If everything can easily be explained with "science", then everyone can make a better sounding amp, no?
It is not that the science is not applicable, it is just that the understanding of the science is low.
I agree that THD can be kept low enough that it should not be a problem.
Judging from the sound characteristics, it seems that TDA7294 has higher output impedance and thus lower damping factor. But what if, the "measured" damping factor of the TDA7294 found higher?
There must be more explanation. Fortunately for me, what I hear is more important than numbers. Numbers only needed to help achieve what I want to hear.But talking about DF, at what frequency you would measure them? 40Hz? Even if the TDA7294 is found to have lower than 180 DP, lets say 100, the DP might not be responsible for the difference in quality. I think that, even if you use an easy speaker load, TDA7294 will still sound like a mosfet amp.
EDIT: Ooppss.... 180 is for TDA7294? It is big enough! You see? It is not the DF.
By the way, your output mosfets are enclosed inside of the negative feedback loop of the inbuilt driver chip core. Do you still think you're hearing mosfets?
Well, if you really are hearing mosfets, this is from noise on the output buffer power pins bus railed into the predrive power pins, which are extremely sensitive and can be mistakenly polluted. The chip deserves a better job than that and something a lot better than the unfortunate datasheet examples. So, please double-check.I would say it is rather conceptual.
In case we are talking about small package like IC that could be intensively mathematically optimized and implemented with high precision by a modern FAB transconductance of BJT's NPN quasicomplimentary output stage seems to be a key factor to me.
LM3886 looks like more clever/professionally designed IC overall.
By the way, your output mosfets are enclosed inside of the negative feedback loop of the inbuilt driver chip core. Do you still think you're hearing mosfets?
I'm not sure if a mosfet has a sound at all
I mean, the sound of this amp is similar to the sound of a mosfet amp in general. I just checked the equivalent circuit, and I cannot see how the driver chip can change the sound of it. Of course, I don't understand electronics as well as you do, but imo this topology can never be able to compete with a discrete complementary design.
Imo, if better result is expected, it could be from adding another front end stage, like LM3886 in Mauro Penasa design.
The Mauro's circuit was the one that had been preferred by many people when compared to a discrete expensive lateral fet design. I was one of a few who prefer the mosfet, but TDA7294 cannot even compete with the discrete fet amplifier.
I can see from circuit topology the potential of an amp. This is not subjective, the prove is in the majority of people preference
The LM3886 MyRef and related designs are optimized designs for LM3886. The only solid conclusion that I can draw from comparison is that one chip is better supported than the other. Only one factor gives LM3886 the win, and that is the availability of full optimization.
It wouldn't have mattered which chip was chosen for such efforts, because whichever amplifier received a fully optimized support circuit wins.
Due to the inconvenient pinout and misleading datasheet of TDA7293 surrounding an excellent part in a veritable fortress of error, I honestly believe that optimization may require creativity with the pliers, a skeptical view of the datasheet and much testing. Yes there will be some problem avoidance, but that is true of every amplifier, both chip and discrete alike.
Both chips are capable of nearly identical high fidelity. At this time, one is easier to apply, thanks to Mauro Penasa
However, the TDA7293 is more conducive to guitar and violin amplifier than LM3886. Expanding the specialty circuits of the guitar amplifier TDA7293 to widen it to cover the entire audio band is a job of power circuit fine tuning. That works well with keyboard amplifiers. Any time there is a significant gap in such analysis, I would suggest that you check over layout and power circuit because a voltage is only as important as its reference, and that concept includes audio signal. Analysis result: It may be more difficult to stretch TDA7293 to cover the entire audio band in high fidelity. This is not impossible.
Your suggestion of Howland Current Pump TDA7293 and/or Nested, Composite TDA7293, seems to be a very good idea. Those can make wonderful amplifiers and the sturdy little chip, if running completely stable and cool, can conveniently support 4 ohm speakers. And, that is a nice feature in a chip amplifier. It also has available Slave Mode Parallel for TDA7293V and TDA7294H, possibly answering the question on how to do a parallel Howland Current Pump at much larger scale. Thank you for suggesting a part that has this facility.
P.S.
If you actually Do want to hear a Fet, build the MooseFet IRF510 triode simulation preamp by Greg van der Sluys of Classic Valve Design and then you will hear a Fet. I have one and it is useful to harmonically filter some types of sources, especially useful with compressed sources such as HD Radio, Sat Radio, MP4, Sat TV, so that instead of a droning, I hear a Fet.
It wouldn't have mattered which chip was chosen for such efforts, because whichever amplifier received a fully optimized support circuit wins.
Due to the inconvenient pinout and misleading datasheet of TDA7293 surrounding an excellent part in a veritable fortress of error, I honestly believe that optimization may require creativity with the pliers, a skeptical view of the datasheet and much testing. Yes there will be some problem avoidance, but that is true of every amplifier, both chip and discrete alike.
Both chips are capable of nearly identical high fidelity. At this time, one is easier to apply, thanks to Mauro Penasa
However, the TDA7293 is more conducive to guitar and violin amplifier than LM3886. Expanding the specialty circuits of the guitar amplifier TDA7293 to widen it to cover the entire audio band is a job of power circuit fine tuning. That works well with keyboard amplifiers. Any time there is a significant gap in such analysis, I would suggest that you check over layout and power circuit because a voltage is only as important as its reference, and that concept includes audio signal. Analysis result: It may be more difficult to stretch TDA7293 to cover the entire audio band in high fidelity. This is not impossible.
Your suggestion of Howland Current Pump TDA7293 and/or Nested, Composite TDA7293, seems to be a very good idea. Those can make wonderful amplifiers and the sturdy little chip, if running completely stable and cool, can conveniently support 4 ohm speakers. And, that is a nice feature in a chip amplifier. It also has available Slave Mode Parallel for TDA7293V and TDA7294H, possibly answering the question on how to do a parallel Howland Current Pump at much larger scale. Thank you for suggesting a part that has this facility.
P.S.
If you actually Do want to hear a Fet, build the MooseFet IRF510 triode simulation preamp by Greg van der Sluys of Classic Valve Design and then you will hear a Fet. I have one and it is useful to harmonically filter some types of sources, especially useful with compressed sources such as HD Radio, Sat Radio, MP4, Sat TV, so that instead of a droning, I hear a Fet.
Last edited:
check out post #481 under follow URL:
http://www.diyaudio.com/forums/chip-amps/226437-optimizing-tda7294-output-49.html
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.