Point 2 Point (no PCB) for TDA7293, TDA7294, TDA7295, TDA7296.

Got the new little CRC power supply done!

It has 2200u first caps, and the tank/reservoir section is 6x2200u per rail. The CRC resistor value is 0.3R and diode bypassed to prevent sag. The dual bridge rectifier is Fairchild Stealth. It has 2.2K drainers at first caps locale. It has orange LED indicators for each rail, near the output side. That board is 2-5/8" x 8" size.
 

Attachments

  • Power 005.jpg
    Power 005.jpg
    41.8 KB · Views: 7,627
Hi there Mr Daniel. this past days a friend came with a 8¨+tweeter active speaker;
it was a discrete amp about 50W, it was so burnt with a big hole thru the board.
So with the same psu as original, put in a 3886GC this i had it almost finished ( less
PS) so just try it! Maybe not sogood design its almost the same at M Feuerbacher
site.Tried it and sounded good,think it needs tweaking, but my friend just loved it and as he needed for playing, just took it like that.
Now reading thru your postings, i´m aware now that maybe the 7293 bridged conf. that i was considering, now i see what your answer was; the ps xformer is the limiting factor,
so i´ll try with a single chip, and go for your first post schem.
my xfrmr is center tapped, good quality , dual bridges cannot be used? trying to find out.
Well enough, have some duties before solder time.Parallel chips dont increase output
cause its an 8 ohm spkr. right? maybe just to make it more reliable? what would you suggest?
Anyway i think i´ll settle with 68k-2k7-220u for the feedback circuit i has a preamp and equalizer integrated at the input board.
Ok thanks for all your attention and time to answer our postings.
Martin Martinnez
Uruapan Michoacan
 
Parallel chips dont increase output cause its an 8 ohm spkr. right?
No, this is not a discrete amplifier.
The theory is correct, except for chip amplifier limiter engagement point of about 45 watts, whereupon the LM3886 will begin screeching an internal hard clipper or TDA7294 will omit the bass via internal soft limiter and in either case, it becomes more difficult to hear bass when you turn it up past about 45 watts or so. The limiter influence is either mild or inactive until you turn it up and wonder why it isn't as good as a discrete amp.

With your transformer, a PARALLEL chip amp and an 8 ohm speaker, the on-chip limiter (45W+45W=90W) can't interfere with your 70W application. Each chip sees half of 70W, 35W apiece and then on-chip limiters never switch on.

So, I have to report that the Parallel TDA7294S/TDA7293 Master/Slave configuration is ready to use and fits your application. This is the easiest parallel chip amp possible. I will be using it to upgrade my receiver as soon as the LM3886's are removed.

Notes:
At this time, I am currently working on "energetic efficiency" for the solo chip. I may have different news to report upon completion of that task. It is far more difficult with just 1 chip. And thus, the 1-chip application does not yet have the chair flipping power of the parallel amp.

It is possible to put the feedback resistor and feedback-shunt resistor(s) trackside, on bottom of the boards, in which case they sound almost identical to a point to point effort (no board related tone).
 
It is such a shame to waste intelligent resources
on these Nasty little integrated chips designed for WalMart applications.

Limiters don't have any specific affinity for program material in the
ordinary case. Tweets can trigger circuit protection as well as Woofs.
All amplifiers sound sickly once the output hits the rails.

Trying to apply magic components to these dogs will not render any
substantial performance improvements. These chips are great for
trying to make a silk purse out of a sow's ear.
 
Or a much smaller silk purse, without the sow's ear problem.

An ordinary design of discrete amplifier offers one practical thing the chip doesn't offer, and that's output transistor linearity at higher power. For example. . .

A chip amplifier's built in output transistors are a bit smaller.
With the solo TDA7294/5/6, your options are:
(listed worst option first)
#1. Push for both bass and treble, causing unpleasant midrange instead
#2. Push for pretty treble and accept a bit less bass as the cost
#3. Push for bass power and install treble boost as a compensation
#4. Use Nested/Composite/Howland for linearity tracker/servo
#5. Under volt it to make the smaller more linear amplifier
Using the TDA7293 adds this option:
#6. Parallel output devices to make the larger more linear amplifier.

Clearly, this is an output transistor size versus linearity issue.
There is why the chips demonstrate strong affinity for good power circuits and the topic of energetic efficiency / dynamic power / peak power / music power / back-emf synchronization, which is a variety of language all used to describe Not riding the brakes on brief transient signal. Little output device can still deliver whopping big transient surges for clarity and bass too. But they never make suitable toasters, aka constant current reliance cripples them.

So, let's try the parameter matching.
Martin Martinez asked about a chip amplifier to use for one 8 ohm speaker and he has a 25+25vac transformer of 125VA. On the list above, there are three linear amplifier options, #4 is overly complex, #5 doesn't work with his transformer voltage, so #6, parallel output devices is the logical choice.
 
thanks again
i dont mind sewing ears
i´ve been using chipamps because that´s what many amps use (instrument amps, power mixers, home systems etc) and also to replace very old amp sections.
They usually sound nicer and louder, and more abuse withstanding.
so its gonna be parallel 7293,p2p, and still deciding the feedback string
Let me thank you again.
time ius very valuable 4 me, so i apreciate much your attention.
martin martinez
uruapan michoacan
 
They usually sound nicer and louder, and more abuse withstanding. so its gonna be parallel 7293, p2p, and still deciding the feedback string
http://www.diyaudio.com/forums/chip...r-slave-style-no-lossy-emitter-resistors.html <--click there :D
I've described it (with the new values) in that parallel amp thread at Post#1, but did not have time to make a new schematic. If you can go by the description to get the part values, the operating voltage is same as yours and should make same results, a storm of hi-fi.
 
Last edited:
Recipe for TDA7294 "mids&treble" amplifier:
120k feedback resistor, but 2k7 resistor series to 500R trimmer for feedback-shunt resistor. Cap sizes at small signal, is 220u+1u electrolytics for nfb-shunt cap, and the input cap is not larger than 0.47uF (preferably smaller). Speaker negative has a series capacitor not larger than 330uF (preferably smaller). And, due to reduced bass current, the amplifier can be run easily from regulators for high quality sound. This Mids&Treble amplifier does quite well at amplifying the little higher pitched audio signals to huge proportions for much amazement. Each chip can drive an array of midranges and tweeters. You'd need very powerful bass amplifiers just to keep up with it--about 500W to 1 kilowatt Class D to a woofer with a passive crossover may be enough, but that depends on speaker efficiencies.
 
Last edited:
ok! that´s interesting; maybe with active xover filter.
And bridged tda´s + bipolar for subs?.
So the hi gain together with input cap value and nfb string
plus the cap in series at the output will make it more dynamic
for hi freq and most lof is filtered?.
Does it need compensation if i use RC filter: c in series with +,
R series w -, at the tweeter?
i use something like 4x series-parallel 2,2uf and 1x 15-20W, 15-52 ohm.
 
Last edited:
Does it need compensation if. . .
It doesn't need external compensation added when:
1). There's good amplifier power layout (not all boards meet that criteria)
2). When run at a gain of in the neighborhood of approximately 38X*
*Both voltage and layout can change that gain factor.

P.S.
It is possible to use compensation as an effect with the mids&treble application, if one wants to make Piezo tweeter arrays sound a bit more like electrostatic/ribbon tweeter sound. Apparently, overdo is much too easily done, so the output would need to be measured.
P.P.S.
So far, compensation experiments have failed to allow a lower gain setting; however, I believe that a LOT LESS power supply voltage will allow a lower gain setting and produce a practical output quality despite lower gain and a lot less power. However, success on that would depend on speaker efficiency. Also applicable to small devices, such as table radios that don't need 70 watts.
 
Last edited:
Thanks!
then maybe is enough with
what is already in the circuit.
need to check for temperature.

well this pair is 2x 75k-2,7k-220uf
noninverting so 29x gain, right?
and the rest goes by your schematic.
It´s gonna be bridged with the
25-0-25 xfrmer and 4x4700uf.
(local 2x220uf on each chip).
as soon as i get back will
finish and test it.
 
Martin, that's my older schematic and designed for under-volting. I did not realize at the time that it was the low voltage that made for good tone despite low gain. If you have some trouble with the tone, replace the 75k with 100k and it should be really very merry (at exactly 38X).

You might try one of them that way and compare to see if there's a useful difference.
For bass amp, see my newer schematic.

Also, first power up needs some precautions:
Light Bulb Tester saves the day! <--link
Check for DC offset before connecting speaker.

P.S.
A bipolar capacitor or back to back polar cap in series with the speaker negative cable, is generally a good to protect test bench speakers, and valuable speakers as well. If used as a permanent install with valuable speakers, right-size the capacitance to promote best bass extension by aiming the roll-off to coincide with, or somewhat lower than, speaker roll-off ("hook the corner frequency"); and, also bypass the output cap assembly with a Nichicon ES 4.7uF or 10uF little green bipolar electrolytic for pure treble.
The approach is very popular to decrease distortion from tweeters, midranges and subwoofers; however, it shouldn't be surprising that it also works for normal woofers. The performance boost is very slight with woofers; however, it is a combination of a little bit more bass at the speaker and a little bit more headroom at the amplifier, depending on signal, of course, since the idea of using the protector for a performance boost involves the efficiency of giving the speaker a bit less of what it can't do. Smaller woofers with more roll-off respond to that a bit more noticeably than larger woofers with less roll-off.
Somewhere in-between the amplifier designer versus speaker designer, this well working protection/efficiency has gone missing. But it remains true that speakers don't require DC and amplifiers don't really require excess loading all the way down to 1hz. So, there is probably some gains to be had by defining the audio band more precisely and corresponding to equipment operating tolerances.
I tried this protector on my test speaker and was pleasantly surprised at the output, so I tried it on all of my speakers, and it worked every time. The main difference is that with the protector in place, cranking up the bass on the EQ results in more real bass instead of more booming. That's fairly appreciable. See also Bob Cordell's website on the topic. If you combine my caps idea with Bob Cordell's active EQ, you'll get a huge amount of real quality bass.
 
Last edited: