A small, low-voltage, battery operated amplifier

Hi,

Here is a small, low-voltage (3V or less) amplifier.

I wanted a signal-tracer amplifier, very small, and battery operated to be used in completely floating situations: like typically to measure the noise across a component.

This amplifier fits the bill. It is rather crude, but has some interesting features: it uses same-sex output transistors, operates efficiently at ridiculously low voltage levels, and yet offers a decent quality: here, it is shown at max output power, and the THD is well under 2%, with a nice harmonic profile.
At lower power, the distortion is even lower.
Also note that the gain is very high, because of the intended application.
More reasonable values would further decrease the THD, and increase the input impedance.
The peak to peak output voltage before clipping reaches 2.5V with a 3V supply, which is quite impressive

No temperature compensation is shown; for optimum stability, R4 should be an NTC (or an NTC should be part of R4).
But even without compensation, the stability is acceptable thanks to the low supply voltage: the total quiescent current establishes itself at ~10mA.

The polarity of all transistors could be reversed, allowing the use of germanium transistors: the only adaptation required is R4, that should be changed to 39 or 47 ohm.

R9 is optional, it slightly improves the THD and the stability.

Have fun!
 

Attachments

  • MiniAmp.png
    MiniAmp.png
    157.4 KB · Views: 1,362
R1 appears to load the output with 4 ohms. Is this correct?
R1 symbolizes the load (speaker). The circuit has been designed for a minimum load of 4 ohm, but in practice it will accept even lower loads.
For an 8 ohm load, BC337-25 are sufficient.
If the amplifier is used exclusively with 32 ohm earphones, all the transistors can become BC547 or similar.

This original version is rather sensitive to the supply voltage.

The following variant eliminates this sensitivity completely at the expense of an additional transistor.
It is highly unorthodox, but it has some side benefits: it also compensates thermally the output stage, and provides a further reduction in distortion.
R4 has to be adjusted individually to set the quiescent current, much like the Vbe multiplier of a "normal" amplifier.
 

Attachments

  • MiniAmpVar.png
    MiniAmpVar.png
    149.3 KB · Views: 1,127

rjm

Member
Joined 2004
Paid Member
I see now. I wasn't sure earlier because a) you indicated this was for instrumentation, so I could see no reason for driving 4 ohms and b) it's tied to the positive rail.

Point b) above rather makes is unsuitable for headphones, unfortunately, at least conventionally wired ones with shared return wire.

Perhaps your circuit might find some application as the driving amplifier for little self-powered computer desktop speakers?
 
Point b) above rather makes is unsuitable for headphones, unfortunately, at least conventionally wired ones with shared return wire.

It does not matter: your headphones don't care about the absolute potential (if such a thing even exists) of the common return wire.
Anyway, the whole thing is supposed to be battery operated = floating.

The only case it could pose a problem is if the return of your headphones is hardwired to something tied to the input ground.
 
Here is the final incarnation of this amplifier in the signal-tracer.

I have used the transistor-compensated version, added a FET buffer, volume pot, input protection, 3 NiCad and their rudimentary charger.

The whole thing is built into the shell of an old, eighties-style touch-phone.

Works like a charm, you just have to unhook to power it, and there is also a possibility to connect earphones or an external speaker.
Quiescent current consumption is 9mA.

The quality is not bad for such a rudimentary circuit: I tried it on the big lab speaker and it made quite an impression. On 8 ohm, it reaches 2.7Vpp before clipping (with a 3V supply)
 

Attachments

  • DSCF0008.JPG
    DSCF0008.JPG
    312.9 KB · Views: 941
  • DSCF0009.JPG
    DSCF0009.JPG
    306.3 KB · Views: 868
Can I use it with a 5v supply?

Yes you can: the output DC level has to be modified to 2.5V with R6, and R4 (quiescent current) needs to be readjusted.

The value on the schematic is only indicative, you have to adjust it individually. The adjustment is quite sensitive, but once it is done it remains perfectly stable.

If a fixed 5V supply is used, then the resistor/NTC compensation is sufficient
 

Attachments

  • Mini5V.png
    Mini5V.png
    156.9 KB · Views: 822
Is MC7805, or maybe MC7806 okay to use?
I forgot to ask about max operating voltage when driving 8 ohm speaker.
Is that amp an NTP input or a singleton input?
And, would it be okay to bridge it or is the THD a bit too high for that?
P.S.
Other problem: It hangs up when it touches my face. :)
P.P.S.
Figured a name for it: Tracerphone! :)
 
Last edited:
Is MC7805, or maybe MC7806 okay to use?
Yes, with a good output bypass cap, like 1000µ. With a regulated supply, the NTC compensation would suffice.
I forgot to ask about max operating voltage when driving 8 ohm speaker.
The same as for 4 ohm. This is a low voltage, low power amplifier and for higher voltages there are lots of (better) alternatives
Is that amp an NTP input or a singleton input?
It doesn't apply, it is an inverting amplifier without a + input.
And, would it be okay to bridge it or is the THD a bit too high for that?
It would be like turbo-charging your ride-on mower: makes little sense

Other problem: It hangs up when it touches my face. :)
It is intended to be mainly used laid on a table. Only occasionally as a "phone", for low-level, hard to catch noises.
Figured a name for it: Tracerphone! :)
Nice one
 
On the schematic in post 9, how to set the gain for a very weak MP3 player?
With the original values the gain will be huge, even for a very very weak MP3 player: it is around 40, meaning you will already clip at input voltages of ~25mV rms for a 3V supply.

For applications other than a signal tracer, the gain should be reduced by increasing R8: for a normal player, 33K or even more should be OK, you'll still have plenty of margin, and the distortion will be reduced by a 10x factor.

For a very weak signal source, you could try 10K: the full power will be attained for an input equivalent to 0.175mW on 32 ohm. I doubt even the weakest of players could play so low.
 
Thank you!
For a very weak signal source, you could try 10K: the full power will be attained for an input equivalent to 0.175mW on 32 ohm. I doubt even the weakest of players could play so low.
Well, I wish to use a weak signal since the digiplayers will distort if set much more than ~70% (although it varies per player). These things all seem to have their sweet spots and most of the digital volume controls are pointless because of only one seemly setting. A rather high gain, or adjustable gain, amplifier seems to help a lot.

P.S.
Can this little amp drive my clipnipper if I use the 1.5v led's and trim a bit more carefully, or is the extraneous current consumption detrimental?
 
Oh, idea.
If R8 is 22k and V+ is 5vdc. . . Then how much input voltage causes clipping?
Perhaps I could find a diode of a bit less than that voltage, put the diode series to a resistor and use it to engage a voltage divider at input?

I would choose to use a diode+resistor load for the upswing and a diode+resistor load for the downswing. The dual asymmetric soft clipper could be set to leave most symmetric signal unmodified but half-soft-clip asymmetric excursions.

If I had the ballpark figure for an input voltage that causes overload, I could then make R8 a trimmer to dial in the soft clipper.

Well, I don't know if this is a good idea. It may be nothing other than an extra complicated way to make a thump stopper (overload protector). But, there's only 2 diodes and 2 resistors to it (d+r for up and d+r for down). Possibly the soft clipper goes parallel to R6?

Edit:
The idea also seems a useful protection for a tracer, for cases where we didn't want to input huge signals.
 
Last edited:
Oh, idea.
If R8 is 22k and V+ is 5vdc. . . Then how much input voltage causes clipping?
Simple arithmetics: the gain becomes 120/22~=5.5, and since the clipping occurs practically at the rails, the input voltage is 5/5.5Vpp=0.91Vpp.
Translated in rms value, that is 0.91/2.828=320mVrms
Perhaps I could find a diode of a bit less than that voltage, put the diode series to a resistor and use it to engage a voltage divider at input?

I would choose to use a diode+resistor load for the upswing and a diode+resistor load for the downswing. The dual asymmetric soft clipper could be set to leave most symmetric signal unmodified but half-soft-clip asymmetric excursions.

If I had the ballpark figure for an input voltage that causes overload, I could then make R8 a trimmer to dial in the soft clipper.

Well, I don't know if this is a good idea. It may be nothing other than an extra complicated way to make a thump stopper (overload protector). But, there's only 2 diodes and 2 resistors to it (d+r for up and d+r for down).
That's up to you....
Possibly the soft clipper goes parallel to R6?
Certainly not R6: it is a virtual ground, with practically no voltage at all.
R7 could be a possibility, but you would need a blocking capacitor, since there also is a DC voltage

Edit:
The idea also seems a useful protection for a tracer, for cases where we didn't want to input huge signals.
No, a tracer doesn't need that kind of subtlety: the protection is very rugged and consists of series resistor and clamping diodes to the supply rails.
For this type of instrument, a bullet-proof protection is essential.
 
Certainly not R6: it is a virtual ground, with practically no voltage at all. R7 could be a possibility, but you would need a blocking capacitor, since there also is a DC voltage.
Well, perhaps I didn't communicate right. So, here's the question as a photograph (attached). The diodes are silicon+1n5819=~0.9v, and then R8 dialed for adjust. Ac voltage to operate the diodes is provided by the source device at input. R8/R6 looks like input voltage divider and R6 looks like input load. But this is wrong (photo attached)?
 

Attachments

  • tracerphone.gif
    tracerphone.gif
    8.3 KB · Views: 335