"Punch" mini-amp with a 32 Ohm speaker?

[Matthew Sanford]

Clint I'm not certain. It worked fine on an 8 ohm, I just figured the 0.25 watt version (I assumed, not marked, bought from china 2 decades ago) didn't have enough to do it. A LM1875 off 18v works fine ...the 386 had 9v, so maybe more power could've helped

[ElectricDruid]

Measurement reads around 30 Ohm, so it is a 32 Ohm speaker indeed.

Here's a picture:

Yes, but only 0.2W.

If the speaker is 32ohm, 0.2W then the original TDA7052 design you posted should be fine. Although the amp can't drive more than about 0.25W at 32ohm, the speaker can't handle more than that anyway, so there's no problem.

Whether a bigger speaker and a more powerful amp might *sound better* is a separate question...sometimes it's nice just to use up the bits we have around.

[Dormammu]

Salut!

I want to make this tiny 1W desk amp (for quick on-hand testing of other guitar things):



However, I have this salvaged 32 Ohm 0.2W speaker lying around without any use, so I would like to use that instead of 8 Ohm one on the schematic, if possible. I have a notion that some tweaking is required to do it, yet completely clueless what exactly it would be.

As always, I'm humbly seeking your advice. Thanks a many in advance!

If you don't want to burn the speaker, I would recommend choosing its RMS power 30-40% more than the amplifier can put out at the level with THD = 10%. And how many ohms its impedance is there — is not so important.
The diagram shown is ok.
When you burn this speaker, its no big deal  — you'll find another one. 

[Dormammu]

What about putting a 32 ohm, larger-wattage resistor across the speaker terminals?  That would bring it down to 16 ohms.  Or a 5.6 ohm resistor?  That would bring it down to just under 5 ohms.

For what purpose would anyone do this?

[PRR]

What about putting a 32 ohm, larger-wattage resistor across the speaker terminals?  That would bring it down to 16 ohms. ....

And the speaker splits with the resistor, gets half the total power. No real change. (Slight loss.)

The TDA7052 is fine for hi-Z at lower voltage. Because bridged. On the same supply, it will deliver twice the voltage (4X the power) of the single-ended LM386. And no large output cap!! At just 6V supply and 32 ohms, the TDA7052 will put out 0.225 Watts, SeneX225's speaker is rated 0.2Watts, a fair fit.

I sure agree that judging guitar tone through a two-inch speaker is like viewing pornography through a pinhole. You don't get the full effect. Even a Six is not big enough, even though Sixes have been sold for g-amps.

[SeneX225]

If the speaker is 32ohm, 0.2W then the original TDA7052 design you posted should be fine. Although the amp can't drive more than about 0.25W at 32ohm, the speaker can't handle more than that anyway, so there's no problem.

Whether a bigger speaker and a more powerful amp might *sound better* is a separate question...sometimes it's nice just to use up the bits we have around.

Thank you! Now I kind of want to try to make both LM386 and TDA7052 versions to see which sounds better (although, I suspect, the difference would be quite miniscule).

I sure agree that judging guitar tone through a two-inch speaker is like viewing pornography through a pinhole. You don't get the full effect. Even a Six is not big enough, even though Sixes have been sold for g-amps.

I understand I should not expect much of such a small speaker. The point of making this amp, however, is not the  quality of sound, but rather a something to plug in-progress stompboxes or audioprobe into without the need to go to another room for a soundcard or a proper amp.

[merlinb]

Then, perhaps, I should use another chip?

The power is limited by the supply voltage, not the chip. Run the same chip off 15V and you will be back to about 1W. Just build it, you can always change the speaker later. 0.25W is not exactly quiet, anywhow!

[Clint Eastwood]

Here, I made this up to fit your speaker. It should put out about 0.25 watt in class A.



Q2 and R1 will dissipate about 1 watt.

The input impedance is around 70kohm, so you may want to put a buffer in front of it.

[antonis]

Here, I made this up to fit your speaker. It should put out about 0.25 watt in class A.

Q2 and R1 will dissipate about 1 watt.
The input impedance is around 70kohm, so you may want to put a buffer in front of it.

Wonderful distortion circuit..!!

P.S.
Q2 Emitter is biased at about 800mV so, even with gain pot set at MIN (for pot value of 100R), any signal of amplitude greater than 5mV RMS should start to clip..  

[rutabaga bob]

What about putting a 32 ohm, larger-wattage resistor across the speaker terminals?  That would bring it down to 16 ohms.  Or a 5.6 ohm resistor?  That would bring it down to just under 5 ohms.

For what purpose would anyone do this?

Merely a comment on the first couple of posts, regarding resistance values.

[Dormammu]

What about putting a 32 ohm, larger-wattage resistor across the speaker terminals?  That would bring it down to 16 ohms.  Or a 5.6 ohm resistor?  That would bring it down to just under 5 ohms.

If we do this, we will get a load on the amp, it is not clear for what purpose, and all this power will pass by around the speaker. A speaker with this impedance is quite capable of handling the circuit shown on its own, as long as we don't try too hard to burn it with too much input.

[Clint Eastwood]

Q2 Emitter is biased at about 800mV so, even with gain pot set at MIN (for pot value of 100R), any signal of amplitude greater than 5mV RMS should start to clip..  

I do not follow you, Antonis.

Q2 emitter is at 6.4volt according to the simulator, and it is usually right.

With gain at minimum,  100ohm pot resistance, clipping starts at 100mv rms input.

With 35mv rms input, and gain set so that output is about half maximum power, THD is less than 1%.

With lots of gain/ little feedback, distortion of course gets high. But it is a nice spectrum, only low order harmonics are high.

I shall build this thing in the weekend and report back on the actual sound.

[antonis]

Q2 emitter is at 6.4volt according to the simulator, and it is usually right.

Dunno about your simulator by my simulator calls for about 800mV ..

Let take it in old-fashion way..
Q3 Base is biased at about 7.7V (12 x 220/340) and Emitter at about 7.10V (roughly)..
Q2 Emitter stands at about 100mV (7.1 x 33/2233)..

So, my simulator was very generous, for your aspect....

P.S.1
Your issue lies on R2/R6 values and ratio..
(with those values, Q3 should be permanently saturated..)
You see, Q3 Collector current is set by its Emitter voltage divided by R6 + R1 and not by Q1 V_EB divided by R2 (ignoring R5)..

P.S.2
Q1/Q2 act as a discrete Darligton (Sziklai actually) p-n-p transistor pair..
In the absence of R6, Q3 & Q1/Q2 arrangement should work as a CFP with almost 100% feedback and R2 value should be crucial for Q3 current hence Q3 h_FE in the way of Q3/Q2/Q1 arrangement should act as a perfect Emitter follower..
(high current gain but no voltage gain..)

[Clint Eastwood]

I don't know what you have fed your simulator, but I think it is not the schematic I posted. What you say about Q2 emitter voltage is just wrong.
Here it is again, I added some bias voltages:


Now tell me again what''s wrong with it? what do you mean with 'Q2 Emitter stands at about 100mV (7.1 x 33/2233)..'?

According to you 3mA should flow through R6 and Q3, but in my simulation it is only 0.3mA.

[antonis]

 

Never mind..
(each one has the right to trust his own simulator..)

If it works for you, I've not right to object..

[PRR]

Dammit, you figured it out while I was thinking.

Here's my rough notes:


The bias is not exactly optimized but close enough for jazz. (R1 could be 45r and output node could be nearer 8.4V.)

Output power is barely likely to exceed 0.2W---- this type resistor-coupled power amp can not much exceed 8.5% audio efficiency, even though other numbers are found in otherwise good books and essays.

[Clint Eastwood]

The bias is not exactly optimized but close enough for jazz. (R1 could be 45r and output node could be nearer 8.4V.)

What is the relation between bias point and music style according to you 

[PRR]

The bias is not exactly optimized but close enough for jazz. (R1 could be 45r and output node could be nearer 8.4V.)

What is the relation between bias point and music style according to you 

I am sorry if my flip remark offended anybody.

Backstory- I roadied for jazz musicians in the period just after Johnny Lytle put out an album of that name. (Well before the Van Morrison song.)

[Clint Eastwood]

Wow, that was around the time I was born.
Not offended at all, I was just curious!

[Rob Strand]

You can greatly increase the efficiency with a push-pull output stage and still get class A.  Here we only need about 60mA of bias current instead of 200mA or so.  You could back off the bias current to 10mA and get class AB with a further increase in efficiency - then you might as well stay with an LM386.

As shown class-A up to 180mW but full output goes upto 350mW.
THD 0.13% for class-A region and 0.31% up to clipping. 
We actually need to clip for the THD to rise to 1%.
(180mW roughly matches the design with the 33 ohm collector load.)

It's possible to simplify a few things but I'm not doing it.   The idea is only to show push-pull
you can get class A and more efficiency with push-pull.