The only stage of the TDA7293 that can be separated is the output stage, as far as i understand the datasheet (I have not made any measurement on this). Then driving this stage by an opamp like the LM4562 the maximum output voltage you can get is about the output voltage of your opamp - practical a little less because the gain of an output stage is always smaller than 1. Probably you can get the extra voltage gain needed by using a discrete design, but i wonder if it is worth the effort - the TDAs input and VAS stages work very well.
I am perfectly aware that the output is the only one that can be used separately and it is precisely the one that interests me .
What I do not know , and my guess is that applies to you as well , is which interior parts of the TDA establish which of its´s limits . so my interest concerning the output stage is to fgure out what its´s frequency-limit is and what the slew-rate without the internal circuitry preceeding it would be.
Concerning the complete input-circuitry yes it performs well but any decent op-amp is superior both in noise and distortion for one and some of them have three to five times higher slew rate and frequency-response.
A discreete vas between the input-opamp and the TDA powerstage , using BF471/BF472 complementary pair just for example , allows extremely high voltages and a transition-frequency of 60 MHz !!! yes I know they are video output transistors but Elektor used them in its superb CRESCENDO in 1983 a mosfet power stage with four Hitachi "SK135/2SD50
I have been repairing and building all kind of electronics from 1971 up to 2014 and from vacuum tubes through to digital circuitry so give me a little credit for all that. Like speakers 3 feet by 3 feet from the floor to ceiling in large discotheques in times where one had to make do with El34 valve amps and no more than 100 Watts for a room holding more than 500 people.
So input opamp like OPA for low noise and distortion , discrete complementary Vas to allow high rails anf frequency , and mosfet output stage from TDA for high slew rate and 50 volt rails could quite possibly be a very good mix and ALL THOSE TOGETHER CONSTITUTE A COMPOSITE AMP.
Keeping an open mind has its merits. I am not attacking nor critizising what you do and develop , I am merely saying there could be more than one good way of doing it. If you are happy with 25 volt rails fine . I am not. I want , at the same time , the highest rails , the best frequency response and the lowest possible noise and distortion.
Otherwise I might , as Tom has said , stay with the LM3886 and be done with it.
What I do not know , and my guess is that applies to you as well , is which interior parts of the TDA establish which of its´s limits . so my interest concerning the output stage is to fgure out what its´s frequency-limit is and what the slew-rate without the internal circuitry preceeding it would be.
Concerning the complete input-circuitry yes it performs well but any decent op-amp is superior both in noise and distortion for one and some of them have three to five times higher slew rate and frequency-response.
A discreete vas between the input-opamp and the TDA powerstage , using BF471/BF472 complementary pair just for example , allows extremely high voltages and a transition-frequency of 60 MHz !!! yes I know they are video output transistors but Elektor used them in its superb CRESCENDO in 1983 a mosfet power stage with four Hitachi "SK135/2SD50
I have been repairing and building all kind of electronics from 1971 up to 2014 and from vacuum tubes through to digital circuitry so give me a little credit for all that. Like speakers 3 feet by 3 feet from the floor to ceiling in large discotheques in times where one had to make do with El34 valve amps and no more than 100 Watts for a room holding more than 500 people.
So input opamp like OPA for low noise and distortion , discrete complementary Vas to allow high rails anf frequency , and mosfet output stage from TDA for high slew rate and 50 volt rails could quite possibly be a very good mix and ALL THOSE TOGETHER CONSTITUTE A COMPOSITE AMP.
Keeping an open mind has its merits. I am not attacking nor critizising what you do and develop , I am merely saying there could be more than one good way of doing it. If you are happy with 25 volt rails fine . I am not. I want , at the same time , the highest rails , the best frequency response and the lowest possible noise and distortion.
Otherwise I might , as Tom has said , stay with the LM3886 and be done with it.
Last edited:
Thank you Bernd , this is great news for me. While I am able to design circuits unfortunately I AM NOT able to optimize them , partly because of insufficient knowledge and partly because being retired I have limited instrumentation when it comes to fine tuning.
My idea is to take a good opamp , at plus minus 15 volts , take its output to the negative rail level by means of say a BC546 which can handle 60 volts , then a complementary VAs based on said BF471/BF472 to drive pin 11 of one or more TDA7293.
I think one can try this with the very VAS stage from that crescendo , of which I have the circuit diagram and exhaustive explanation in the original ELEKTOR magazine. Greetings
My idea is to take a good opamp , at plus minus 15 volts , take its output to the negative rail level by means of say a BC546 which can handle 60 volts , then a complementary VAs based on said BF471/BF472 to drive pin 11 of one or more TDA7293.
I think one can try this with the very VAS stage from that crescendo , of which I have the circuit diagram and exhaustive explanation in the original ELEKTOR magazine. Greetings
here is my idea - just an idea. But this will surely work when the TDAs output stage has the gain of ~1. If you look at the values....R14 should be at least 2k.
And for sure some more finetuning will be needed.
And to bring some light into your knowledge you can read the books about building and optimizing Audio Amplifiers. Just search for the names "Douglas Self" and "Bob Cordell"
And for sure some more finetuning will be needed.
And to bring some light into your knowledge you can read the books about building and optimizing Audio Amplifiers. Just search for the names "Douglas Self" and "Bob Cordell"
Attachments
Yeah. Those were annoying.
The OPA1642 is quite a bit higher noise than the LME49720, though. Higher distortion too. The OPA1656 is solid, though. It should be noted that while the typical specs for the OPA1656 are on par with those of the LME49720, there are hardly any test limits on the OPA1656. Basically only offset, offset drift, CMRR, AVOL, and supply current are guaranteed specs for the OPA1656 whereas the LME49720 adds THD+N, output noise, slew rate, and gain-bandwidth product to the list of guaranteed specs.
Guaranteed specs are those with either a minimum or maximum (or both) defined in the spec table of the data sheet. These values are tested in production. Typical specs are derived from device characterization.
The OPA1656 also has this really annoying phase wobble in the 1-10 MHz range that can make compensation a challenge in a composite amp.
Tom
Attachments
Analyzing the TDA7293 and studying its data sheet there are several things that jump to the eye.
Actually it is a double chip with separate input and output stages since the output of the input and vice versa are freely accesible on pin 11.
1.) the first possibility and the more obvious one is that by putting the chip into slave-mode one has a complete and independent mosfet output stage which has its own biasing circuitry , protection circuitry etc all built into the chip and so can be used with any circuit feeding it a signal provided the output is included in the feedback-loop of said circuit.
2.)The signal available at pin 11 can not only drive another TDA but any other external output-stage as long as the current needed to drive that stage does not overload the circuit inside the TDA which feeds pin 11. This opens up possible external output-stages driving loads as low as one or two ohms using discreete mosfets.
in short , using ones imagination there is a lot of things one can try and develop based on this chip-amp whose V.A.S ouput and powerstage-input are accessible.
Actually it is a double chip with separate input and output stages since the output of the input and vice versa are freely accesible on pin 11.
1.) the first possibility and the more obvious one is that by putting the chip into slave-mode one has a complete and independent mosfet output stage which has its own biasing circuitry , protection circuitry etc all built into the chip and so can be used with any circuit feeding it a signal provided the output is included in the feedback-loop of said circuit.
2.)The signal available at pin 11 can not only drive another TDA but any other external output-stage as long as the current needed to drive that stage does not overload the circuit inside the TDA which feeds pin 11. This opens up possible external output-stages driving loads as low as one or two ohms using discreete mosfets.
in short , using ones imagination there is a lot of things one can try and develop based on this chip-amp whose V.A.S ouput and powerstage-input are accessible.
Last edited:
Hi bepo .
I like the circuit idea from post 65 though I think that a LM49720 for the input opamp is plenty good enough. You get that for 2,50 euros instead of the almost 11 euros of the LT1028.
The pair BD139/BD140 though are too limited voltage-wise as they only go to 80 volts , not enough for 50volt-rails what do you think about MJE340/MJE350 for example?
I like the circuit idea from post 65 though I think that a LM49720 for the input opamp is plenty good enough. You get that for 2,50 euros instead of the almost 11 euros of the LT1028.
The pair BD139/BD140 though are too limited voltage-wise as they only go to 80 volts , not enough for 50volt-rails what do you think about MJE340/MJE350 for example?
post 65 shows only the topology i have in mind - i actually have something similar (discrete CFP-MOSFET-output stage) to hear my music with. But as i want to build an active dipol- speaker, similar to Linkwitz LX521 with one amplifier for each chassis, i no longer want to adjust anything and so i found my way to the composite amp.
I had a look at TDA7293 datasheet again, especially the diagram where the SOA is shown for the paralleled version: if you have low impedance speakers you should use three of the TDA7293 in parallel when having a power supply of +/- 50Volt. This is not what I want to do. I just want to use one TDA per channel - and having all speakers that never drop below 4 Ohm a maximum supply voltage of +/- 35Volt is the limit for me. Nevertheless the topology shown should be tried to see if its any better.
Bernd
I had a look at TDA7293 datasheet again, especially the diagram where the SOA is shown for the paralleled version: if you have low impedance speakers you should use three of the TDA7293 in parallel when having a power supply of +/- 50Volt. This is not what I want to do. I just want to use one TDA per channel - and having all speakers that never drop below 4 Ohm a maximum supply voltage of +/- 35Volt is the limit for me. Nevertheless the topology shown should be tried to see if its any better.
Bernd
Hi Bernd ,
Thats precisely my point Bi-Amping , Triamping and even 4 channel-amping needs 50volt rails for a 12"bass so if the circuit is designed for that to begin with one can always use it with lesser rails wheras the opposite is a no-go.
A composite amp with the layout for three TDA, can be used with only two or even only one ! And the boards from china , say JLCPCB , do not cost so very much. The work is practically the same , the pcb a little more expensive but good for all purposes downscaling then is easy.
Greetings Michael
Thats precisely my point Bi-Amping , Triamping and even 4 channel-amping needs 50volt rails for a 12"bass so if the circuit is designed for that to begin with one can always use it with lesser rails wheras the opposite is a no-go.
A composite amp with the layout for three TDA, can be used with only two or even only one ! And the boards from china , say JLCPCB , do not cost so very much. The work is practically the same , the pcb a little more expensive but good for all purposes downscaling then is easy.
Greetings Michael
Another point you have mentioned is the SOA of the TDA7293. I have already written about this in another thread:
My solution is simple , mount one or more TDA directly on the Heatsink isolating the Heatsink from the chassis as it will be at negative rail voltage if internal or...
if the heatsink is external mount the chips on a 4mm copperstrip with 10 square cms per chip and mount that on the heatsink with insulation .
Both versions bring the chip temp down by up to30 degrees compared to a chip mounted on insulation by increasing the heat transfer from the die to the heatsink making RTH case to heatsink a lot lower
My solution is simple , mount one or more TDA directly on the Heatsink isolating the Heatsink from the chassis as it will be at negative rail voltage if internal or...
if the heatsink is external mount the chips on a 4mm copperstrip with 10 square cms per chip and mount that on the heatsink with insulation .
Both versions bring the chip temp down by up to30 degrees compared to a chip mounted on insulation by increasing the heat transfer from the die to the heatsink making RTH case to heatsink a lot lower
Last edited:
Hello Mike,
i actually do not understand why you do not build the amplifier you have in mind by yourself.
With your knowledge and practical solutions you describe it would be an easy thing for you to do. The minimum equipment you need is a signal source and an oscilloscope - other tools like simulation- or EDA programs are available for free. For further measurements you can use a high quality external soundcard and the program ARTA, which is fully functional in the demo version... and i am sure you know all that- so why dont you give it a try? And if any problems arise you an always place the relavant question(s) here....
i actually do not understand why you do not build the amplifier you have in mind by yourself.
With your knowledge and practical solutions you describe it would be an easy thing for you to do. The minimum equipment you need is a signal source and an oscilloscope - other tools like simulation- or EDA programs are available for free. For further measurements you can use a high quality external soundcard and the program ARTA, which is fully functional in the demo version... and i am sure you know all that- so why dont you give it a try? And if any problems arise you an always place the relavant question(s) here....
easy one that .
I´m a "faultfinder" by 50 years of work , not a designer . I am perfectly able to analize a given circuit and finding out what is there and modify by trial and error . I have no idea where the design of a pcb is concerned !!! Then I am a 67 year old pensioner with a wife in a place consisting of a living-room and a bedroom , plus kitchenette and small bathroom and so my means are limited . in every way you can imagine . I can and do work at a friends house and can build an amp or a box but there is a limit of what he´ll put up with as well . So I try to see if younger people than I , maybe with less experience but with more modern knowledge might be interested in colaboration and , in this way , overcome my own limitations. Thats all there is to it.
I do have instrumentation including a Hameg Scope , dual channel 60 Mhz but nothing of whats necesary to check and measure low level noise and distortion
Most limiting my better half , not the youngest herself either , wont let me turn the flat into a workshop and I do not want a war at home.
I´m a "faultfinder" by 50 years of work , not a designer . I am perfectly able to analize a given circuit and finding out what is there and modify by trial and error . I have no idea where the design of a pcb is concerned !!! Then I am a 67 year old pensioner with a wife in a place consisting of a living-room and a bedroom , plus kitchenette and small bathroom and so my means are limited . in every way you can imagine . I can and do work at a friends house and can build an amp or a box but there is a limit of what he´ll put up with as well . So I try to see if younger people than I , maybe with less experience but with more modern knowledge might be interested in colaboration and , in this way , overcome my own limitations. Thats all there is to it.
I do have instrumentation including a Hameg Scope , dual channel 60 Mhz but nothing of whats necesary to check and measure low level noise and distortion
Most limiting my better half , not the youngest herself either , wont let me turn the flat into a workshop and I do not want a war at home.
Last edited:
My space to work on is also limited to a table 1,2m by 1,5m. And my tool to check noise and distortion is a high quality sound card. An analyzer, be it "Audio Precision" or "Rohde und Schwarz" is too expensive for hobby use only. I also started with electronics as a "faultfinder" and i am building circuits ever since. My knowledge in audio circuitry comes from analizing circuits that already existed, reading books (Self, Slone, Cordell), reading all kinds of datasheets and application notes, simulation and practical tests. When doing PCB work always have a look where the current is flowing , keep the traces as short as possible, especially on high impedance knots. Especially when high currents are flowing route the return path of the current on the opposite site of the PCB, so that the fields can cancel.
Another book that helped me to understand electronics circuitry is "Tietze und Schenk, Springer Verlag (written in german)"
Another book that helped me to understand electronics circuitry is "Tietze und Schenk, Springer Verlag (written in german)"
Reading and studying I have come across something for which
I would like to have opinion on to see if I might be right.
There is a very good design for the LM 3886 which , modified accordingly , should be aplicable to the TDA293 as well. I am referring to the my_ref design by Mauro Penasa , see the following link :
amp-lm3886-mauro-vidalgo/docs at master * diyaudioby/amp-lm3886-mauro-vidalgo * GitHub the last sub-link being in english
The gerbers can be found at :
amp-lm3886-mauro-vidalgo/gerbers at master * diyaudioby/amp-lm3886-mauro-vidalgo * GitHub
I would like to have opinion on to see if I might be right.
There is a very good design for the LM 3886 which , modified accordingly , should be aplicable to the TDA293 as well. I am referring to the my_ref design by Mauro Penasa , see the following link :
amp-lm3886-mauro-vidalgo/docs at master * diyaudioby/amp-lm3886-mauro-vidalgo * GitHub the last sub-link being in english
The gerbers can be found at :
amp-lm3886-mauro-vidalgo/gerbers at master * diyaudioby/amp-lm3886-mauro-vidalgo * GitHub
Last edited:
the my_ref design is for sure different to common designs - and it is well described. But the data stated at the end of the description dont really differ from common design. And as there are no measurements shown it is hard to say if this design sounds any better than common designs.... i will see if i find the time to analyze the circuits topology by simulation...
If adapted properly this should for sure also work with the TDAs.
The soundcard i am using is a terratec phase X24 FW (Yamaha GO46 is the same hardware) - but this is only possible by using an old PC with Windows XP with fire-wire-connector. (A real good hardware with a real horrible control-software)
Another card i had to test is the terratec DMX6fireUSB, Not as good as the X24FW but also very useful - and works with Windows10 (but absolutely not with ubuntu!)
by the way - the boards have arrived
If adapted properly this should for sure also work with the TDAs.
The soundcard i am using is a terratec phase X24 FW (Yamaha GO46 is the same hardware) - but this is only possible by using an old PC with Windows XP with fire-wire-connector. (A real good hardware with a real horrible control-software)
Another card i had to test is the terratec DMX6fireUSB, Not as good as the X24FW but also very useful - and works with Windows10 (but absolutely not with ubuntu!)
by the way - the boards have arrived