TDA7293 TDA7294 TDA7293V TDA7294V without Zobel Network by driven Band Pass Subwoofer

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I want to introduce an appropriate compensation network like a zobel network/boucherot network (for prevent damaged TDA7293V/TDA7294V) so as a serial network consist of a (maybe 20 winding) coil and resistor for a commercial active bandpass subwoofer.

By a power module from a friend I note the absense of any parts between speaker cable and power output pin 14.
I cannot believe, that this is is a right design, because I don't discover any other reasons for the destroy of this chip.
Thank you very much for the appropriate design rules.
 
Last edited:
Ways to blow out the TDA7294 (what not to do):
Heatsink insulator failure
Shoulder washer failure
Omission of the ~220u caps from the predrive power pins
Omission of the ~picofareds RF blocking caps/RC's
Running the output overcurrent too long
Using a fake device
Setting the gain below 20 or other instability causing
Using very dirty power
Not using standby via time delay per datasheet
Having the standby engage before the mute
Having the outputs run lower voltage than the predrive

There's a big list of what not to do.
However, when set up decently, the TDA7294's are extremely durable and the chip itself is capable of high-fidelity. You'll have to search to find good schematic examples--they're not in the datasheet.

TDA7293/4 Parallel is a decent choice for a subwoofer amp, but a solo TDA7293/4 is a bad choice for a subwoofer amp. The onboard protection decreases low bass output when the amp is at full blast. Not only is that counterproductive for a subwoofer amp, but the protection is insufficient. The datasheet shows undersize bootstrap cap and undersize NFB cap, all of which are counterproductive for subwoofer amplifiers; however, the footnotes in the datasheet contain needed information, so ignore the bad schematics and read the good literature.
 
I want to introduce an appropriate compensation network like a zobel network/boucherot network (for prevent damaged TDA7293V/TDA7294V) so as a serial network consist of a (maybe 20 winding) coil and resistor for a commercial active bandpass subwoofer.

By a power module from a friend I note the absense of any parts between speaker cable and power output pin 14.
I cannot believe, that this is is a right design, because I don't discover any other reasons for the destroy of this chip.
Thank you very much for the appropriate design rules.

It says on the first page of the datasheet no output network needed. However, there is this caveat on the schematic, also on the first page.

"The Boucherot cell R6, C10, normally not necessary for a stable operation it could be needed in presence of particular load impedances at VS<±25V"

So what voltage was your chip running from?
 
Last edited:
Ways to blow out the TDA7294 (what not to do):
Heatsink insulator failure
Shoulder washer failure
Omission of the ~220u caps from the predrive power pins
Omission of the ~picofareds RF blocking caps/RC's
Running the output overcurrent too long
Using a fake device
Setting the gain below 20 or other instability causing
Using very dirty power
Not using standby via time delay per datasheet
Having the standby engage before the mute
Having the outputs run lower voltage than the predrive

There's a big list of what not to do.
However, when set up decently, the TDA7294's are extremely durable and the chip itself is capable of high-fidelity. You'll have to search to find good schematic examples--they're not in the datasheet.

TDA7293/4 Parallel is a decent choice for a subwoofer amp, but a solo TDA7293/4 is a bad choice for a subwoofer amp. The onboard protection decreases low bass output when the amp is at full blast. Not only is that counterproductive for a subwoofer amp, but the protection is insufficient. The datasheet shows undersize bootstrap cap and undersize NFB cap, all of which are counterproductive for subwoofer amplifiers; however, the footnotes in the datasheet contain needed information, so ignore the bad schematics and read the good literature.
Thank you very much for examples of possible reasons.
I have check all marked resons in italics. This I exclude.
Capacitors between predriver voltage rail pins and power output voltage rail pins are not present - also no low pass resistor between this pins. PIN 7+13 (positive rail) so as 8/15 (negative rail) are shorted.

I don't understand this term:
Running the output overcurrent too long
Means this, that the internal protection could be faulty ?
The bass transducer is a Vifa M22 - go to
FS: Vifa 8" M22 Midbass/Subs - CarAudioForum.com
and
2 Vifa M22 W-09, Selbstbau / DIY - HIFI-FORUM
using in a band pass subwoofer (Fundament 1.1, Bresc Audio, formerly company name: Gass & Dexheimer, Kirchberg/Hunsrück)
there are no additional parts to optimize the impedance (perhaps a reactive load cause parasitic inductance). Therefore I will introduce a zobel network (boucherot network) from output to GND (10 ohms, 150nF). Perhaps a big 200 ohm resistor in parallel to the load is additional useful.

"Not using standby via time delay per datasheet" and
"Having the standby engage before the mute" I will check - therefore I must create the schematic for the St-By control.
The power supply provides +/-36 until 37 volts dependend of the mains voltage (measured without power output, i. e. without signal modulation).
The filter consists of a 12 db variable Sallen Key low pass filter but additional 2x 12 db high pass filter with two different cut-off frequencies (protection against ultra low bass signals).

P.S. How looks the curvature for a bandpass subwoofer impedance in general?
perhaps like that from the last pages about
http://www.eighteensound.com/staticContent/applications/kits/18Sound_kit18_dual.pdf
or
http://www.diysubwoofers.org/projects/home/inf10bp/
 

Attachments

  • Vifa M22WR09-front.jpg
    Vifa M22WR09-front.jpg
    78.1 KB · Views: 654
  • Vifa M22WR09-rear.jpg
    Vifa M22WR09-rear.jpg
    98 KB · Views: 649
Last edited:
Capacitors between predriver voltage rail pins and power output voltage rail pins are not present - also no low pass resistor between this pins.
What????? Capacitors don't go between since that would block dc and never work. I'm sure you would have noticed zero power. What actually goes between is your choice of either a solid wire or a diode. Normal configuration is solid wire or trace.
http://www.st.com/internet/com/TECHNICAL_RESOURCES/TECHNICAL_LITERATURE/DATASHEET/CD00000017.pdf <<read it
Note: Please consult the TDA7294 datasheet or other schematics to find plausible locations for capacitors.
PIN 7+13 (positive rail) so as 8/15 (negative rail) are shorted.
Having the predrive power pins in a bus power confiuration with the output buffer's power pins is perfectly acceptable normal configuration for this chip.

The datasheet shows it right, except that their signal caps are somewhat too small and their power caps are much too big.
I don't understand this term:
Running the output overcurrent too long Means this, that the internal protection could be faulty ?
The internal protection is ALWAYS slightly deficient. A 4 ohm speaker will cause a single chip to blow. A 2 ohm load will cause a parallel amp to blow.
They die of overcurrent, but not immediately.
"Not using standby via time delay per datasheet" and "Having the standby engage before the mute" I will check - therefore I must create the schematic for the St-By control.
Easy way:
Use 10k to power the mute (just like the datasheet, except omit the cap).
Power the standby per the TDA7294 datasheet's suggestion of resistor with cap delay.
 
Last edited:
Capacitors between predriver voltage rail pins and power output voltage rail pins are not present - also no low pass resistor between this pins. PIN 7+13 (positive rail) so as 8/15 (negative rail) are shorted.
What????? Capacitors don't go between since that would block dc and never work. I'm sure you would have noticed zero power. What actually goes between is your choice of either a solid wire or a diode. Normal configuration is solid wire or trace.
http://www.st.com/internet/com/TECHNICAL_RESOURCES/TECHNICAL_LITERATURE/DATASHEET/CD00000017.pdf <<read it
Note: Please consult the TDA7294 datasheet or other schematics to find plausible locations for capacitors.
You haven't understand, sorry for not clearly description. There are two caps behind the rectifier (Value: each 10.000uF/50V) - one between the pos. rail and earth and a second between neg rail and earth. But no extra caps through low pass filter resistors to the rails respective an independend power supply for the predriver pins. This means, that the power driver pin and predriver pin (13/7-15/8) are shorted at both rails - as show on page 1.17 about
http://www.st.com/internet/com/TECHNICAL_RESOURCES/TECHNICAL_LITERATURE/DATASHEET/CD00000017.pdf
Bypass-caps are also not in use.

Datasheet for the used driver (4 ohm version) you will find here:
http://www.diyaudio.com/forums/multi-way/107629-looking-m22wr-09-08-datasheet.html#post3157683
Thank you for the other interesting facts and advices.
 
Last edited:
The bus rail topology is not a cause of breakage.
Lack of 220u or 330u caps at the amp board makes the amp run hotter.
Noise at the amp board (layout) makes the amp run hotter.
Lack of RF filtering makes the amp run hotter.
Poor enclosure ventilation makes the amp run hotter.
Lackluster power filtering makes the amp run hotter.
Running a 4 ohm speaker from a single chip makes the amp run MUCH HOTTER and will wear out quickly.

The cause of breakage in that subwoofer is overheat+overcurrent: They should have used a parallel amplifier when driving a 4 ohm speaker.

SO, if you want to know how to fix it, install this: TDA7293 Parallel Kit
 
I want to introduce an appropriate compensation network like a zobel network/boucherot network (for prevent damaged TDA7293V/TDA7294V) so as a serial network consist of a (maybe 20 winding) coil and resistor for a commercial active bandpass subwoofer.

By a power module from a friend I note the absense of any parts between speaker cable and power output pin 14.
I cannot believe, that this is is a right design, because I don't discover any other reasons for the destroy of this chip.
Thank you very much for the appropriate design rules.

Please, post electronic diagram. It may not be necessary to attach zobel network/boucherot network, but to tell you exactly I need to see electronic diagram.
 
The power supply provides +/-36. . .
And with a 4 ohm speaker? WOW!! Max spec for TDA7293 is 100w, and with 4 ohm speaker the rail max is +/-32.7vdc, or approximately a 22+22 transformer. Also, max spec is not for long term use (except with the most careful design possible, including overcurrent management). For lasting, it should be run less than max.

However, your subwoofer amp designer chose to use more than max. OOPS! Cause of breakage is confirmed as overcurrent failure (short lifespan, wearing out).
 
And with a 4 ohm speaker? WOW!! Max spec for TDA7293 is 100w, and with 4 ohm speaker the rail max is +/-32.7vdc, or approximately a 22+22 transformer. Also, max spec is not for long term use (except with the most careful design possible, including overcurrent management). For lasting, it should be run less than max.

However, your subwoofer amp designer chose to use more than max. OOPS! Cause of breakage is confirmed as overcurrent failure (short lifespan, wearing out).
The most easy solution therefore seems to be the use of only one half winding of the power supply transformer to create a supply voltage of approximately +/- 18V. That means 6db less undistorted maximum sound pressure level (because only a quarter of the output power is at disposal).
But then the reliability would be sufficient even at 4 ohms load impedance.
A solution for the mass/GND creating I have for such cases (the winding midpoint now isn't longer available by the transformer).
Please, post electronic diagram. It may not be necessary to attach zobel network/boucherot network, but to tell you exactly I need to see electronic diagram.
The developer
http://www.thommessen.com/english/thommessen_frameset2.htm
don't published information like service manual and schematic diagram.
But the topology of the power amp section is very close to that one from the data sheet of the TDA7293V.

BTW - by my own powered subwoofer projects I use always an additional zobel network suitable to the to the particular bass driver(s) mostly 10-30uF and 8R2
 
Last edited:
The most easy solution therefore seems to be the use of only one half winding of the power supply transformer to create a supply voltage of approximately +/- 18V.
Either cut the power or. . . why not double the chip? http://www.diyaudio.com/forums/chip...r-slave-style-no-lossy-emitter-resistors.html That is really easy and supports full power.

P.S.
Generally, supporting 4 ohm speakers with a chip amp needs "Parallel Chip Amp"
Simply double the chip for double-current capacity.
 
Last edited:
Last edited:
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.