Tube Arcing

falconcry72

Arced my first power tube today and scared the crap outta me. I've had a couple tube preamps over the years, but I got my first tube power amp (an ARC VS-115... frickin' gorgeous) about a month ago. Well, I wanted to test the limits with my maggies... and I found the limit. These speakers have such low sensitivity...

Anyway, I checked the bias on all tubes and sure enough one of them measured 0.00mv. Audio Research hand marks all their tubes with exact performance specs, so I'll be able to get a perfect match, but how common is it for power tubes to arc? Is cranking the volume the only cause or can other factors lead to failed tubes?

Might have to choose between the amp and the speakers. This amp can drive "normal" sensitivity speakers to levels as loud as I would ever want with no problem. The LSiM-705's, for example, are no problem at all. These maggies are "rated" at 87db, but reviewers have measured them in room as low as 82. That's low. The VS-115 can get them up to about 90db sustained with peaks in the low 90s, but that's it. Try mid-90's sustained and **** gets weird.

falconcry72

DSkip wrote: »

The impedance is a bigger issue with tube amps. Higher current amps handle impedance drops better. Since tube amps are biased towards high voltage/lower current, they simply do not do as well as SS amps on hard to drive speakers. Maggies are hard to drive speakers. Perhaps you should consider looking at SS for those pups?

The 3.7's are ruler flat at 4ohms (small peak to 8ohms, but they never dip below 4). The VS-115 has an output tap for 4ohms, and it measures very low distortion into 4ohm loads from that tap, even at its full rated output (less than 0.5%) by independent reviewers. I don't think impedance was the culprit in this particular case, given the specs of the speakers and the amp; I think I was just driving too hard and it ran out of steam.

As far as looking into SS amps, yes I have owned many SS amps paired with maggies (Adcom, Yamaha, Rotel, Parasound, Bryston, Krell, BAT...). The VS-115 replaced a BAT VK-500 that did quite well. I still have the BAT, but I've promised it to my dad, and I'm ready for a change anyway. Up until a month ago when I got the VS-115, the speakers were the one part of my system that I thought was "done". Now that I've heard what the VS-115 is capable of, I am re-thinking that stance. Now the question is: do I find the perfect amp for my speakers, or do I find the perfect speakers for my amp? I'm leaning towards finding a sweet pair of monitors for this amp (micro utopias, harbeths, etc), and adding a pair of rythmiks. Or maybe stick with the maggies and find a Pass X250.5. decisions decisions...

tonyb

You crossed the beams....never cross the beams.

falconcry72

tonyb wrote: »

You crossed the beams....never cross the beams.

Definitely did. Do tubes ever arc due to karma? I could see that being an issue.

mhardy6647

The only thing that makes Magneplanars hard to drive is low sensitivity; they're essentially purely resistive loads. Any transformer-coupled vacuum tube amp should be happy enough driving a Magneplanar as long as the ampifier is powerful enough -- most of them, frankly, aren't. Here's one that should be OK:
http://www.manley.com/neo500.php

The voltage/current thing that folks here tend to think of really lives mostly in the realm of direct-coupled (usually solid state) amplifiers. I don't want to minimize the importance of appropriate power supply design in vacuum tube amplifiers, but transformer-coupled amplifiers (tube or solid state) use the output iron to ensure that the output device "sees" an appropriate load. That is (probably) part of the reason that McIntosh stuck with transformer (actually autoformer) coupling even in solid state amplifiers for quite a while. Heck, I think they might still make some autoformer coupled amplifiers.

As to an output tube arcing -- well, anything that takes them outside their design limits is asking for trouble. Usually that is due to maladjustment (e.g., of bias, if the bias is adjustable) or component failure causing bias, plate, or cathode voltage or current to be "nonoptimal"), but some amplifier designs run active components right on the ragged edge. I am more used to seeing HV rectifiers arc than output tubes, but the latter certainly happens. I am guessing that the amp in question is running modern power output tubes; I don't think that most, if any, of them are as robustly made as those of five or six decades ago, unfortunately.

falconcry72

Thanks for the input, mhardy. My output tubes are KT-120s. I was literally testing the limits of the amp when it arced... it's because I was asking too much of it. If these maggies really are 82db sensitive, how many watts would it theoretically take to get them to 95db in room with 100db peakes? That's what they were doing.

The amp tests at about 130wpc at 0.5% distortion into a 4 ohm load.

mhardy6647

The formulae are pretty simple. Doubling distance of the microphone or ear from the radiator subtracts 3 dB, and stereo (two speakers/two channels delivering identical power instead of just one) will add 3 dB of SPL. Every desired increase of 3 dB of SPL (historically considered to be "clearly but barely" audible in the industry, believe it or not) will require twice as much power. An increase of 4.8 dB -- call it 5 :- ) -- requires three times the power, an increase of 10 dB (= 1.0 Bel) requires ten times the power, an increase of 20 dB (=2.0 Bel) requires one hundred times the power, etc. It's exponential (i.e., the dB is logarithmic).

Thus:

...if the sensitivity is based on the most typical standard: 82 dB at 1 meter for 2.83 VAC input (which is two watts into four ohms)...

A stereo pair would deliver 85 dB at 1 meter with 2 watts per channel

You want 95 dB. At 1 meter from the speakers, a desired 95 dB SPL would mean that you'd need 10 dB more power, which is 10 times more (20 watts per channel).

If you want 95 dB at 2 meters from the speakers, you'd need 40 watts per channel; at 4 meters from the speaker, you'd need 80 watts per channel, etc.

100 dB peaks would mean you'd need about triple the power points above. A 4.8 dB increase in SPL correlates to about triple the power required. For example, 100 dB SPL (pretty darned loud) at 2 meters from a pair of 82 dB sensitive, four ohm stereo speakers requires in theory 120 watts per channel.

These are of course "theoretical" -- the real world in terms of acoustic environment and the nature of the AC waveform (i.e, the music signal) will also have some effect.

mhardy6647

Occam's Razor and all that, I'm thinkin' a bum output tube.

Emlyn

mhardy6647 wrote: »

Occam's Razor and all that, I'm thinkin' a bum output tube.

Me too. I'd check with the amp manufacturer for advice on what occurred.

mhardy6647

good advice, that.

falconcry72

DSkip wrote: »

Did the sound ever strain? According I that formula and the tubes you use, watts weren't the issue.

I could not hear any distortion, but it was somewhat compressed sounding. It was very loud... about as loud as I would ever drive these speakers with any amp. Also keep in mind the sensitivity rating I quoted is based on one particular review and it may be very different in my room from my listening position.

To give you an idea of the real world difference in sensitivity between the LSiM-705's and the 3.7's, take these numbers into consideration: to get the polks to what I consider a normal listening volume (mid 80dbs) my preamp's volume control is set to 20. To get the 3.7's to that same volume, the preamp volume control has to be set to 40. To get the polks to what i consider "loud" (mid 90dbs with 100+ peaks), the volume control is set at 30. To reach this same loud volume with the maggies, the volume control was at 60, and that's when the tube arced.