thermionic said:
Putting a relay in the signal path is not my idea of an ultra-reliable design, though
How often do you see them fail?
I've seen many failed relays. Out of my head, Otari tape recorders, Soundcraft broadcast mixers.
I can think of items that are still going strong after 20+ yrs which contain tens of relays. I've owned cars with tens of relays in them; everything else failed, but I don't recall having to change a relay.
I had...it was the horn relay, not a big deal, I'm not a honker.
What is the usual failure mode for relays? Does this apply to small-signal relays which are sealed, supposedly filled with inert gas?
Now we are getting closer to the issue. The answer is, it depends. Power relay contacts arc and eventually fail to make contact, but before that, the residues may create leaks. Low-signal relays usually end up with contact resistance increasing and becoming intermittent or in need of a more or less gentle tap. They also generally are combined with low-power coils and after years of use, the mechanical wear makes it more difficult to attract the armature correctly, losing contact pressure. But I would say the worst situation is medium power application that sit between the small-signal relays with super-duper plating and protection, and the power relays that rely on arcing for contact reliance. In these in-between applications, all types of failures can happen.
There are loads of mastering items with relays in them. Do these fail regularly?
Probably not regularly. And the consequences of failure may not be catastrophic.
But I know that if I have to choose between a switch or a relay triggered by a switch, I prefer the former in terms of reliability.
edit - a quick Internet search suggests relays are regarded as being on-a-par with physical switches in terms of quoted operations. Does that mean that, in order to obtain 'reliable' status we all have to use solid-state switches, triggered by touch-screen (yuk) interfacing?
The standard industrial evaluation method for reliability is MTBF.
Relay manufacturers have a variety of approach regarding this issue.
On the one hand, Omron declare they won't publish MTBF because it depends on too many factors that are not under their control, such as, current, voltage, load, temperature,...
On the other hand, Teledyne go to great lengths producing a formula where all these factors are taken into account. One can see that they attribute 30% more life expectancy for general purpose medium power relays over low-power types and 20% less for sensitive.
Indeed the factor goes from 1 to 100 depending on construction type.
The number of contacts also influences MTBF; the more complex the contact arrangement, the more risks of failure.
Operating the relays at 45°C (113°F) ambient instead of 25°C (77°F) reduces MTBF by about 20%.
And obviouly, the number of operations also changes MTBF.
This is very important in commercial products, where the failure of one relay generates frustration and aggravation, in addition to service costs.
If you can troubleshoot and conclude it's a relay and you know where it is and have one ready, it's a minor inconvenience. If you don't check all these boxes and have to hire a service guy, or worse, ship the unit to a service center, it's a proper sanctified PITN.
Going back to the specific case presented by the OP, I wince over the fact that he has put two electromechanical components where one would be enough...