Hello everybody . I have read through a large number of threads and I am aware that it will be difficult to come up with something new as both the LM 3886 and the different versions of the TDA series appear in a good many of them.
But I still would like to start a new thread about the TDA 7293 in particular because I think that there is a lot of potential in it provided it gets the same kind of attention and dedicated effort the 3886 has got for so many years.
On the other hand I do not like the sound of Class D amps at least for now but looking around it seems to be the future trend. I do not want to say there are no good ones around but I prefer analog amplification.
I think there is a trend to active speakers at the same time there is an ever growing number of so called "digital" amps . Just as with LED lighting the magical word lately seems to be efficiency, not perfection anymore. Eventually this might lead to the dying out of traditional chip amps though I hope it will be a long while as yet. But their price will rise........
So to get back to my starting point and TDA 7293:
If all the people here in this forum were to get together and design both an optimized circuit and optimized PCB for a modulat system based on the TDA 7293 where one could add one or two of the chips , then bridge two of those , so driving the TDA by a buffer that can invert or not we could not only get the very maximum out of these amps but possibly help other folks as well who now buy any one of those EBAY-and/or Amazon sold SUPERAMPLIFIERS from China .
Make a one time recipe to cook up an excellent TDA7293. Concentrated in one thread with the obvious advantages for people new to all this ?
What do you say ? Michael
But I still would like to start a new thread about the TDA 7293 in particular because I think that there is a lot of potential in it provided it gets the same kind of attention and dedicated effort the 3886 has got for so many years.
On the other hand I do not like the sound of Class D amps at least for now but looking around it seems to be the future trend. I do not want to say there are no good ones around but I prefer analog amplification.
I think there is a trend to active speakers at the same time there is an ever growing number of so called "digital" amps . Just as with LED lighting the magical word lately seems to be efficiency, not perfection anymore. Eventually this might lead to the dying out of traditional chip amps though I hope it will be a long while as yet. But their price will rise........
So to get back to my starting point and TDA 7293:
If all the people here in this forum were to get together and design both an optimized circuit and optimized PCB for a modulat system based on the TDA 7293 where one could add one or two of the chips , then bridge two of those , so driving the TDA by a buffer that can invert or not we could not only get the very maximum out of these amps but possibly help other folks as well who now buy any one of those EBAY-and/or Amazon sold SUPERAMPLIFIERS from China .
Make a one time recipe to cook up an excellent TDA7293. Concentrated in one thread with the obvious advantages for people new to all this ?
What do you say ? Michael
Last edited by a moderator:
Concerning active speakers .
For any reasonably decent size 2 way ( or more way )speaker it is better and cheaper to use two or more chip amp modules and an active filter than a large passive one , and I say large because to get good coils for the inductors one needs aircoils , and relatively large and expensive foil capacitors as well.
An electronic crossover using op-amps and a number of resistors and small capacitors including its stabilized rails at plus minus 15 volts can be done on a 50x60 milimeter pcb and be 24dB linkwitz Riley as well while one is at it for a 2-way and 50x90mm pcb for a three-way without too much effort and be connected to the amps rails por power. The materials will be cheaper than a good passive crossover and the frequencies adaptable to different drivers , with the added advantage of absolutely no cross-influence between speakers.
For those concened with power and dissipation the Midrange and Tweeter could be run on analog chip amps while the bass could be a class D for efficency if that is important.
there are so many ways to play this once one has good amps for mid and high frequencies.
For any reasonably decent size 2 way ( or more way )speaker it is better and cheaper to use two or more chip amp modules and an active filter than a large passive one , and I say large because to get good coils for the inductors one needs aircoils , and relatively large and expensive foil capacitors as well.
An electronic crossover using op-amps and a number of resistors and small capacitors including its stabilized rails at plus minus 15 volts can be done on a 50x60 milimeter pcb and be 24dB linkwitz Riley as well while one is at it for a 2-way and 50x90mm pcb for a three-way without too much effort and be connected to the amps rails por power. The materials will be cheaper than a good passive crossover and the frequencies adaptable to different drivers , with the added advantage of absolutely no cross-influence between speakers.
For those concened with power and dissipation the Midrange and Tweeter could be run on analog chip amps while the bass could be a class D for efficency if that is important.
there are so many ways to play this once one has good amps for mid and high frequencies.
Last edited by a moderator:
"Done right" is almost a trademark of Neurochrome. I have not the least doubt that Tom has done extremely well in particular on measurable parameters. He may have had a head-start from his time at TI and may still have connections so he can measure THD below -120dB. That can absolutely be said to be "done right". For me "done right" implies as "good as possible" and there we are going to face two problems: measuring THD below -100dB and, as best possible includes the perception of the sound quality, finding reference persons who can tell what sounds best. When we operate at such perfectionist levels below -100dB, it is difficult to find a suited reference for evaluating improvement attempts.
I have no doubt that also TDA7293 can be used to very high standards if someone(s) cares to do the work. As even change of a signal capacitor from one seeming good type to another good type may affect the sound, the work is more extensive that what may be assumed for a start.
A good initiative!
I have no doubt that also TDA7293 can be used to very high standards if someone(s) cares to do the work. As even change of a signal capacitor from one seeming good type to another good type may affect the sound, the work is more extensive that what may be assumed for a start.
A good initiative!
Last edited:
A single TDA7293 with DC-servo has been done by Dibya,xrk971&jhofland, it measures pretty good and reviews are good too.
A modular approach could be worthwile to some and be something "new".
Looking at the AN-1192 and fig.17 for example you could have a PCB with an onboard bridging adapter like in fig.2 here:
Bridging Adapter For Power Amps
Follow OUT+ with two paralleled TDA7293 and connect OUT- with a shielded cable to an identical second PCB where you´d skip the input buffer.
I thought about paralleling chips but thought its not worth the effort as it sure makes PCB-design more complicate.
To route your supply voltages to 2 chips with low-inductance traces/polygons isn´t trivial (IMHO). But then I haven´t tried it and of course you have the possibility to go to a 4-layer-board which can be had very cheaply these days.
I am aware of the fact and I have two hopes first one being that Tom will forgive me for using these words in this threads title and the second one being my hope that he might be willing to participate in this in some way or other.
What I mean by "modular approach" could , for example, be to design a PCB for the TDA7293 which instead of being mounted in the traditional horizontal position would be mounted vertical in such a way that two or more of them could be "stacked" horizontally with the necesary conections from the master-board to the slave-boards being straight wires through the according holes in the PCB´s.
The TDA´s are predestined for parallel use and the parasitic capacitances and inductances of short wires which are perpendicular to the PCB´s should be less of a problem than the equivalent traces on the bords ? The sensitive inputs are only on the master PCB.
So if the same PCB has provisions to be used as both master or slave by simply leaving components out and to add the two jumpers to ground the inputs on the slave-PCB´s then all that is left is to place the six holes for the wires crossing from one board to the next.
In this way one only PCB could be used universally for any possible combination and could be kept relatively small at the same time.
The TDA´s are predestined for parallel use and the parasitic capacitances and inductances of short wires which are perpendicular to the PCB´s should be less of a problem than the equivalent traces on the bords ? The sensitive inputs are only on the master PCB.
So if the same PCB has provisions to be used as both master or slave by simply leaving components out and to add the two jumpers to ground the inputs on the slave-PCB´s then all that is left is to place the six holes for the wires crossing from one board to the next.
In this way one only PCB could be used universally for any possible combination and could be kept relatively small at the same time.
I personally would design a PCB for 2 chips in parallel.
Much easier to construct.
Your solution can be done I guess but there´s more to consider also.
If you need inspiration about stacking PCBs, have a look at Peter Daniels gainclone constructions; for example:
GAINCLONE - strona 377 - DIY - Audiostereo.pl
Not very service-friendly but his constructions show very creative ways of building amps (wiring and thermal design).
Much easier to construct.
Your solution can be done I guess but there´s more to consider also.
If you need inspiration about stacking PCBs, have a look at Peter Daniels gainclone constructions; for example:
GAINCLONE - strona 377 - DIY - Audiostereo.pl
Not very service-friendly but his constructions show very creative ways of building amps (wiring and thermal design).
to SESEBE :
you thread is very interesting but it has nothing to do with what I wrote above.
What I am asking is can the output stage of a TDA be used as the powerstage of another VAS , like say another high quality high voltage op-amp , by feeding that amps output signal into pin 11 of a slave-configured TDA so that the only part of the TDA used is its mosfet output-stage
That way one could use several TDA simply as a cheap well-compensated dual mosfet transistor pair. And never mind the TDA´s input stage.
What I am after is an amp with a gain between 1 and 4 otherwise impossible with the TDA7293
you thread is very interesting but it has nothing to do with what I wrote above.
What I am asking is can the output stage of a TDA be used as the powerstage of another VAS , like say another high quality high voltage op-amp , by feeding that amps output signal into pin 11 of a slave-configured TDA so that the only part of the TDA used is its mosfet output-stage
That way one could use several TDA simply as a cheap well-compensated dual mosfet transistor pair. And never mind the TDA´s input stage.
What I am after is an amp with a gain between 1 and 4 otherwise impossible with the TDA7293
Last edited:
Have you read this? :
https://www.diyaudio.com/forums/chip-amps/197398-tda7293-single-bridge-parallel.html#post2729693
And before the question arises and you want to bother panson_hk, I asked and he did not pursue this project,
so the information in that thread is all there is for now!?
https://www.diyaudio.com/forums/chip-amps/197398-tda7293-single-bridge-parallel.html#post2729693
And before the question arises and you want to bother panson_hk, I asked and he did not pursue this project,
so the information in that thread is all there is for now!?
Last edited:
Yes, it goes in the direction you want to head.
I personally think sesebe´s example is actually nearer to a "BPA200"
or at least easier to implement.
See how he´s paralleled the chips in a single-layer-PCB.
Leave out the Mosfets, put a buffer/inverter or let´s call it a bridging adapter in front and you´re done.
Make this a 2- or even 4-layer board and it will be as compact as possible.
In case you want to bridge solder a cable to a 2nd (identical) PCB.
And yeah, you could make those stackable too.
I personally think sesebe´s example is actually nearer to a "BPA200"
or at least easier to implement.
See how he´s paralleled the chips in a single-layer-PCB.
Leave out the Mosfets, put a buffer/inverter or let´s call it a bridging adapter in front and you´re done.
Make this a 2- or even 4-layer board and it will be as compact as possible.
In case you want to bridge solder a cable to a 2nd (identical) PCB.
And yeah, you could make those stackable too.
well the direction is not yet fixed.
My idea to make this "stackable" is because :
1.) I want the board (s) to be as small as possible.
2.) It should be scalable and as flexible as possible
3.) It should be possible to use the slave-boards with another IC as driver, using pin 11
So the necesary decoupling caps must be populated on each board.
4.) The input components only on the master board of course and the feed-back resistor
soldered directly on pins 2 and 14 to avoid trace inductance.
in short as universally usable as possible while not sacrificing performance. 'Of course stacked boards using only wires perpendicular to them are not exactly service-friendly but in a well-designd and rugged circuit that should not be too big a problem especially for DIYers for whom to to it without too much money is important!?
My idea to make this "stackable" is because :
1.) I want the board (s) to be as small as possible.
2.) It should be scalable and as flexible as possible
3.) It should be possible to use the slave-boards with another IC as driver, using pin 11
So the necesary decoupling caps must be populated on each board.
4.) The input components only on the master board of course and the feed-back resistor
soldered directly on pins 2 and 14 to avoid trace inductance.
in short as universally usable as possible while not sacrificing performance. 'Of course stacked boards using only wires perpendicular to them are not exactly service-friendly but in a well-designd and rugged circuit that should not be too big a problem especially for DIYers for whom to to it without too much money is important!?
Last edited:
I already see this kind of project same were bat I do not remember were.
For me the cip power IC are a close subject because the slewrate is veri low for my preference. You can see the measured slewrate when paralel 2 IC in my topic. Is under 10V/µsec.
A slewrate of 10V/uSec allows for a swing rail to rail of above 100Khz at 40V-rails and by bridging two chips that same slewrate becomes 20V/uSec. I would say nobody can hear a difference between an amp with that and another with twice that for example . What I pursue is an exellent sounding amp , not a champion in otherwise useless measurements.
Besides what good is a slewrate of XYZ if then you have to limit the frequency-response of the amp to keep it stable and free from oscillations??----Snake-oil---------
Anything, say up from 50 to 60 KHz,, for arguments sake, is not only totally inaudible but beyond any tweeters capabilities as well and only heating said tweeters voice- coil for nothing reducing the useful power capacity of the tweeter.
The amps and speakers I am building are for people , not the bats in neighbours chimney.
If I were to build FM-transmitters I believe I would have to go to some other forum
Besides what good is a slewrate of XYZ if then you have to limit the frequency-response of the amp to keep it stable and free from oscillations??----Snake-oil---------
Anything, say up from 50 to 60 KHz,, for arguments sake, is not only totally inaudible but beyond any tweeters capabilities as well and only heating said tweeters voice- coil for nothing reducing the useful power capacity of the tweeter.
The amps and speakers I am building are for people , not the bats in neighbours chimney.
If I were to build FM-transmitters I believe I would have to go to some other forum
Last edited:
The TDA circuit is pretty simple so not easy to improve on.
My only bad experience was from poor pcb layout where I hadn't kept the feedback resistor path short and I got some oscillation.
I really dont like parallel amps so I would go for a bridged mode setup.
It should be easy enough to come up with a dual tda pcb and a phase splitter front end for bridging.
My only bad experience was from poor pcb layout where I hadn't kept the feedback resistor path short and I got some oscillation.
I really dont like parallel amps so I would go for a bridged mode setup.
It should be easy enough to come up with a dual tda pcb and a phase splitter front end for bridging.
Heh. Well, not quite. If I wanted to actually claim trademark rights to the term, I would have added the TM symbol. Done Right™. But you can't trademark commonly used words, so it'd have to be DoneRight™ or something similarly ugly.
I appreciate the notion, though. Thanks for the chuckle.
It's worked so far.
The title did catch my eye for exactly that reason, but I also noted the long pause and the question mark.So in summary: No offence taken, but I hope you'll use the term in the spirit in which it was intended, i.e. to get the best darn TDA7293 amp that physics will allow.
I'm not quite sure what your goal is, so it's hard for me to provide specific input at this point, but here're some thoughts:
- Output current limit: 6.5 A. That rules out bridging (see the bottom of page 11 of the data sheet) and requires two TDA chips (or lower supply voltage) for full output swing into 4 Ω. You can see my math for the LM3886 here: Taming the LM3886 Chip Amplifier: Output Power – Neurochrome. Just plug in 6.5 A instead of 7 A and rerun the numbers. Note that the 6.5 A is the typical number (i.e., it may be higher or lower for the chip that you get) whereas the output current of the LM3886 is 11 A typical, 7 A guaranteed min. (across temperature, device variation, supply voltage, etc.).
- For a large package, the TDA7293 has remarkably high thermal resistance from junction to case. 1.5 K/W, worst case (not guaranteed, page 11). That will limit performance with high power supply voltage.
- The data sheet contains remarkably little information. Where's the AVOL curve for example? How about a graph of output current vs temperature? That alone is enough for me to shy away from using it. But that's me.
That said, for those who are interested in characterizing the chip before using it (or more plug-and-see-what-happens rather than plug-n-play than I am) it certainly looks like it could be the basis for a capable amplifier.
I've said it before and I'll say it again: You will need a very large heat sink for this. Note the 1.5 K/W theta_jc above. To keep the device below 150 ºC at 25 ºC ambient, you'll need a heat sink that can dissipate 62 W and only increase in temperature by 25 ºC. That means you need a heat sink specified for: 25/(62*1.257) = 0.32 K/W. This is assuming music reproduction (14 dB crest factor) with the peaks at clipping levels and a 4 Ω load. For sine wave, you're looking at 89 W dissipated, so you get down to 0.22 K/W. That means fan cooling for sure. It's no coincidence that discrete 100 W rated power amps tend to use 2-3 output devices, per side, in parallel.
The 1.257 factor is the compensation factor that accounts for the temperature differential between the heat sink and the ambient air. Heat sinks are more efficient when they're hotter, so manufacturers commonly specify them for 70 ºC temperature rise. The 1.257 factor applies if the heat sink is 30 ºC hotter than ambient air (yes, I said 25 ºC above, but the table only went to 30 ºC).
You can solve the thermal issue in many ways. By far the easiest is to lower the supply voltage, but that gets you down into LM3886 territory. The LM3886 is vastly better characterized (i.e., has a much more detailed data sheet), so why not just use that then? Unless you're going for a "same but different" type of design, which is valid too. I bet you could make a very capable amp if you used two TDA7293s in parallel (data sheet Fig. 11) and lowered the supply voltage to ±36 V. That's coincidentally a common SMPS output voltage, so maybe power it with an SMPS and squeeze it into a ModuShop Mini Dissipante 3U (0.40 K/W) chassis. That could be nice.
Use can also use many devices in parallel. The TDA7293 is well set up for that. You don't have to stop at two devices either...
I'm also intrigued by the "high efficiency" schematic in the TDA data sheet (Fig. 8). That's technically Class H rather than Class AB, but whatever. Before careening down that path, I recommend reading the relevant chapters in Self's and Cordell's books. Class G and Class H are commonly plagued by a switching transient when the output stage transitions from the lower rails to the higher ones. That can wreck performance. I'm thinking the two 1 uH inductors in Fig. 8 are to deal with this.
Then you add the PCB layout. You'll need something better than an eBay connect-the-dots layout if you want the full performance from the chip. I would go for four layers for that reason. I'd dedicate a large swath of three layers to V+, V-, and power ground. But that part is more straight-forward once you get the amp prototype working.
It sounds like a fun project. Best of luck with it.
Tom