Unnecessarily complicated perhaps. Where will you get the 5V from? The segment of circuit I put up, with a series input resistor of 470K or so, as suggested by others, will work with any amp supply and keep the component count low.
Using TDA7293 clip detector pin
Hi guys, I just made and tested a simple circuit designed to use the clip pin 5 of TDA7293. The circuit uses just 2 transistors and can be used at any supply voltage. It not requires a stabilizer. I dimensioned the components to be able to detect short clips holding the LED on for more than 0.5s.
+V is the positive supply voltage of TDA7292, GND is the amp supply GND. If +V is more than 40Vcc, increase the 82K resistor to be the double of voltage. (i.e. for +V = 45V 90Kohm). LEDs can be shorted with no damage to the circuit.
If you use a parallel or/and bridge configuration, tie together pins 5 of all TDA7293 used by same audio channel.
LEDs can be mounted on the front pannel of the amplifier
Hi guys, I just made and tested a simple circuit designed to use the clip pin 5 of TDA7293. The circuit uses just 2 transistors and can be used at any supply voltage. It not requires a stabilizer. I dimensioned the components to be able to detect short clips holding the LED on for more than 0.5s.
+V is the positive supply voltage of TDA7292, GND is the amp supply GND. If +V is more than 40Vcc, increase the 82K resistor to be the double of voltage. (i.e. for +V = 45V 90Kohm). LEDs can be shorted with no damage to the circuit.
If you use a parallel or/and bridge configuration, tie together pins 5 of all TDA7293 used by same audio channel.
LEDs can be mounted on the front pannel of the amplifier
If you can do it with an LED then you can do it with a jfet or LDR to reduce the amplitude of the clips. Exception: You'll need to add a cap, such as 10u to the power-side of the jfet or LDR so as to slow it down enough to avoid modulating the treble excessively.
Result: Volumax style headroom boost. Its more useful than flashing indicators.
Result: Volumax style headroom boost. Its more useful than flashing indicators.
You are right Daniel, if brute undistorted power on output is the goal. But sound "purists" don't accept anything else than "pure" sound, keeping the dynamic of the source untouched. Just keeping the LEDs below flashing at louder passages of music, the amp will reproduce the dynamic of the source with fidelity. Of course, at the expense of not delivering the max power continuously.
No sir. At most a variance of 2db at the speaker AFTER clip detector activation, and AFTER filling a delay cap so as to determine that clip detection is desirable and needed to reduce the duration of clipping.
This is not brute! Assuming the indicators have worked passably or in any way relevant: After clip detection LED has lit up solid, mild mitigation should not be worse!!!
I did mention to do something more useful than the indicator light. If we were to confine assumptions to only that much, it may be more clear that I wanted a slightly effective delayed slow compressor that cannot do anything when the clipping indicator LED's are off.
Can we please confine our assumptions to that which may be useful?
For example, a little emergency de-clip function is useful to a rather short extent.
Logically, if modifying the signal while the error indication lamps are lit is not a good thing, then the indication circuit is critically defective. Please determine which is the case.
A mitigation circuit should never exceed the capacity of its error detection circuit, and I have no idea why anyone would assume otherwise.
Doing a little better would be far closer to perfect. I'll bet that sounds good too.
In any case, I'm just saying that if the clip indication pin is as unreliable as you have just described, then we won't and couldn't be needing an indicator lamp for it. Please clarify?
This is not brute! Assuming the indicators have worked passably or in any way relevant: After clip detection LED has lit up solid, mild mitigation should not be worse!!!
I did mention to do something more useful than the indicator light. If we were to confine assumptions to only that much, it may be more clear that I wanted a slightly effective delayed slow compressor that cannot do anything when the clipping indicator LED's are off.
Can we please confine our assumptions to that which may be useful?
For example, a little emergency de-clip function is useful to a rather short extent.
Logically, if modifying the signal while the error indication lamps are lit is not a good thing, then the indication circuit is critically defective. Please determine which is the case.
A mitigation circuit should never exceed the capacity of its error detection circuit, and I have no idea why anyone would assume otherwise.
Doing a little better would be far closer to perfect. I'll bet that sounds good too.
In any case, I'm just saying that if the clip indication pin is as unreliable as you have just described, then we won't and couldn't be needing an indicator lamp for it. Please clarify?
Last edited:
I can't say that the clip pin is unreliable, but for sure when a logic "0" is output by this pin, the signal was already distorted. Pin 5 signal is a consequence of clipping not a prediction. When clip LED flashes "damage" of sound was done, what you can do is just reduce the volume to prevent clipping on the next loud passage.
So, after clip detector indicator has lit up, there would be a good time, after tiny delay, to have a jfet automatically decrease the input volume a little bit.
Yes, Daniel. You are absolutely correct. If "audio purists" don't want signal limiting they should never listen to an amplifier that is clipping anyway🙂 Indeed, as you stated so succinctly, once the clip detector is active the "damage" is done.
So... why not take corrective action? In fact, I have been employing this technique for many years in my 7293 amp designs.
I experimented with several different arrangements including FET, modulating the mute pin, and LDR.
The winner: NSL-32 from Silonex. (not NSL-32R or R3)
The characteristics of this particular LDR are just about right. It has a fast attack and slow decay... no "pumping" effect is audible.
I drive it straight from the pin 5 (can handle 20ma) with a series resistor and pot to trim for best sound.
In fact, by using this trick along with mixed mode (current AND voltage) feedback to create high output impedance (about 5 ohms) I have a very tube like character overall.
So... why not take corrective action? In fact, I have been employing this technique for many years in my 7293 amp designs.
I experimented with several different arrangements including FET, modulating the mute pin, and LDR.
The winner: NSL-32 from Silonex. (not NSL-32R or R3)
The characteristics of this particular LDR are just about right. It has a fast attack and slow decay... no "pumping" effect is audible.
I drive it straight from the pin 5 (can handle 20ma) with a series resistor and pot to trim for best sound.
In fact, by using this trick along with mixed mode (current AND voltage) feedback to create high output impedance (about 5 ohms) I have a very tube like character overall.
@spare_parts Can you post a schematic?
Did you try also VTL5C2 ? Datasheet data looks very similar to NSL-32 including decay time. Also I'm curios to know if you tried LCR0202/3 more affordable but datasheet say anything about decay time.
Did you try also VTL5C2 ? Datasheet data looks very similar to NSL-32 including decay time. Also I'm curios to know if you tried LCR0202/3 more affordable but datasheet say anything about decay time.
@cmorariu:
I never bothered to draw one 🙂
The resistive element of LDR simply goes in parallel with the input resisor divider leg to ground. Only addition other than that is some series resistance before the divider (start around 1-2k)
I like the NSL-32; I did try the VTL5C2 and it is workable... see if you like it. Try a "custom" LDR tuned for purpose by choosing an LED with a light wavelength peak output to match a particular CDS cell and pot the assembly in epoxy. This circuit is more art and tuning than math, especially sicne the LDR manufacturing has sloppy tolerance around 20% in my experience. A pot in series with the LED allows trimming of the response. The attenuation really digs in at around 10ma through the LED.
I never bothered to draw one 🙂
The resistive element of LDR simply goes in parallel with the input resisor divider leg to ground. Only addition other than that is some series resistance before the divider (start around 1-2k)
I like the NSL-32; I did try the VTL5C2 and it is workable... see if you like it. Try a "custom" LDR tuned for purpose by choosing an LED with a light wavelength peak output to match a particular CDS cell and pot the assembly in epoxy. This circuit is more art and tuning than math, especially sicne the LDR manufacturing has sloppy tolerance around 20% in my experience. A pot in series with the LED allows trimming of the response. The attenuation really digs in at around 10ma through the LED.
Last edited:
@spare_parts
Thank you for the details, I wait for the NSLs and LCRs to come and I will give it a try.
Thank you for the details, I wait for the NSLs and LCRs to come and I will give it a try.
Inbuilt compressor/limiter feature?
A resistor series to pin 3 (pass) and a resistor series to pin 4 (shunt), could form a voltage divider during mute.
After that is figuring out how to drive the mute feature from the clip detect feature--Try to "mute on clip". How do we hook that up?
A resistor series to pin 3 (pass) and a resistor series to pin 4 (shunt), could form a voltage divider during mute.
After that is figuring out how to drive the mute feature from the clip detect feature--Try to "mute on clip". How do we hook that up?