T&L Publications

Oh-OK, I thought you were referring to the output stage.
Give me a day to locate the original schematic. if it doesn't show up, I will print one up and upload it. I am working on a project right now that I had to build up a slightly modified version for some test setups. One thing that caught my attention is that in one of the tests I have to drive a completely reactive load at several watts and it was completely stable across the range. Also, I hope you aren't one of those nuts that demands 0.001% harmonic distortion , as this one is rated for 0.5% distortion under all conditions.

Thanks Robert. I look forward to seeing what one looks like.

"0.001% THD" - no no no no no, more like 0.000000000000000001%!!!

Couldn't resist! Gotta have my wire with gain...

Thanks again,

CeaSaR

CeaSaR wrote:Anyway, if you were to move the transistor base(s) to the output of the op-amp, would that be more like a "Losmandy"? That's the impression I am getting. This also seems to be the way many "digital", or class D amps, are set up. Either that or I am in mental overload and am mixing to many fleeting images together.

Consider that when you are building one of these general purpose op-amps directly driving the bases of a complementary symmetrical output section, the op-amp is likely current limited to 20mA of output. A pair of output transistors like the TI41C/42C may have a current gain of just 20. This limits total current output to 400mA. Into an 8 ohm speaker, supplies 1.28 watts peak (P = I^2 * R), or just 640mW RMS. HALF that for a 4 Ohms.

If you want to correct crossover distortion by using diodes, biasing the output transistors gets a bit more complicated. If you want to use Darlington outputs for more current gain, crossover distortion gets worse.

Quoting Rosanne Rosanadana: "It's always something".

I think that the word you were fishing for could be "beta multiplier", the two transistor pair that works like a darlington but uses one PNP and one NPN. That is the name we used circa 1970 for those audio output transistor pairs.

Caesar
Here is the Losmandy circuit and a modified one. As stated I built up a couple of the originals years ago for my stereo system and a couple of modified ones for other uses and they all have a good track record. The original is built on a 4" x 4" x 2" finned heat sink with a small circuit board attached to and mounted under it.The 4009 looks like an ordinary metal can op-amp with a small clip on heat fin attached. The modified version will put out just over 10 Watts into an 8 ohm load with a +/- 16 volt supply.
http://www.postimage.org/image.php?v=PqhBBR9
I am going to do a search for the 4009 or a suitable replacement and will let you know if anything turns up. This is shown with a split supply but will work just as well with a single supply as long as R3 is ac coupled to ground as well as ac coupling the out put to speaker. But the split supply has all the advantages.

EDITED 1-30-10

I did find the OEM and EIA# for the 4009. It is made by National Semi.
Compensated version:
LM143 +/- 40 volt supply
LM343 +/- 36 Volt supply
Slew Rate 2.5 V/ us
Uncompensated Version:
LM144 +/- 40 Volt supply
LM344 +/- 36 Volt supply
Slew Rate 10V/ us
These are obsolete and the best price I could find on them to date is $15 a pop.Unfortunately this makes the price of Nationals Overture series much more attractive as you get the whole amp in one package ,albeit higher cost.
There is one newer device that is available at reasonable cost-- OPA548T. Has a +/-30V supply rating and good slew rate. This one would conceivably allow the amp to put out 45+ watts rms / 8 ohm.They are available thru Mouser for $10.

Robert,

Thanks for the info. I looked into a couple of the opamps that are included in my copy of CircuitMaker and
used the LF353 and your updated version (04's, 06's and 55's) at +-16V. Talk about flat! 300mV input, R2
set to 42K gives 10.3 W and the response curve shows a spike at about 17k-18kHz, but the scale reads
32.7 dB from top to bottom on the Y axis using auto view, so the "spike" is actually extremely minute.
Manual view shows a flat line all the way across. Heck of simulation.

Looking around I came across the TDA1562, datasheet, that uses a 14.4V single ended supply for up to 70W!
It is for use in an automotive environment, but it is really hard to beat something like that. Kinda humbles a
guy, if you know what I mean.

Hey Bob, I'm looking at IC's!

CeaSaR

**Edited for clarification re: waveform scale.

Hello Ceasar
In actual use, the full power response checks out DC to >25 KHz dead flat w/8 ohm load.
That Phillips Op-Amp is pretty impressive, but still for automotive use, it is limited to 14 volts total supply. And as with all auto amps - rated with a 4 ohm load. At 8 ohms the power is cut in half (35 watts in this case). Also internally are two amps bridged so that the speakers cannot be returned to common. This may or may not be a problem. I wonder if this chip is still available as most Phillips stuff is obsolete and only sold in Asia by the box load. If I were looking for HIGH power in a single chip, I would investigate the National Semi Overture series with 68 Watts being the leader of that group and line operated higher voltage supplys.
One thing to keep in mind though is that if you have an amp running at 50 watts and then increase that power to 100 watts, thats only a 3 db increase and will only sound a little louder. After all once you get to LOUD, increases beyond don't sound a whole lot louder. I think the higher power becomes important when using inefficient speakers or playing classical music with tremendous dynamic range.

CeaSaR wrote: Hey Bob, I'm looking at IC's!

CeaSaR

Finally!

Bob Scott wrote:Finally!

Thought you might be relieved.

CeaSaR

Robert,

I edited my previous post to you for clarity re: fequency response.

I was using the TDA1562 as an example of where a real engineer can go when they are given such constraints.

I've seen the Nat.Sem. Overture series and they are quite nice. They might be in my future...

As for the power factor, I have usually been satisfied with things in the lower powered range, for I have been
fortunate to have fairly efficient speakers (hovering around the 90dB range at 1W - manu's specs). Most of my
listening is done with less than 20W, probably less than 10W 90% of the time.

Yep, I am aware of the ever decreasing power vs. loudness curve.
For example, starting arbitrarily at:
1w=90dB
10w=100dB
100w=110dB
1000w=120db
etc.

At this point, when it comes to the higher powered stuff, I have been thinking in terms of music creation (guitar
amps, etc.) because of son #1's propensity for his "electric twanger" (50 points if you can get those 2 references).
I want to build something that might take him for a few more years before he hits the "real" world. Besides, who
wouldn't like to make a name for themselves in the music world? "Dude, I just got a new Dad's amp and it kicks..."
Oh the dreams....

CeaSaR

Got your edit and makes more sense to me now-thank you. The freq. response I gave you was more or less from memory, so I just ran a Quick sweep with an 8 ohm load and actually it was PERFECTLY flat from DC to 70KHz. Only down 3 db at 100 KHz. I know its meaningless for anybody but a twisted audiophile, but it does give this amp a lot of versatility such as servo systems and high power Ultra sonic systems. Also seems to handle radical phase shifts well.

Wait! Ultrasonics is another thread...

It is a very parts friendly design. Definitely worth consideration. I have to find a new sim software or find out how
to put new parts into this one. I'd love to be able to use the TIP series or others to see what can be done.

Thanks,

CeaSaR

I was/am interested in the original (well, v6.0?) circuit for use as a headphone amp and also to power a set of Philips SBP1100
speakers I picked up at Walmart. Pretty decent small 2" speakers rated 1.5W that I'd like to use with my laptop mostly. I wanted
something better than the usual LM386-type chip amp for these unusual hemispherically-enclosed speakers.
Hmmm... is that Losmandy amp stable into 2 ohms?
"...most Philips stuff is obsolete..."? I thought they just spun off the semiconductor division and rebranded it as NXP. I am having
trouble finding a datasheet for the LM2822M. Anyone know where it's hiding?

Stable into 2 ohms - yes. Phillips has spun off, but I believe most of the older thru hole chips are no longer made (at least in the specialty versions).

Scratch