Light bulb attenuator/compressor

[midwayfair]

I'm trying to figure out a way to make sure that the amp always sees a load in case a bulb blows.

So the only way I could think of to keep a resistor in the circuit all the time is to put it in parallel with the bulb ...

red speaker wire > 20W (?) lightbulb in parallel with a 20-50W 8ohm resistor > speaker
black wire as normal.

But this doesn't work. The bulb's resistance goes up as the wattage increases, which means the load drops. That's the opposite of what I want.

Any ideas?

I'm specifically interested in building a lightbulb compressor, not any other type of attenuator.

[tca]

I'm specifically interested in building a lightbulb compressor, not any other type of attenuator.

Something like a Wien bridge (with other resistor values):



I wonder how much compression you can get out of it.

Edit: Actually is more like an expander than a compressor, much like you describe, but it has the advantage that if any of the lamps fail, nothing "bad" will happen. To get a compressor you should put the lamp in series with the speaker.

[artifus]

http://sound.westhost.com/project45.htm

[PRR]

> a load in case a bulb blows

OR gets hot?

IIRC, a 120V 60W lamp is 16 ohms cold but over 50 ohms at a dull glow (and obviously 240 ohms at full 120V bright-white).

So the amp "loses its load" any time that the lamp is warm.

I think you want something like this:



OR: drive the lamp and dummy-load with an LM386, then run that signal into a real amplifier. Because a lamp-limiter in the speaker side *must* absorb *most* of the power that you dragged to the gig. If you need 10dB of limiting (not a lot), then a 50 Watt amp puts 5 watts in the speaker.

[midwayfair]

> a load in case a bulb blows

OR gets hot?

IIRC, a 120V 60W lamp is 16 ohms cold but over 50 ohms at a dull glow (and obviously 240 ohms at full 120V bright-white).

So the amp "loses its load" any time that the lamp is warm.

I think you want something like this:

THIS looks more like what I'm after.

It occurs to me that I could borrow a friend's Night Light and see how it does the compression setting. I think that has a constant load regardless of the use of the bulb.

Edit: is that a 4-ohm speaker? Do I make any changes if it's an 8 ohm?

[chptunes]

Hey folks.. ran across this old thread while searching for "light bulb compressor" schematics.  Has anyone achieved good results from PRR's circuit, or similar 'light bulb' arrangements?

Thanks, -Corey

[brett]

Hi
off the top of my head, I can't see a reason why this wouldn't work in a stompbox.
The 'driver amp' might be an LM386 and the bulb from a 6V flashlight. Replace the speaker with a 'sensing' load (a JFET source follower or BJT emitter follower). Or even another 386 for a compressed stompbox micro-amp.
Surely this has been done before?
cheers

[pinkjimiphoton]

hey jon,

if memory serves, you can wire 4 60 watt 120v lightbulbs in series/parallel and get a reactive load that will handle 200 watts or so. each lightbulb is a reactive load, and close enough to 8r for rocknroll.

i asked kevin o'connor (TUT) about doing this about 10 years ago, and he said it would be fine and not hurt my amp.

never actually tried it tho... just fwiw

[midwayfair]

hey jon,

if memory serves, you can wire 4 60 watt 120v lightbulbs in series/parallel and get a reactive load that will handle 200 watts or so. each lightbulb is a reactive load, and close enough to 8r for rocknroll.

i asked kevin o'connor (TUT) about doing this about 10 years ago, and he said it would be fine and not hurt my amp.

never actually tried it tho... just fwiw

The issue was ensuring that there can never be no load if the bulb blows. Using four bulbs might make it less likely they'd all fail, but not necessarily fool-proof in the way that a resistor is. I'm still not sure if PRR's schem means that if the bulb blows there's still a load. I don't know how speakers work. I scrapped this idea because I'm too afraid of damaging my amp and just haven't had time to learn what I need to know to ensure it's not damaged. I'd RE my friend's Night Light to build myself, but it's not like Swart overcharges for the thing (it's cheaper than many other attenuators).

[slacker]

In PRR's scheme the bulb is in series with the speaker and they are both in parallel with the 8 Ohm resistor. If the bulb blows it breaks the connection to the speaker but the amp still has the 8 Ohm resistor as a load.

[PRR]

> if memory serves, you can wire 4 60 watt 120v lightbulbs in series/parallel and get a reactive load that will handle 200 watts or so.

Don't try that.

One 120V 60W lamp is 16 ohms cold. Great! I wired it to a 10 Watt tube amp, cranked, and got WAY too much voltage, and a dim glow. Turns out this much power drives the lamp to about 100 Ohms, the amp is nearly un-loaded.

Four series-parallel just gives you the same deal at 40 Watts. I wouldn't even expect that to be a "good load" at the 10W level.

BTW: the lamp is not significantly reactive, and we don't want reactive loads for basic amp *testing*. Yes, speakers have complex reactances; melding test-resistor results to loudspeaker results is yet another layer of fun.

The lamp *attenuator* plan I posted will not leave the amp un-loaded (I know better).

No, it is not generally useful except E-V used it as tweeter protection. The lamp heat/cool times do little for music. But Jon wanted to try it, was not interested in any other type compressor, so WTF.

[pinkjimiphoton]

well, as we both know, i'm far from an ee paul!!! lol

but... unloaded would be an open circuit i would think, not 100r!  amps can usually handle much higher resistances better than too low...as i recall from my more adventurous days (read: willing to blow up high voltage tube stuff ) an 8r speaker goes ALL OVER THE PLACE as a "resistance" (i don't fully understand the concept of impedance, other than it changes at different frequencies)... i mean, from 40-50r to zero ohms at some point... zero ohms, it's no longer a resistance, it's a conductor. while you (or me)'d think that would be an issue (and it would be if it stayed there) but apparently it's not cuz it's intermittent.

when i first got the idea of using light bulbs in place of speakers, i contacted kevin o'connor (author, The Ultimate Tone and London Sound amp guru) about it and asked him if indeed i could hurt anything trying it. he said it would be fine, the amp doesn't care if it's a speaker or a lightbulb, as long as it has a load on it. i asked him about the series/parallel thing, and he said that would be fine.... not to argue the point, cuz we both (all) know i don't know $#I7 about stuff, but if ya think about it, just reading the resistance of a light bulb isn't enough. like a speaker, to find the impedance you would need to input a 1k sine wave. at other frequencies, an 8r speaker could be anywhere from 4. some odd ohms to 12, and still be an 8r load.

speaker line level is up to 70v or so, so a 60-100 watt lightbulb that can handle 120 + volts shouldn't burn up.

the 60watt bulbs i'd measured were reading about 8r, but i'd imagine it could vary by brand and size etc.

again, not to argue, but just to learn more is why i'm bringing this up. i've seen tube amps that were designed just to drive relays, which are definitely not like speakers!! (i've taken them and hooked up speakers to the same output transformers that were for the relays and it worked fine, indeed it's currently one of them as the OT in my tweaked out princeton)

i find it interesting to see two completely different viewpoints from two such respected techs as you and kevin.

anyways. 

[midwayfair]

But Jon wanted to try it, was not interested in any other type compressor, so WTF.

Aw, I hardly think 'not interested in any other compressor' described me. There are very few compressors I'm not interested in. This was just a type I hadn't built yet, and you already co vinced me that it wasn't a great method. 

[pinkjimiphoton]

 

[PRR]

> Aw, I hardly think 'not interested in any other compressor' described me.

I know you have wide-ranging interests.

However in *this* thread someone put these words in your post:

I'm specifically interested in building a lightbulb compressor, not any other type of attenuator.

If you let the thread expand to cover "all" compressors, I can point you to about 4 decades of bedtime reading.

Many grizzled designers think that Power Amps are the toughest job in audio design (mostly the blow-up proofing issues). I'm starting to think dynamic range processors are a strong #2. They don't blow up (usually), but they don't just "pass signal", they have to jack-around the signal yet slip under the listener's radar (sonar?), and never slip-up (burp, fade, swell noticably, dropout, garble, bobble, fuzz....).

So in that sense I have STRONG interest in SIMPLE compressors that are 90% good with <10% parts of a "proper compressor". Selected LDRs are a low-risk easy path.

[PRR]

You are right that 100 ohms is not the same as infinity.

Kevin may have blown-up more amps (or seen more blown-up amps) than I have. One blow-up mode is when there is not enough resistance to damp the kick-back of an OVER-driven tube amp's output transformer. It doesn't have to be Infinite ohms to puncture the winding insulation. It *may* have to be more than 10X the design load. And from 16 to 100 is less-than 10X. But all these safety margins are very approximate. And the OT is a major expense. I prefer not to get even close.

Lamp impedance versus frequency.... I have tried this, so have many others. Small lamps are used as test-loads for small radio transmitters, At 2MHz or 5MHz they become a little inductive. Over the audio band they are dead-flat. For another evidence: common 240V 3W lamps are used to stabilize audio oscillators over the 20Hz to 200KHz bands. If there were any deviation of impedance, the output level would change.... it doesn't.

Speaker impedance should NEVER be "zero". Speakers without crossovers will never dip below their DC resistance. See speaker impedance curves. Here is one, bottom of the page. DC resistance Re (incorrectly reported as impedance) is 6 ohms. Audio impedance is above 8 ohms at 45hz, peaks at 50 ohms at 88Hz, drops to 7 ohms 200-300Hz, rises to 8 ohms at 800Hz, and 20 ohms at 5KHz.

Tubes used to drive relays are not normally transformer-coupled, nor fed raw audio.

70V speaker lines are a somewhat different situation.

We are not so much worried about the lamp Burning-Up, as it not being the expected value. All physical resistances change with temperature-- that's the way atoms are. if you heat copper about 100 degrees its resistance doubles (and that matters in some high-power loudspeaker work). Resistor-metal is picked for a lower resistance/temperature change, but even so if a resistor heats 100 degrees we usually say "Wow! Hot!" and go no further. However a lamp filament normally, safely, heats a thousand degrees. 10X more than a hot resistor. And Tungsten has a pretty high resistance/temp rise (which is key to light-bulb stability). So even 100 degrees in a lamp, far-far below glow, is a Large change of resistance. And a poor test-load.

Use a fry-pan. You won't easily get it wow-hot, and you can feel the heat directly instead of through a vacuum.

[pinkjimiphoton]

thanks for not tearing me up on this brother. you make all valid points.

but one question.... when i was playing with stuff a long time ago, i had a fluke i used to read the resistance of the celestion in my princeton... and it was definitely hitting 0 ohms at times while i was playing... i just had it hooked up and was watching while i was playing (silly hippy is dangerous to be around i think). i watched it go up and down.... if it's hitting 0 ohms (or something so low as to be close enough for the meter to think it was zero) does that mean the OT is passing dc? if it is, i'd assume it's shot.... or was i having one of them big hallucinations again?

[brett]

Hi
there is enough DC resistance, inductance(s) and capacitance(s) in most '8 ohm' speakers to allow low AC resistance (a couple of ohms). (PJP: the OT is passing AC, not DC. Transformers cannot pass significant amounts of DC (and near-DC)).

I've not heard that going 'open' or a bulb going high resistance is much of a problem for an amp's output section.

Re: LDRs
If only LDRs were better behaved, they'd be great. But they have slow turn-on and slow turn-off (at different rates, so it doesn't help to just increase the compressor's sensitivity to the slew-rate).

I wonder what the rate of resistance change is like for the electricity to heat conversion in the lamp? Faster or slower than an LDR? More or less symmetric? More or less 'musical'?
cheers
 

[gritz]

Hi
there is enough DC resistance, inductance(s) and capacitance(s) in most '8 ohm' speakers to allow low AC resistance (a couple of ohms). (PJP: the OT is passing AC, not DC. Transformers cannot pass significant amounts of DC (and near-DC)).

I've not heard that going 'open' or a bulb going high resistance is much of a problem for an amp's output section.

Re: LDRs
If only LDRs were better behaved, they'd be great. But they have slow turn-on and slow turn-off (at different rates, so it doesn't help to just increase the compressor's sensitivity to the slew-rate).

I wonder what the rate of resistance change is like for the electricity to heat conversion in the lamp? Faster or slower than an LDR? More or less symmetric? More or less 'musical'?
cheers
 

From memory the rise time of a "typical" automotive indicator filament lamp is in the order of 100 - 200 milliseconds. No idea about decay times, but I assume it's all related to filament mass v. surface area, with skinny low current filaments quenching quicker. An internet search may yield manufacturer data for lamps used in applications where this may matter.

Regarding slow LDRs: Yeah, cadmium sulphide ORP12s and the like are pretty glacial, but vactrols designed for higher speed applications use different photosensitive compounds (cadmium selenide?) I use these:

www.silonex.com/datasheets/specs/images/pdf/104058.pdf‎

As with all vactrols the actual attack and recovery times depend on the amount of current you hit them with (and for how long), but attack times of a few ms in combination with a recovery time of a couple tens of ms are vaguely ballpark. More current = faster attack + slower decay and vice versa. As a bonus just 1mA liberates a resistance of about 300R. I like 'em!

If any newbies are reading this thread I'd like to point out that running a valve amp without a vaguely suitable load across the output transformer secondary is only marginally less silly than poking your genitals into a bee's nest. It's a bit like closing your eyes while driving at 70mph and counting to ten before opening them again. It may not end in tears, but no good will come of it.

I don't know if there's such a thing as a "musical lamp". Perhaps the singing arc would be a better choice.

[pinkjimiphoton]

 

that was sig-worthy.


i know the OT is ac on the secondary, but what i'm talking about is the speaker kinda going thru zero for lack of a better explanation.
i remember the meter being set on low scale r,  and watching it go up, and down, and occaisionally hitting zero with a "-" sign displayed.
it is entirely possible i was misunderstanding it, but i took it as the speaker varying from an inductor (8r-ish) to a conductor (0r-ish).

was i misreading this? i mean, we're talking 14-15 years ago, and i knew even less than the picayune things i've learned now...

if it's conducting, even for a few brief milliseconds...slow enough to be noticeable on a digital meter.... wouldn't that mean the speaker is putting out dc at that point? wondering if that's what makes some speakers blow, by pushing them past their point of induction and popping the voice coil.
i mean, guitar signals thru an o-scope, even if you don't know what you're doing like me is pretty cool to watch.... and way more complicated than a 1k sine wave.

oy... my head hurts at this point

anyways, generally the only time i've seen dc on an output transdormer other than that experiment was if the primary was shorted to it or arcing. usually that sounds like a blown speaker, right around F# above and below middle c. you hit those notes, or CLOSE to those notes, and suddenly the amp craps out. lay off the note, it comes back on. (i hate it when that happens. )