protecting an FET

I'm sure Martin has long experience with how valuable an asset electrical isolation is when dealing with audio signals, and even more so when dealing with multiple rooms. The 2n7000 is a 5-ohm (max) part, among the smallest of the common MOSFETs, and although it's a pretty rugged part, capable of switching 500mA, larger more robust FETs are available in many SSR opto-couplers.

For example, the NAiS Aromat AQV252 is a 1.4-ohm (max) 60V opto- relay in a 6-pin mini-DIP package. It's made up from two 0.7-ohm FETs wired back-to-back, with about 150pF of switch capacitance when open with 5V across its terminals (it has less than 3pF of coupling capacitance to the controlling logic circuitry).

NEC's PS710A uses even larger MOSFETs, and has 0.1-ohms of typical ON resistance. Surprisingly, its self-capacitance is only about twice that of the much smaller NAiS part. It's also available in 6-pin mini-DIP through-hole and surface-mount packages.

Both of the above types are "ac" parts, which means they act as switch closures despite the direction of current flow. If one can insure careful control the polarity of the wiring, photo-transistor opto-couplers can be used. These come with up to four switches in a 16-pin package, for as little as 8 cents per switch, qty 1k.

I read in sci.electronics.design that Walter Harley wrote (in ) about 'protecting an FET', on Fri, 30 Sep 2005:

I'd make protection quite certain by putting a miniature relay at the sending end. May not be the cheapest but you will need to test any other solution to see IF it resists surges and ESD.

Interesting that you guys are suggesting that approach. I'm doing exactly that for some other closures in the same station, relay closures in one room that end up shorting across the 'play' button contacts of a minidisk player in another room. In that case it's the MD that needs to be protected, and full isolation seems desirable; the MD's designers never intended its internals to be exposed in this way.

I'd considered using the photomos couplers to handle the mixer output protection as well, and I've ordered enough parts to do it that way (see Win's post about 5 becoming 25...). The only real downside is that it means I need a power supply somewhere - one more piece of equipment to be mounted, to be cabled and documented, and to potentially fail.

But it seemed kind of like overkill and I wondered whether a passive circuit would do just as well, for the mixer control outs, which are after all intended to be exposed to the outside world. Maybe not - Martin's observations about the importance of the mixer not failing are exactly right. And it's nice to use the same solution everywhere possible; avoids confusion.

Thanks!

-walter

Half-way there but lacking creativity... And what is the necessity of including the whole original post in your reply? This Harley whiner was so worried about consuming too much disk space off Google- how fortunate for Google they have the non-entity worrying about their resources.

I don't know, actually. The front-panel switch is a momentary closure to ground, so we just wire in parallel with it. I've never bothered to trace the circuitry inside the MD. If I had to guess, given that it's consumer electronics, I'd guess a microcontroller port pin with a pull-up resistor to Vcc.

Not at this location. (At my home, it's another story...) There are some light-industrial businesses on the same block, though, so I'd assume there are inductive spikes on the AC lines. And the punch blocks are all mounted on the studio wall, completely exposed to whomever leans up against them after shuffling across the carpet - there's some plan to enclose them but it depends on volunteer carpentry labor. The joy of a community station.

They've not had any problems so far; this is just belt-and-suspenders stuff, provoked by the fact that for the first time they're running control wiring not just within a studio, but between studios.

Thanks, -walter

How about:

BIG diode PTC !!-----!

snip moan

Uncreative definately, but it is solid, bulletproof.

It's quite normal for broadcast stuff to be in operation for years without powering down. You want it as reliable as possible. One radio station I worked never had a breakdown that required a power down in it's MCR console for 5 years (in my time there), and it's bypass system was a box of relays. 1 hour downtime could mean the loss of say

Ok, its not as dangerous as Joergs' field of medical devices, but there is an awful amount of money involved. And dead air is highly embarasing

I could have suggested that Walter use a nasty Pic and a RS422 port, programmed by the janitor's son, to control the remote device. Creative but not advisable.

But the Chief Eng at Walter's place should know this. If he didnt, he has no right to be called a Chief Engineer, (sorry Walter)

martin

"Walter Harley"

Let me reinsert that "moan" :

And what is the necessity of including the whole original post in your reply? This Harley whiner was so worried about consuming too much disk space off Google- how fortunate for Google they have the non-entity worrying about their resources.

Snip it again and it comes back 50x over- How's that?

Nah- not at all.

This is what optocouplers are all about-all the channels are isolated- shorting to other circuits has no effect whatsoever. And if it's a layout problem where you are unable to place a receiver at the destination then no method eliminates the possibility of blowing that circuit out.

On Fri, 30 Sep 2005 22:14:45 -0700, Walter Harley wrote: ...

no

Lose the varistor. Their knee is very "soft", and they wear out. Use Transzorbs:

You should be able to get them almost anywhere. They're faster, with a sharper knee, and last forever. I don't get any kickbacks, I'm just a satisfied customer. :-)

At first, I ass-u-me-d that you were talking about protecting the _input_ side of the FET, in which case I'd have suggested about 1K in series with the input line, and 1N4148s or 1N914s to the rails.

Cheers! Rich

...

On Sat, 1 Oct 2005 09:04:44 -0700, in sci.electronics.design "Walter Harley" wrote: snip

Whats the i/p circuit of the MD? If it's 5V logic stuff you might be able to put 2 reverse diodes, across the rail to switch and switch to ground, and a 1K R in series to the outside world, this is reasonable bullet proof. It depends on the switch pull ups etc. If it is a cpu port pin, who knows! Do you have much RF interference around you?

That will save the external supply, and all those hassels, since you already have an isolated cct.

martin

Actually, if it's used correctly, only the tips where the wires were cut off are exposed -- that's less exposure than any other termination I've seen or heard of. The idea is that you press the insulated wire into the slot, and the knife edges cut through the insulation and slightly into the conductor, making an airtight seal and a low-resistance connection. You can put several wires into each slot for multiple connections.

These things are much easier to use than screw terminals and such. Why does no one but the telcos use them? I can see the usual expensive ceramic-based blocks not being in common use, but why not a plastic cousin? It would seem they'd be more reliable (another telco emphasis) and possibly cheaper.

Anyone know what, if any, disadvantages they might have?

John Perry

Sounds good, but subscribe/lurk to some of the broadcast newsgroups,they are much more useful than SED. I'm fortunately out of the commercial side of broadcasting now, and a few years out of date on which gizmo is the best for "your sound".

One thing, keep those DC control punch blocks physically separate from the progam punch blocks, same applies to comms, and well marked, and there are never ever enough parallel strips on the jackfield

enjoy, while you still have the enthusiasm

martin

What about the rest of the people who repliled to this post?

Stop moaning, there are better things to do with life. Get a girl/boy friend, go out for a meal, even a BIGmac would do, take a bus and look at the countryside

have some FUN,

martin