Phantom power crowbars

[NewYorkDave]

A semi-random thought came to me today... I've looked at a number of phantom power supply circuits over the years and crowbars seem to be relatively rare. That seems odd to me; I'd certainly want to protect the mic in the case of regulator failure. Even just a zener with a breakdown voltage a few volts above 48V, across the supply, would be better than nothin' at all...

 

[analag]

Good point. But then if you look at a lot of Phantom power schemes you will notice that it's not even a proper P.S. Mostly a voltage doubler which is half wave and a bunch of other crap surrounding that, then it's off to powering up your esoteric transistor mic.

analag

 

[kvintus]

never thought of that... A zener would probably be a good idea.

What's wrong with a voltage doubler/tripler if it's properly regulated and filtered?

/Anders

 

[Boswell]

What do you want to protect against? The problem with voltage multipliers is usually getting enough volts at a low enough impedance, so any failure is going to result in a reduction of output volts. There isn't any way they can produce a higher voltage by mistake, unless you put 240V on the primary of a transformer strapped for 110V, but then you would have other problems.

If, however, you're talking about a phantom power rail being derived by regulation from the HT line of a valve circuit, then, I agree, the designer should incorporate a crowbar or independent voltage limiter to protect in the case of regulator failure.

Bear in mind that most capacitor microphones will withstand considerable overvoltage of their phantom power supply without sustaining damage.

 

[Svart]

Use a TVS and be done with it. Most crowbars won't catch anything faster than a few ms. I've tried this for other types of circuits and a crowbar will work ok for slow, large spikes but anything into the ps range will need a TVS.

I can give some examples of circuits if you want.

 

[mobyd]

[quote author="Boswell"]What do you want to protect against? The problem with voltage multipliers is usually getting enough volts at a low enough impedance, so any failure is going to result in a reduction of output volts. There isn't any way they can produce a higher voltage by mistake, unless you put 240V on the primary of a transformer strapped for 110V, but then you would have other problems.
[/quote]

Funny you should mention that - just spent a considerable amount of today replacing a metric buttload of surface mount NE5532s in a small soundcraft spirit mixer that went consideranly overvoltage on ALL RAILS when the reference voltage to the 3842 PWM driver to the SMPSU chopper FET went haywire (before blowing the chip and FET to bits). Oddly, the TLO72s & 4s all survived.
M

 

[Freq Band]

[quote author="mobyd"]
......replacing a metric buttload of .........
M[/quote]

.....I could not find that measurement in my Pocket Reference Guide.


=FB=

 

[Boswell]

Freq Band wrote:

mobyd wrote:

......replacing a metric buttload of .........
M

.....I could not find that measurement in my Pocket Reference Guide.

http://home.earthlink.net/~terrafied/work/mbload.html

 

[Freq Band]

[quote author="Boswell"]

Freq Band wrote:

mobyd wrote:

......replacing a metric buttload of .........
M

.....I could not find that measurement in my Pocket Reference Guide.

http://home.earthlink.net/~terrafied/work/mbload.html[/quote]
...ok, I penciled that in, inside the back cover, next to my other notes : "gazillion,zillion, and bijillion".

 

[Viitalahde]

Say, how would a relay work here? Use a relay with a larger coil rating (a 24v relay for a 15v rail), possibly put a resistor in series with the coil to make sure the relay acts only in real overvoltage situation.

A 24v relay will for sure flip at 18v. When flipped, it'd short the regulator to ground and blow the fuse..?

For phantom power.. I quess you could use a voltage divider or something.

 

[Svart]

it works too slowly for most protection. You need something fast.

I usually use a TVS and thermal fuse

I've also use a true crowbar, A TVS, a zener, SCR and Thermal fuse.

What happens is the TVS catches fast transients that the thermal(pico) fuse and SCR won't react to >picosec. If the fault is slow and prolonged enough the thermal fuse blows. If the problem is overvoltage the zener conducts to the gate of the SCR and the SCR pulls the rail to ground tripping the thermal fuse until the fault clears and the system reset.

I don't know how this would work on signal lines since the signal would have to go through the fuse.

 

[Viitalahde]

Hey Svart, thanks for the tip on TVS's. :thumb: There seems to be a nice selection of this stuff available.

I found this tutorial to TVS's: http://www.microsemi.com/micnotes/100ser/103.htm

All clear to me. But when used for a negative rail, I suppose you turn the diode upside down like you need to do for normal diodes?

I also never heard of crowbar circuits before.. Now I know. http://www.kbt-dc-supplies.com/crowbar.php

 

[Brian Roth]

One of my "pet peeves" about the power supplies in many/most commercial recording desks...no "OVP" (Over Volt Protection).

The worst example I can remember was an A-H desk when the nominally +/- 18 V audio rails began spitting out +/- 30 VDC "raw" (from the rectifier) DC and smoke literally poured out of 32 I/O modules.

Bad design to begin with, having that high of "raw" rails, but having no OVP was totally stupid. I spent a few weeks replacing chips at a cost of many thousands of dollars.

FWIW, the "open frame" linear modules (Power-One, Internationa Power, Condor, etc) all have optional OVP protection for a few bucks extra.

Bri