Your output inductors makes me worry ... 6amp rated ... Let's say you trying to put 4 ohms at 36v , how much amperes you get ? 9 ! . D-class is switch mode thing , and peak currents are very high and then comes delay ,pause ,deadtime , can call it so . Overall average current may look small ,like 1 ampere ,but actual current through mosfets may be 10ampere .Inductor should not saturate, or it loses its inductance and current increase many times ,that can also make ic blow easily. I think you should change output inductors to some big , 25amp rated , and not use yellow / white cores like in atx psu - they are crap ,heats a lot ,losses big amount of energy. Try some with dark blue rings , Dtmss model ,poland made , available on tme .
This is a very good suggestion.
As said, I tried different inductors and also putting them in parallel. For sure a single choke solution with high rating would be the better and simpler solution.
What I've seen today is something related strictly to how much the woofer is dampened into the box, changing this has changed very much the clipping threshold level.
The next days I want to try another woofer with bigger excursion, because some months ago I repaired a very big subwoofer, always 4 ohm, dual coil 46mm height and in the end I benchmarked it with the client. We tried it with a chinese version of the TDA7498 (not E) evaluation board made by Arceli. Sincerely I don't remember this kind of clipping issue.
As said, I tried different inductors and also putting them in parallel. For sure a single choke solution with high rating would be the better and simpler solution.
What I've seen today is something related strictly to how much the woofer is dampened into the box, changing this has changed very much the clipping threshold level.
The next days I want to try another woofer with bigger excursion, because some months ago I repaired a very big subwoofer, always 4 ohm, dual coil 46mm height and in the end I benchmarked it with the client. We tried it with a chinese version of the TDA7498 (not E) evaluation board made by Arceli. Sincerely I don't remember this kind of clipping issue.
Hi,
the TDA7498E is limited in some regions, like minimum load resistence and maximum supply voltage. If you use a "usual" 36V Lion Pack, it should have 10 cells in series (10SxP). Fully charged this will have a real output voltage of 46V. 36V is only the nominal, medium voltage during discharge! The TDA7498 (E) will tollerate this, until you give it a heavy input signal. When the current limiter kicks in, the ampere taken from the supply will be limited (it's a "limiter"!), but as it is build for 36v, it will blow at 46V if you insist to drive it with full input (like 4V). Because the limiter is made to tollerate short transients, you can still kill the chip if you insist on maximum power. You can not build such a chip idiot save, there is allways an idiot more clever than the best protection.
You described what you do and how the IC reacts. For me this seems quite normal, if you ignore data sheet limits. Not the chips fault. I have some TDA7498E boards from China, around 20 € now (was much less 2 years ago). Red with a small fan on the heat sink and the option to PBTL them. They sound (sorry to say that) simply fantastic and are very reliable. Maybe get one of these (be sure to buy the "E" version!) and see if they work in your configuration, if the explode like usual in your construction, look for the faults you made. Much cheaper than keeping on with your TDA fireworks.
Read the data sheet and understand how supply voltage and maximum power output are connected and how a low Ohm load will effect this.
Read the data sheet of the TDA7498, as it has some important information the "E" data sheet has missing!
https://www.ebay.it/itm/125355476614
the TDA7498E is limited in some regions, like minimum load resistence and maximum supply voltage. If you use a "usual" 36V Lion Pack, it should have 10 cells in series (10SxP). Fully charged this will have a real output voltage of 46V. 36V is only the nominal, medium voltage during discharge! The TDA7498 (E) will tollerate this, until you give it a heavy input signal. When the current limiter kicks in, the ampere taken from the supply will be limited (it's a "limiter"!), but as it is build for 36v, it will blow at 46V if you insist to drive it with full input (like 4V). Because the limiter is made to tollerate short transients, you can still kill the chip if you insist on maximum power. You can not build such a chip idiot save, there is allways an idiot more clever than the best protection.
You described what you do and how the IC reacts. For me this seems quite normal, if you ignore data sheet limits. Not the chips fault. I have some TDA7498E boards from China, around 20 € now (was much less 2 years ago). Red with a small fan on the heat sink and the option to PBTL them. They sound (sorry to say that) simply fantastic and are very reliable. Maybe get one of these (be sure to buy the "E" version!) and see if they work in your configuration, if the explode like usual in your construction, look for the faults you made. Much cheaper than keeping on with your TDA fireworks.
Read the data sheet and understand how supply voltage and maximum power output are connected and how a low Ohm load will effect this.
Read the data sheet of the TDA7498, as it has some important information the "E" data sheet has missing!
https://www.ebay.it/itm/125355476614
Hello Turbowatch2,
yes the 10s4p I'm using for reference now, when fully charged, shows only some decimal more than 42V and never exceeded this.
When running as normal STEREO mode on the two coils, sound is optimum and nothing to say about it, it does its work very well also with the bass guitar as expected due to the mechanical characteristics of the portable amplifier as now developed and patented (until the clipping is reached).
So no problem on the board, here I've did my best following indications on ST application note.
The actual level is controlled on the preamp board via regulable soft-clipper + low pass crossover to set a reasonable working point.
Clearly anybody, trying to get out more signal from it, could raise the source+preamp volume until the clipping would be reached again (at that point, loudness level on the tweeter is something like unlistenable, you must run away).
In the actual situation, this same clipping effect, clearly and not subtly audible, would be saying to the user "hey man, it'st too much for me, please turn down" and this can be expressly written into the operating manual.
The board you linked up can be very useful for the eventual purpose of PBTL operation. I want to catch one on Amazon or from other supplier because this would came in Italy in september and I'm not supposing to take so much time on this issue. My actual needing is to complete the first prototype, so I'm going straight for STEREO mode, the only working until now, trying today another woofer with different Xmax. I want to exclude this trivia as soon as possible.
Thank you very much for your participation and suggestions!
yes the 10s4p I'm using for reference now, when fully charged, shows only some decimal more than 42V and never exceeded this.
When running as normal STEREO mode on the two coils, sound is optimum and nothing to say about it, it does its work very well also with the bass guitar as expected due to the mechanical characteristics of the portable amplifier as now developed and patented (until the clipping is reached).
So no problem on the board, here I've did my best following indications on ST application note.
The actual level is controlled on the preamp board via regulable soft-clipper + low pass crossover to set a reasonable working point.
Clearly anybody, trying to get out more signal from it, could raise the source+preamp volume until the clipping would be reached again (at that point, loudness level on the tweeter is something like unlistenable, you must run away).
In the actual situation, this same clipping effect, clearly and not subtly audible, would be saying to the user "hey man, it'st too much for me, please turn down" and this can be expressly written into the operating manual.
The board you linked up can be very useful for the eventual purpose of PBTL operation. I want to catch one on Amazon or from other supplier because this would came in Italy in september and I'm not supposing to take so much time on this issue. My actual needing is to complete the first prototype, so I'm going straight for STEREO mode, the only working until now, trying today another woofer with different Xmax. I want to exclude this trivia as soon as possible.
Thank you very much for your participation and suggestions!
Hello everybody,
tested two hour ago same circuit, STEREO mode with another dual coil 4+4 with decent linear excursion, in free air. Test was pretty destructive for another IC because something like 6V p/p of input signal was used but worth of, as the average current circulating on the board reached 4.6A without any clipping and then I brought back volume to the minimum.
The whole board was build brand new, and IC didn't blowed during the test.
It blowed after I returned (with power switched off) to the small dual coil woofer used until now for reference, and then switched on with the input signal to minimum.
Now I can say Xmax matters in this specific case with this kind of tiny class D monster and have parameters enough to set the complementary circuitry to reach the best also avoiding damage.
Still thank you for your contributions and precious advices!
tested two hour ago same circuit, STEREO mode with another dual coil 4+4 with decent linear excursion, in free air. Test was pretty destructive for another IC because something like 6V p/p of input signal was used but worth of, as the average current circulating on the board reached 4.6A without any clipping and then I brought back volume to the minimum.
The whole board was build brand new, and IC didn't blowed during the test.
It blowed after I returned (with power switched off) to the small dual coil woofer used until now for reference, and then switched on with the input signal to minimum.
Now I can say Xmax matters in this specific case with this kind of tiny class D monster and have parameters enough to set the complementary circuitry to reach the best also avoiding damage.
Still thank you for your contributions and precious advices!
Maybe your dual speaker had burned coils , and one of coils or both were no longer 4 Ohms ? Then it would consume a lot of current and give small power output.
The woofer used until yesterday works and it's installed into the box. The one used at the second "turn on" was another piece of the same model and brand, but new, never used. I measured slighlty lower impedance (3.2ohm) but both coils were good. I don't know why it happened. Maybe I asked too much to the input pins during the test with this new speaker gently provided me by a seller for the trial.
I didn't mentioned that the test was produced with a pure sine tone of 50Hz.
I didn't mentioned that the test was produced with a pure sine tone of 50Hz.
Last edited:
It is recommended that you try TDA2030A.
In practical application, you will find that TDA2030A is very good. The only drawback is that it is too cheap.
Maybe it's only $0.20. There is no way to sell it very expensive.
But it's really very good.
Or tda7850
In practical application, you will find that TDA2030A is very good. The only drawback is that it is too cheap.
Maybe it's only $0.20. There is no way to sell it very expensive.
But it's really very good.
Or tda7850
Hello Lil Jim,
in substitution of the expensive LM1875, I would think you are referring to.
I've found some difficulties to find TDA2030A on the marked especially at this price, so I went straight for the one said as "the best" between these ICs.