Question about increasing current draw..

[Govmnt_Lacky]

I built a Ruby amp with the optional headphone output (via Zobel network a-la GGG) and I noticed something quite disturbing 

I built the Gain trimmer onto the PCB and have the Volume as an adjustable pot. I have the whole circuit tweaked just right.

Yesterday, I was trying to fit a 9V battery into the circuit to make the amp portsable but first I wanted to know how much current the whole thing was pulling down. I threw the whole amp onto a DC power supply that measures current draw. I was shocked to see that the initial draw of the circuit was 75mA.....

Even worse, when I adjusted the volume pot, the current draw went as high as 200mA 

Does this sound right for this circuit OR did I do something BAD?

[PRR]

'386 should pull 14mA no-signal.

Link to schematic and detail your changes.

[Govmnt_Lacky]

'386 should pull 14mA no-signal.

Link to schematic and detail your changes.

Built from GGG schematic and PCB transfer:

http://www.generalguitargadgets.com/pdf/ggg_ruby.pdf?phpMyAdmin=4a28f86a515b7883e7bc35a68d4e7b6d&phpMyAdmin=78482479fd7e7fc3768044a841b3e85a

Built circuit with the 1K Gain as a trimmer and the Volume as a 16mm pot. Included SWITCHABLE headphone jack w/ Zobel network according to GGG layout.

Speaker used was  4 1/2 inch 8ohm.

Like I said above, the amp works great (a bit louder in the headphones but its all OK) however, when I hooked it up to the DC power supply, the current draw went WAY UP when I increased the volume 

P.S. I used ALL of the values and transistors in the GGG layout. The ONLY deviation I have is that I am using a lamp as an on/off indicator INSTEAD of an LED.

Ideas?

[Govmnt_Lacky]

Found an interesting article here to back up what I am saying. Looks like the current draw DOES increase with the volume. 

http://tech-tut.com/?p=1219

Any ideas on how to remedy this so I can get more than 1 or 2 hours out of a 9V battery?

[R.G.]

Like I said above, the amp works great (a bit louder in the headphones but its all OK) however, when I hooked it up to the DC power supply, the current draw went WAY UP when I increased the volume 
P.S. I used ALL of the values and transistors in the GGG layout. The ONLY deviation I have is that I am using a lamp as an on/off indicator INSTEAD of an LED.

Whenever a solid state amp mysteriously pulls a lot of current, suspect oscillation. Put a scope on it and eliminate this first.

What current does the lamp draw?

[PRR]

> I threw the whole amp onto a DC power supply that measures current draw. I was shocked to see that the initial draw of the circuit was 75mA.....

If you used a 60mA lamp (instead of the LED we might assume) then 75mA idle is fine. You should pop out the lamp while researching.

FWIW, the '386 sheet says 4mA idle typical, not 14mA idle as I recall.

> Even worse, when I adjusted the volume pot, the current draw went as high as 200mA 
> Looks like the current draw DOES increase with the volume.

OK, just what are you testing?

With _NO_ signal and/or _NO_ load, the amp consumption should not vary much.

If consumption rises with NO signal and/or load, especially if it shoots up suddenly, as R.G. says it probably DOES "have signal": hypersonic oscillation.

WITH load AND signal, the amp consumption WILL rise. There is no free lunch. Power has to come from somewhere. The '386's idle power, 9V*4mA, 0.036 Watts, is much less than we want in a speaker. So at BIG output power it sucks harder. To put 0.75W into a load takes 1.0W-1.5W from the battery.

As a rule-o-thumb, a good amp with 8 ohm load and full sine power will suck like a 50 ohm resistor. 9V/50= 180mA. 12V/50= 240mA= 0.240A. This agrees with the Tech-Tut article +/- the fact that loud guitar is not a perfect sine (but on average, not too very different).

> get more than 1 or 2 hours out of a 9V battery?

How much "more"?

Do you have data on your 9V batt? Would be helpful.

I'm seeing numbers like 9V 300mAH (at high current, 100mA).

Also that a 9V battery is NOT intended to be sucked at 100mA. They were designed for transformer-coupled pocket radios making about 0.1 Watts max or about 25mA played "LOUD" (about 5mA-10mA played soft).

They are better than they used to be.

Still the older '386 sheets show 16 ohm loads; any lower is a strain on the battery and also causes excess loss inside the '386.

9V 300mAH is 2.7 Watt-Hours of DC. Efficiency (re: Sine) won't be far over 50%. You can have 1.3 Watt-Hours of audio output. This can be 1+ Watts for an hour or 0.1 Watts for 13 hours or 0.02 Watts (tranny radio under the pillow) for 26 hours. Indeed 100mW pocket radios ran 10-20 hours on a battery.

Let's say you can live with less than the whopping 0.7 Watts that a '386 at 9V and 8 ohms will make. How would you get less power? Use Ohm's Law! Lower the voltage. Raise the resistance (load impedance).

Side-note: you don't want WATTS, you want AIR-SHAKE. You can't stir a lake well with a teaspoon, and you can't make heavy sound well with a teeny speaker. It may seem silly, but a good hot Twelve-Inch will make a lot "more" from a part-watt than a Three-inch. This trick alone can give same/better sound three times longer from the same battery. A Full Stack (8 zingy Tens) is even better (but won't fit on the bus).

If we are thinking Less Volts, we REALLY need to question "out of a 9V battery". The 9-Volt is your WORST battery buy (after watch, dog-collar, and PC batts). Yes, it is the most convenient and compact way to hit 9V, but you PAY for that convenience. If load is so light that battery cost does not hurt, OK. But your question implies 9V batt cost DOES hurt. Then the question is not "per 9V batt" but "per battery Dollar".

Here is a hastily-compiled list of battery cost and value:

AAA __________ $0.83/each ____ 1,200mAH ____ 1.8WH ____ $0.46/WH
AA ___________ $0.60/each ____ 2,700mAH ______ 4WH ____ $0.41/WH
C ____________ $1.62/each ____ 8,300mAH _____ 12WH ____ $0.13/WH
D ____________ $1.62/each ___ 21,000mAH _____ 31WH ____ $0.05/WH
9V ___________ $3.50-$1.80 _____ 300mAH_____ 2.7WH ____ $0.70/WH
6V Lantern ____ $4.00 _______ 11,000mAH _____ 66WH ____ $0.06/WH
6V HD Lantern _ $8.30/each __ 26,000mAH ____ 156WH ____ $0.055/WH

Or sorted by bang-per-buck:

D _____________ $1.62/each ___ 21,000mAH _____ 31WH ____ $0.05/WH
6V HD Lantern _ $8.30/each ___ 26,000mAH ____ 156WH ____ $0.055/WH
6V Lantern ____ $4.00/each ___ 11,000mAH _____ 66WH ____ $0.06/WH
C _____________ $1.62/each ____ 8,300mAH _____ 12WH ____ $0.13/WH
AA ____________ $0.60/each ____ 2,700mAH ______ 4WH ____ $0.41/WH
AAA ___________ $0.83/each ____ 1,200mAH ____ 1.8WH ____ $0.46/WH
9V ___________ $3.50-$1.80 ______ 300mAH_____ 2.7WH ____ $0.70/WH

The 9V is by far your worst buy. Even the pathetic (though compact) AAA is 1.5 times better. C-cell is 5 times better. D-cell is fourteen times better.

$/WH is abstract. We really think in dollars per hour.

At 0.5W output, 0.75W drain on battery:
D _________ 31WH ____ $1.62 ____ 41 hours _____ $0.04/hour
6V HDLan _ 156WH ____ $8.30 ___ 208 hours _____ $0.04/hour
6V Lan ____ 66WH ____ $4.00 ___  88 hours _____ $0.045/hour
C _________ 12WH ____ $1.62 ____ 16 hours _____ $0.10/hour
AA _________ 4WH ____ $0.60 ___ 5.3 hours _____ $0.11/hour
AAA ______ 1.8WH ____ $0.83 ___ 2.4 hours _____ $0.34/hour
9V _______ 2.7WH ____ $2.00 ___ 3.6 hours _____ $0.55/hour

For comparison: international calling costs $6-$25/hour, movie theater is $17+/hour, video rental $5/hour, drinking beer on the lawn is $3/hour (your milage may vary). By those standards, $0.55/hour to play guitar in the park is NOT "expensive fun". However $0.04-$0.10/hour is a lot more fun per buck.

You want to use D-cells. Four Ds will make 0.5W in 8 ohms for 164 hours for $6.50.

If that's too big/heavy, look at AA-cells.

9V should be your last resort.

[Govmnt_Lacky]

WITH load AND signal, the amp consumption WILL rise. There is no free lunch. Power has to come from somewhere.

Thanks for the good info Paul 

This statement pretty much sums up what I have been looking for. It appears that the rise in current draw is "to be expected" or parallel with the rise in volume. I have found several reports/articles on the interwebs that confirm this with respects to this particular build.

My only question is "Why are there no threads/reports of the Ruby amp devouring 9V batteries?"

[R.G.]

My only question is "Why are there no threads/reports of the Ruby amp devouring 9V batteries?"

At least partially because the 9V battery is going the way of the dodo. At over $2 a battery, it only takes ten to equal the cost of a 9V wall wart.

And this observation does get you your chance to become immortal as the first person to report an observed phenomena! 

[Govmnt_Lacky]

And this observation does get you your chance to become immortal as the first person to report an observed phenomena! 

Immortaliy is not what I was hoping for.

A portable practice amp that does not suck the life out of a battery faster than Uncle Sam reaches in and pulls money out of my pockets would have been suffiecient! 

Remember.... work hard.... Millions of people on welfare are depending on you! 

[arawn]

6 d cells is still pretty portable

[R.G.]

Immortaliy is not what I was hoping for.
A portable practice amp that does not suck the life out of a battery faster than Uncle Sam reaches in and pulls money out of my pockets would have been suffiecient! 

Hmm. OK, you're going to have to go Class D to get any better. Historically, the first audio amplifiers were class A. Class B was invented not least to cut the total amplifier power wasted to produce a given output. Class C is unlistenable to most people. Class D is an end-run around the limits of linear amplification.

But diminishing returns gets you. A decent Class B amp has an efficiency that peaks at about 78% at full power, and is more like 50% at medium powers. This is compared to the audio power out. The speaker itself eats the 78% of power at full power, leaving only 22% in the amp itself. At medium powers, it's about half and half. So even if the amp was 100% efficient, you'd still only extend full-power battery operation by a fifth (roughly). You'd double it if all operation was at half-power. Neither of these approximations are completely true any time, but it gives you a place to start digging.

You can mostly eliminate the amplifier part of the problem by going to a more complex Class D amp.  That will get more than 90% of the power going to the speaker at all times. Reports vary about the quality of the sound, depending on the design.

Remember.... work hard.... Millions of people on welfare are depending on you! 

That scares me a little. No, a lot. 

[PRR]

> My only question is "Why are there no threads/reports of the Ruby amp devouring 9V batteries?"

I have no idea. It isn't a VD, that you have to report it.

> It appears that the rise in current draw is "to be expected"

This could be inferred from the data-sheet. Idle is about 9V and 4mA (0.004A) or 0.036 Watts. Full power is (from graphs) 0.7W to load and 0.25W of "dissipation": the chip gets hot. Where is this 0.7W+0.25W= nearly a whole watt coming from? Must come from battery. 0.9W at 9V is about 0.1A. Rise from 0.004A to 0.1A happens when full signal is applied.

It could also be inferred from boom-car observations. Put a 1,000 Watt amplifier in the trunk. Every bass-hit, the headlights dim. Or from boom-car sales poof: they sell Farad capacitors to store energy for the amplifier so sudden thuds don't dim the lights.

> It appears that the rise in current draw is ...parallel with the rise in volume.

This is an astute observation. Taking "volume" as "signal voltage", a class B amp's current rises directly as signal voltage. 10% signal voltage is 10% of maximum current.

However its Power rises as square of signal. So 10% voltage is 1% power.

> A decent Class B amp has an efficiency .... At medium powers, it's about half and half.

This may be a good guide for boom-cars and guitar-amps. For unclipped speech/music, the average voltage is under 10%. Power demand is 10% of max. Power output is 1% of max. So the average efficiency of near-clipped speech/music is 7%. And worse and worse if not played to clipping. Only a few percent of your precious battery juice reaches the speaker. Class D can, ideally, suck "only" the power which the speaker requires. This would give a HUGE boost in battery life. (May be why my new MP3 player runs 20 hours while the old one pooped in 2.) In non-battery systems, the heatsink advantages are also compelling.

[Gurner]

AAA __________ $0.83/each ____ 1,200mAH ____ 1.8WH ____ $0.46/WH
AA ___________ $0.60/each ____ 2,700mAH ______ 4WH ____ $0.41/WH
9V ___________ $3.50-$1.80 _____ 300mAH_____ 2.7WH ____ $0.70/WH

Modern day 9v Alkalines typically deliver about 550maH (over 80% more than you've listed)

IMHO, the table doesn't emphasise the main issue .....sure, an AA @ 2700maH is mighty impressive, but that's only at 1.5V....so you then have to scale up on the number of batteries to get a useable volatge ....four as a minimum for an audio amplifier (but even then, if regulating, then you're likely perilously close to the dropout voltage out the blocks!)  - ideally more batteries should be used, but then you're possibly running into space problems.

I agree that 9V battereies are feeble.....if I were putting together an audio amp that *had* to use batteries, then I'd make sure the enclosure was large enough to host at least six AA batteries, which as you've illustrated nicely, do represent the best bang per buck.

[R.G.]

> A decent Class B amp has an efficiency .... At medium powers, it's about half and half.
This may be a good guide for boom-cars and guitar-amps.

Actually, it's for continuous sine wave testing at medium powers. Real music and speech has a crest factor which varies, but is often about 20-30db, as you note.

I find these discussions of how to get more sound out of the same amp ( Like dude, can I make my amp more powerful by hooking it up to 240Vac?) a little funny at times. Not this one - it's focused on reasonable questions - but most of them. The lowest efficiency device in the signal chain is the speaker. It's almost criminal how little sound gets out for the electricity spent in the speaker. Changing speakers can double the apparent amplifier power easily enough.

And if you really want to get lots of sound at low amplifier power, go to horn loading on the speaker. This can cause a massive increase in speaker efficiency by literally matching the acoustic impedance of free air to the much higher acoustic impedance of the speaker cone with the horn acting as a matching transformer. (I know you know this, Paul. It's for the rest of the readers.)

That makes portability tougher, but it would run the necessary amplifier power for a practice amp right down into the mud.

[PRR]

> table doesn't emphasise the main issue ... only at 1.5V

It's in there, if not emphasized: the "WH" figure is V*A*T and is the total energy available.

> ... only at 1.5V

Yes, the native electrochemical voltage is low for amplifiers. You typically want a stack. The 9V gives that to you, convenient. But the 1.5V lumps allow you to stack only as high as you need..... if you want 0.5W in 8 ohms, you can use four cells for 6V, avoid wasting 1/3rd of the energy of a 9V (actually, a musician will play 1.5 times louder and use the WH 1.5 times faster).

> if regulating, then you're likely

I do not understand: why "regulate"? Over-voltage is not a problem. Linear regulators are dead-loss of energy which could be used for output. If you actually want to limit output to a fraction of the battery voltage, a "power" amp can usually do that without added parts.

There are other paths. You "can" make loudspeaker power from as little as 1.5V. 1.5Vpp is 0.5V RMS. For half-watt out you want a 0.5 ohm load. Perhaps eight 4-ohm cones, a mini-stack. Biasing is tricky but do-able. Alternatively do a Bridge with 2 ohm load, though the "small" loss-voltages sure add up. You can use a transformer to lever the amplifier to convenient loads and also reduce voltage loss. You can run a flyback booster to turn 1.5V into 6V 9V 12V DC etc. (Disposable flash-cameras wring 300V out of a AA-cell.)

> almost criminal how little sound gets out for the electricity spent in the speaker.

Indeed.

> Changing speakers can double the apparent amplifier power

Not just "apparent". For same amplifier power, four Tens will put double the actual acoustic power in the air as one Ten (100Hz-800Hz).

> go to horn loading

OK, you be the roadie this tour.

And, within limits, arrays of cones can approach horn efficiency. The face-area will be comparable. At 30%-40% efficiency, the total enclosed volume will be comparable, but the box is a more convenient package than the horn.

There seem to be good reasons guitarists do NOT use horns. IMHO the "woodiness" of a paper cone is musically useful for hard-body guitar sound. A well-coupled horn throat damps such modes. Guitar should speak from a single place, not woofer/tweeter style; this is more practical with cones than simple horns.

[Gurner]

I do not understand: why "regulate"? Over-voltage is not a problem. Linear regulators are dead-loss of energy which could be used for output. If you actually want to limit output to a fraction of the battery voltage, a "power" amp can usually do that without added parts.

I didn't mean regulate for the audio amp, but other stuff  that maybe in your circuit which might need a known reference level (or derivatives thereof)....comparators, MCUs etc.

As it goes I do a lot of stuff with PICs working in conjunction with audio amplifiers.....often I run a PIC at 5V ...but that needs to be a regulated 5V.... 4 x AAs are only providing 6V, & that's lot leaving a lot of leeway to the downside before your into the regulator's dropout voltage. So what I'm saying that for all you can get away with 4 x AA batteries, but for some that's insufficient....then you're looking at five or six AAs..and then you're into real estate issues in some cases .....if you've a decent source of dependable 9V for a reasonable price, whilst not optimum, often the space factor wins!