I got a trafo which is rated at 22-0-22 and its designed by a industrial guy. He said that on load it will deliver the rated values. The trafo is EI and rated at 9Amps and almost weighs 10kg. I have opa549 and I have this good trafo at this moment. After rectification I will be getting 31.1 V may be on over or under voltage fluctuations Its possible to get something 32 at max. So I doubt that is OPA549 stable at this voltage? what is the maximum limit.
what if I drive the opamp in 4 ohms imp? what should be the supply voltage to drive 4 ohms drivers?
what if I drive the opamp in 4 ohms imp? what should be the supply voltage to drive 4 ohms drivers?
apply mains power to your transformer. Measure the input voltage and the output voltage.
It is a small step to then determine the worst case (highest mains voltage & low load current) smoothing capacitor voltage.
Compare that worst case voltage to the 549 specification.
BTW,
this is not stability. This is overvoltage to the power pins.
It is a small step to then determine the worst case (highest mains voltage & low load current) smoothing capacitor voltage.
Compare that worst case voltage to the 549 specification.
BTW,
this is not stability. This is overvoltage to the power pins.
Max Voltage on a 549 is +/-30v so this will be over spec. See datasheet http://www.ti.com/lit/ds/symlink/opa549.pdf.
Is the weather hot by you? The line Voltage varies with the weather (AC load). At our house I've seen line Voltage as low as 99 though it's typically 121 or so. Your 31.2 is already over the 'absolute maximum' recommended by TI. If your line Voltage goes up it will be even worse. While you can often exceed the manufacturers ratings you do so at your own risk.
Haven't you downloaded the data sheet? Burr Brown's sheets were always very good as were Motorola's (now ON Semiconductor) and Analog Devices (Excellent). The 'applications information' is well worth the time to read and understand. Any examples of how to use it are always helpful.
Your amplifier into a 4 ohm load with +/- 30 Volt supplies will need 3-4 Amps PER CHANNEL to run continuously at full power with a sinewave. Your 9 Amp will be fine for music even at 4 ohms though temporary peaks up to 16 Amps but the average will be within the transformer rating. The nice thing about larger transformers is that they have a lower output resistance which suffers less droop under load. Have you considered toroidal transformers? Antek makes some pretty fine units at very reasonable prices. They take less space, are less lossy and weigh less than conventional EI core units.
G²
Maximum ratings are just that. Stability doesn't come into it, you are exceeding the rated limits of the devices and substrate die.
A simple regulator/ripple filter could easily be made to loose a few volts but you always have the issue of the heat generated under full load. Example, 5 volts across the series pass device at 5 amps is 25 extra watts to get rid of.
A simple regulator/ripple filter could easily be made to loose a few volts but you always have the issue of the heat generated under full load. Example, 5 volts across the series pass device at 5 amps is 25 extra watts to get rid of.
Understanding Class-D amplifier power supply requirements
Article is for Class D, but the maths are still appropriate.
As for the amp, I'd just build a Sysmasym or BabyAKSA. Both very simple and will do exactly what you want without all the tedious mucking around with regulated supplies.
Yes, the "method" is capacitance multiplier made with HexMosfets - it will "eat" 4V of each rail and lower the ripple voltage by about 50dB
Attachments
You can use BJTs in a capaciptor multiplier and by adjusting the resistor values trim the output voltage to a fixed %age of the input voltage.
eg add a 10k in parallel to C9.
Change R9 to 1k.
The voltage on C9 is now 10/11ths of the input voltage.
The output voltage will be ~600mV below that 10/11ths voltage on C9.
Fopr BJTs I would omit R10.
eg add a 10k in parallel to C9.
Change R9 to 1k.
The voltage on C9 is now 10/11ths of the input voltage.
The output voltage will be ~600mV below that 10/11ths voltage on C9.
Fopr BJTs I would omit R10.
The ripple filters won't get AS hot as the power chip but they will need substantial heatsinks. And yes, they could share with the amplifier.
G²
I really don't follow your logic on this one...
It's like saying the heatsinks on a 100watt rms class B amp can be kept really small because the average power on music is low.
No,
a 100W amplifier that has a quiescent current of just 20mA and +-50Vdc supplies will have a Pq of 2W. That alone would need a tiny heatsink.
Now play some music at -20dB ref maximum power, average level. The output stage will have an average dissipation of maybe 5W to 10W. The heatsink must be able to keep the amp cool for this.
If very compressed music is played at -10dB average level then the sink will have to dissipate maybe 10W to 20W and still play properly without overheating. In the low Pq amplifier , the music playing dissipation determines the sink size.
Now compare that to a high quiescent current amplifier say 200mA with Pq = 20W.
The sink must be bigger.
Now consider music playing. The dissipation hardly changes even for -10dB average levels. The same sink that copes with Pq also copes with music reproduction. In the High Pq amplifier the music playing dissipation can be very similar to the Pq dissipation.
Let's return to the cap multiplier.
The output voltage tracks the input voltage.
The Vdrop of the multiplier averages 4V (for this particular case).
The quiescent draw of the chip amp is ~ 40mA
When playing music on +-25Vdc supply rails the average current may rise to approximately 100mA (5W dissipation for 0.5W of music output) . In this situation the cap multiplier average dissipation is 4V times 100mA. Increase the output to 5W (-10dB) average then the current may go up a further 50% to 150mA. Or reduce -30dB and it drops to ~80mA.If the music is turned down to -30dB the average.
For all these audio signal situations the cap multiplier sees an average current draw of 40mA to 150mA.
The sink does not need to be big.
Even at 150mA for a 5W of average music power the dissipation is only 0.6W. A 20C/W sink will rise by ~ 20 *0.6*1.5 ~18C.
That is not a problem.
If a high current device is being used then the Rth j-a could well be low enough that it can be used to power a chipamp without any heatsink.
a 100W amplifier that has a quiescent current of just 20mA and +-50Vdc supplies will have a Pq of 2W. That alone would need a tiny heatsink.
Now play some music at -20dB ref maximum power, average level. The output stage will have an average dissipation of maybe 5W to 10W. The heatsink must be able to keep the amp cool for this.
If very compressed music is played at -10dB average level then the sink will have to dissipate maybe 10W to 20W and still play properly without overheating. In the low Pq amplifier , the music playing dissipation determines the sink size.
Now compare that to a high quiescent current amplifier say 200mA with Pq = 20W.
The sink must be bigger.
Now consider music playing. The dissipation hardly changes even for -10dB average levels. The same sink that copes with Pq also copes with music reproduction. In the High Pq amplifier the music playing dissipation can be very similar to the Pq dissipation.
Let's return to the cap multiplier.
The output voltage tracks the input voltage.
The Vdrop of the multiplier averages 4V (for this particular case).
The quiescent draw of the chip amp is ~ 40mA
When playing music on +-25Vdc supply rails the average current may rise to approximately 100mA (5W dissipation for 0.5W of music output) . In this situation the cap multiplier average dissipation is 4V times 100mA. Increase the output to 5W (-10dB) average then the current may go up a further 50% to 150mA. Or reduce -30dB and it drops to ~80mA.If the music is turned down to -30dB the average.
For all these audio signal situations the cap multiplier sees an average current draw of 40mA to 150mA.
The sink does not need to be big.
Even at 150mA for a 5W of average music power the dissipation is only 0.6W. A 20C/W sink will rise by ~ 20 *0.6*1.5 ~18C.
That is not a problem.
If a high current device is being used then the Rth j-a could well be low enough that it can be used to power a chipamp without any heatsink.
Last edited:
What size of cap multiplier sink is required for continuous power testing of the amplifier?
+- 50W into 8r0 is equivalent to 2.5Aac into the test load.
the supply rail sees a duty cycle of ~50%.
The cap multiplier now sees 4V & 1.25Aac, or about 5W during maximum power testing.
Do you/we design the amplifier heatsink and the cap multiplier heatsink for maximum power testing on a continuous duty?
I don't !!!!
+- 50W into 8r0 is equivalent to 2.5Aac into the test load.
the supply rail sees a duty cycle of ~50%.
The cap multiplier now sees 4V & 1.25Aac, or about 5W during maximum power testing.
Do you/we design the amplifier heatsink and the cap multiplier heatsink for maximum power testing on a continuous duty?
I don't !!!!
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.