Switches, relays and long-term reliability

[jdbakker]

Hi all,

Quick question: which have given you more maintenance headaches, rotary switches or (latching) relays ?

Context: I am DIYing a portable multi-track recording solution (stay tuned for schematics et al!). Right now I'm trying to decide whether to handle the mike input gain setting with rotary switches or relays. I have found a bunch of brand-name latching relays (so I won't have to worry about power consumption). I would prefer to use these relays, but I am worried about reliability issues. My earlier work experiences have made me detest all electromechanical parts, especially in devices which experience as much vibration as this portable gizmo undoubtedly will see (as I'll have to use my bicycle to get it to most local recording venues). I have trouble extrapolating data sheet figures to real-life reliability, hence this post.

Any horror stories to share ?

JDB.
[naturally, the vibration is only transport-related -- I wouldn't use latching relays in a system that's subject to heavy shocks while in use !]

 

[Rochey]

JD - you have a lot of experience with digital stuff... why not use a digital control analog mic pre?

I hate to come accross as the walking talking sales guy here... (after my ADC thread...). But devices like the PGA2500 are designed for the kinds of application your designing.

Let me know if you're interested, and I'll see if I can hunt down an EVM.

thanks again

Rochey

 

[jdbakker]

[quote author="mediatechnology"]Secondly, make sure that the relay can operate reliably at very low currents.[/quote]
I've read about that, but never actually experienced it. What does it sound like ? My relays (Aromat/NAIS TK1-L2-4.5V) have a specced minimum current of 10uA. I plan to use a Green/Cohen topology without the huge DC-blocking cap, which means that with a 100R gain resistor 1mV of transistor offset is sufficient to get that 10uA current flowing anyway.
[quote author="mediatechnology"]I've been having pretty good luck with the NKK rotaries for mic preamp gain. Mouser has them around $12 each for a 12 position.[/quote]
Thanks, but my mic pre has two adjustable gain stages, which would either mean a two-deck switch (large) or two separate rotary switches (large and not pretty). I still have this fantasy that I can fit an 8-track recorder inside 25x35x10cm for maximum portability, so size does matter.
[quote author="mediatechnology"]You can hear it zip a little during rotation at very, very high gain but not objectionable.[/quote]
I don't really care about silent switching. 9 out of 10 times I'll be recording myself, and there'll be noone to ride the gain. On my current setup I use an infinite hold peak meter during the sound check, adjust gain to match and just hit record. I don't expect this to change.

A bit more context: the mic pre is directly hooked to an ADC. This means that the gain range is relatively limited: on the lower end it's determined by not clipping a 2Vpeak mike signal (and *almost* driving the full range of the ADC), on the hi-gain end it makes little sense to amplify the mike's resistive noise much farther than, say, 20dB over the converter's noise floor. For a modern 120dB ADC this translates to a gain range of 0-40dB. So a dozen 3dB steps should be OK.

JDB
[and I'm working on a DC-coupled discrete version, too ;-)]

 

[jdbakker]

[quote author="Rochey"]JD - you have a lot of experience with digital stuff... why not use a digital control analog mic pre?[/quote]
See, that's what I'm trying to use the relays for ;-)
[quote author="Rochey"]I hate to come accross as the walking talking sales guy here... (after my ADC thread...). But devices like the PGA2500 are designed for the kinds of application your designing.[/quote]
True. I had discounted the '2500 for battery-powered ops since it's pretty power hungry, but most of my proto amps have ended up consuming similar amounts of power. Low noise, headroom, low power: pick any two...

I have been looking at the PGA231x devices recently, but I found that in a small box shared with lots o' digital bits (for the recording) you really want to keep the analog path fully differential from XLR to ADC.
[quote author="Rochey"]Let me know if you're interested, and I'll see if I can hunt down an EVM.[/quote]
That would be very useful. I actually have a few '2500s sitting on my desk, but Work has put the board design I submitted on indefinite hold. Come to think of it, an EVM might help in showing them the error of their ways, too.

Thanks,

JDB
[quite interested in that PCM4222 you mentioned]