant with the current fluctuations.

My relays will be used for phase phantom pad and DI.
 
buildafriend said:
great discourse.

That word has a strange hidden implication to it. Read a philosophy book and you'll find more made up words than you want. Big words often have big explanations.

too late to switch to SS relays. I'be gone too far with messing with these panosonic latching relay. Great to know they work for us though!

After bread boarding  a regulator circuit they seem to be non latching even though they are listed as latching on mouser. Maybe I have the definitions criss crossed? I have a link to these on page one of this thread. I'm on a phone or I would link you. you guys are right about them not being picky. They are 24VDC rated but as I watch they caps discharge they release at about 6VDC (and when simply disconnected). Regardless I understand how to make these work now so viola.

http://groupdiy.com/index.php?topic=58396.msg743034#msg743034

like I said a week ago... Mouser has mislabeled them...  per the data sheet they are non-latching

In another thread kingston mentioned that they are best used with regulator circuits. I'm assuming thats to keep the voltage constant with the current fluctuations.

My relays will be used for phase phantom pad and DI.

I don't know about using solid state relays for low level audio signals.

Have fun...

JR
 
Is it good practice to keep the transformer you're powering your relays off of isolated or does that only go for the ground? I'm wondering about sharing a transformer for phantom and relay regulators.
 
The most important thing is not to share current path between clean and dirty stuff, audio signals and relays in this case, the clean path would be after the regulator for the audio PS, so the traces going out of the regulator shouldn't touch any relay in it's life, if the relays are powered by the same source (transformer, diode bridge, first smoothing caps) than the regulator the regulator will take most of it's noise out which should be enough to keep noises out of the audio. It's probably overkill having two transformers, and quite problematic (you need to put both of those transformers somewhere) but using a separate rectifier and smoothing caps will keep noises even more out of the equation and is not a big deal in most cases, if you only need an amp or less, having a few amps rectifier plus a one or two mF at 12V or 24V shouldn't be a problem, quite cheap, and if you take all your wiring from there you can be pretty sure the currents won't meet at any point in the circuit. Of course you need to meet both grounds at some point to keep them referenced to each other, unless you are using optocouplers or some insulation to take the control signal from one side to the other if you need to, and even then, you should reference both of them to an "absolute" potential, call it mother earth or the plane chassis.

JS
 
buildafriend said:
Is it good practice to keep the transformer you're powering your relays off of isolated or does that only go for the ground? I'm wondering about sharing a transformer for phantom and relay regulators.

I did not use a lot of relays but for bypasses in studio products I would power the relay off the same power transformer but use a separate diode and capacitor (and separate ground return path). I intentionally used a small capacitor to power the relay so it would drop out quickly after power was removed from the circuit, to bypass the audio before the active circuitry became unregulated and misbehaved. I used a delay in the relay turn on so again the audio path would be settled down before connecting to the outside world..

JR
 
Thanks guys, i would be dead in the water without you.

for experimenting purposes I'm keeping each circuit ground plane isolated but linkable via jumpers.

I'm looking into virtual grounds for the relay section.



 

Attachments

  • separate grounds.PNG
    separate grounds.PNG
    78.5 KB · Views: 96
buildafriend said:
Thanks guys, i would be dead in the water without you.

for experimenting purposes I'm keeping each circuit ground plane isolated but linkable via jumpers.

I'm looking into virtual grounds for the relay section.
I can see your ground planes are ruining any effort at organizing hierarchical ground. Particularly noticeable on the lower left side, where the negative side of the bridge rectifier goes to the reference of the regulator first, then to the output connector and finally at the negative pole of the smoothing cap after a large detour around the board!
The grounded side of the rectifier (negative for positive supply, positive for negative supply) should go to the smoothing cap first, then this "ground trace should go to the regulators reference and then to the output capacitor and output connector, in this order. The ground pour should not be allowed to interfere with this organization.
I don't know what PCB design software you use, but it should not allow you to use these smaller traces from the connector to the bridge (lower right side) and then bigger ones for joining the diodes. Current is maximum at the connector, than splits in two; traces should be sized accordingly. Unfortunately, PCB software often allow just one size per net class.
You should not let all these doglegs, as around D17 & 18.
Finally, there is room for trace length reduction; remember that the shorter the traces, the more copper available for ground. There's never too much ground in audio.
 
thanks for looking so intently abbey road!

top left plane is not even needed. it's helpful for shielding but there is no real ground connection needed there.

everything is linkable, i like this for experimenting. I plan on having a REV 2 so that will based on how many flaws this has in the end. my goal is obviously the least possible. I have bread boarded each of the circuits and they worked well. all but the HT one that is (top right of board) but it should work fine. this whole thing is an isolated PSU section from the main tube preamp board for improved EMR from the transformers. if I'm lucky maybe I'll get enough out of it to daisy chain another unit.

what do you mean dog leg? is that an isolated ground plane area or maybe you mean the bus spokes from the ground sides of component footprints?

This is advanced circtuits PCB artist. Kicad was driving me nuts so I kept with this software which is at least good for etching. I've seen great working boards made by this company. not sure how good or bad the pricing is though.. i've never compared to anything else.
 

Attachments

  • 24V section.PNG
    24V section.PNG
    98.7 KB · Views: 70
I don't want to go on too much a tangent from the thread title. maybe i'll start a new thread for all around PSU PCB design.

again thanks a million for looking and helping to everyone.
 
Try to visualize the current flow,,, especially in ground traces. Between the reservoir cap and  diodes, transformer winding is a nasty hum filled current pulse repeating at 1x or 2x the mains frequency.

The regulator ground pin will pretty much define your PS 0V node. Keep the dirty current on the PS side of that node, and the clean current on the clean side.

I'm sure my comment is an over simplification but stop trying to think as blocks or neighborhoods and visualize the current moving around.

JR
 
thanks!!

The grounded side of the rectifier (negative for positive supply, positive for negative supply) should go to the smoothing cap first, then this "ground trace should go to the regulators reference and then to the output capacitor and output connector, in this order. The ground pour should not be allowed to interfere with this organization.
 
abbey road d enfer said:
buildafriend said:
what do you mean dog leg?
See attachment. You may think I'm picking at trivial details, but it is good practice to minimize length and kinks.

Oh, yeah, depending on process this might get etched away and the trace there will be too thin.

-a
 
Andy Peters said:
abbey road d enfer said:
buildafriend said:
what do you mean dog leg?
See attachment. You may think I'm picking at trivial details, but it is good practice to minimize length and kinks.

Oh, yeah, depending on process this might get etched away and the trace there will be too thin.

-a

I self etch my boards, I used to curve the traces so they were all smooth, since the only problem I had with sharp turns in thin traces was them to lift off the board, now I use 45º angles everywhere unless the traces are really thin (10 mils or less) when I curve them a little, but they are already pretty hard to self etch, I'm starting to experiment with new techniques to make them, but I guess the lifting problem will remain. I usually get away with a 12-12 rule where 45º knees are more than enough, in any case I would avoid a shape like that for this reason and for a more professional look, that looks like someone skipped that detail...

JS
 
I've etched probably around 50 or so boards now.

Things I make sure I do are:
make the traces and gaps a bit bigger than normal
heat the boards up very well for the transfer/toner paper
make sure the blank pcb is very clean and buffed shiny for the toner paper (simple green works well)
use a laser printer and make the printing settings are so it is as thick and dark as possible
don't irritate the PCB in the ferric chloride too much, just a bit if needed
dip it in hot water to separate the PCB from the toner paper and wait until the paper frees it's self
 
Back
Top