Hi,
I'm experimenting with three TDA7294 amplifier boards that I've puchased on advice in this forum from this vendor. The PSU is a dual channel bench supply that is nominally doing 30Vdc/3A per channel (closely similar to this one). There is either a voltmeter and an ampmeter per channel. Actually it does about 32 V and 3.2A with all controls at eleven. I also have a function generator, a pair of 8 ohms/400 watts dummy loads, and a Rigol 'scope.
I've attached the chips to 0.8 K/W heatsinks, using the silicon isolation pads that were provided with the kits. No additional silicon grease was used.
At idle each chip draws something between 40 and 50 mA@ +/- 32V.
When I load the chips by 8 ohms and drive by a 1 kHz sine signal, the output level right before the onset of clipping is 59Vpp, which ist about 50 watts@8 ohms. Appears to be serious. The meters in the PSU read 32.3 V/1.14 A for the positive, and 32.0 V/1.25 A for the negative rail, resp. Hence, the total DC power consumption is about 77 watts. All ok, so far.
When I double the load for 4 ohms, something very strange happens: The output signal is clipping, the meters in the PSU read 27.1 V/1.66 to 1.8 A (varying!) for the positive, and 25.9 V/2.14 A for the negative rail.
After I've reduced the input level right before clipping, the output level is 45 Vpp, and the PSU reads 28.8 V/1.6 A for the positive, and 28.0 V/1.5 A for the negative rail.
Some seconds later the output signal is periodically reduced and clips again at about 20 Vpp.
All three chips behave in about the same manner.
Remember that the PSU's current ability is 3.2 A!
What's going on here? What makes my PSU reduce the output voltages, even if it isn't overloaded?
Best regards!
I'm experimenting with three TDA7294 amplifier boards that I've puchased on advice in this forum from this vendor. The PSU is a dual channel bench supply that is nominally doing 30Vdc/3A per channel (closely similar to this one). There is either a voltmeter and an ampmeter per channel. Actually it does about 32 V and 3.2A with all controls at eleven. I also have a function generator, a pair of 8 ohms/400 watts dummy loads, and a Rigol 'scope.
I've attached the chips to 0.8 K/W heatsinks, using the silicon isolation pads that were provided with the kits. No additional silicon grease was used.
At idle each chip draws something between 40 and 50 mA@ +/- 32V.
When I load the chips by 8 ohms and drive by a 1 kHz sine signal, the output level right before the onset of clipping is 59Vpp, which ist about 50 watts@8 ohms. Appears to be serious. The meters in the PSU read 32.3 V/1.14 A for the positive, and 32.0 V/1.25 A for the negative rail, resp. Hence, the total DC power consumption is about 77 watts. All ok, so far.
When I double the load for 4 ohms, something very strange happens: The output signal is clipping, the meters in the PSU read 27.1 V/1.66 to 1.8 A (varying!) for the positive, and 25.9 V/2.14 A for the negative rail.
After I've reduced the input level right before clipping, the output level is 45 Vpp, and the PSU reads 28.8 V/1.6 A for the positive, and 28.0 V/1.5 A for the negative rail.
Some seconds later the output signal is periodically reduced and clips again at about 20 Vpp.
All three chips behave in about the same manner.
Remember that the PSU's current ability is 3.2 A!
What's going on here? What makes my PSU reduce the output voltages, even if it isn't overloaded?
Best regards!
With 4 ohms you are grossly over driving the chip amp.
Its a 100 watt chip amp and into 4 ohms you are currently getting 190 watts !!!!
Its a 100 watt chip amp and into 4 ohms you are currently getting 190 watts !!!!
Sure? Maximum power output is mainly defined by the rail voltages. Would they remain at +/- 32V also if the chips were loaded with 4 ohms, the output power were about 100 watts and the DC current about 2.5 A.
Best regards!
Best regards!
Power out is V squared/ speaker impedance.
If you have 32 volt rails that is 32*32/4 which is 256 watts which is grossly over driving the chip amp.
Your rails are dropping due to this and even with 28 volts rails you are getting 192 watts.
Into 4 ohms to match 100 watts you need no more than 20 volt rails.
Pout is RMS V squared / speaker impedance!
32 volt rails give 64 volt pk-pk; minus bias voltage of output stage.
So not much more than 20 volt RMS is left, which is 100 watt @ 4 ohm.
Datasheet says it is 100 watts "music power" which isn't RMS.
VERY HIGH OPERATING VOLTAGE RANGE
(±40V)
DMOS POWER STAGE
HIGH OUTPUT POWER (UP TO 100W MUSIC
POWER)
MUTING/STAND-BY FUNCTIONS
NO SWITCH ON/OFF NOISE
NO BOUCHEROT CELLS
VERY LOW DISTORTION
VERY LOW NOISE
SHORT CIRCUIT PROTECTION
THERMAL SHUTDOWN
Power is defined with RMS voltage, so nigelwright7557 calculated wrong power, but was right.
TDA7294 can deliver about 70W on 4 ohm load with +/-27V and limited current (which obviously occurred on the test after heating up of the chip).
with 10% thd can deliver about 100W on +/-29V with 4 ohm load.
TDA7294 can deliver about 70W on 4 ohm load with +/-27V and limited current (which obviously occurred on the test after heating up of the chip).
with 10% thd can deliver about 100W on +/-29V with 4 ohm load.
So the chip »senses« for 4 ohms loads and decreases the rail voltages of a regulated (!) power supply? How does it manage this?
Best regards!
Best regards!
somehow it happens 🙂
Overload in current (in chip sense circuit) can make strange behaviour and probably cause some strange peaks in current and measuring current instrument can't show fast changes but current limit of the power supply will decrease voltage (maybe).
Try measuring with some fast instrument (scope) maybe it will be seen.
Overload in current (in chip sense circuit) can make strange behaviour and probably cause some strange peaks in current and measuring current instrument can't show fast changes but current limit of the power supply will decrease voltage (maybe).
Try measuring with some fast instrument (scope) maybe it will be seen.
I've also observed that paralleling the onboard rail buffer 'lytics (which really aren't that big at 100 µF...) by 6.8 mF elcos increases the 4 ohms output power and the rail voltages under this condition significantly. So can I suspect the bench supply's current regulation/protection being a bit too fast?
Best fegards!
Best fegards!
maybe,
to test my theory, put small resistor in series with +/- supply and measure current with scope (hope you understand how to measure on shunt).
I had some strange behaviour on linear power supply when current limiting was reducing voltage, but avearge current was within allowed current.
current peak and shape was the cause (current limit was 2A in my case).
Best regards
to test my theory, put small resistor in series with +/- supply and measure current with scope (hope you understand how to measure on shunt).
I had some strange behaviour on linear power supply when current limiting was reducing voltage, but avearge current was within allowed current.
current peak and shape was the cause (current limit was 2A in my case).
Best regards
I am right for "music power" which is what the chip is rated at, 100 watts music power.
Instead of using RMS value you just use the peak voltage value.
Its just a way making amp look like it is more powerful than it is.
please check page 10 from this datasheet https://www.st.com/resource/en/datasheet/cd00000017.pdf
"From fig. 20, where the maximum power is around 200 W, we get an average of 20 W, in this condition, for a class AB amplifier the average power dissipation is equal to 65 W. The typical junction-to-case thermal resistance of the TDA7294 is 1 o C/W (max= 1.5 o C/W). To avoid that, in worst case conditions, the chip temperature exceedes 150 o C, the thermal resistance of the heatsink must be 0.038 o C/W (@ max ambient temperature of 50 o C). As the above value is pratically unreachable; a high efficiency system is needed in those cases where the continuous RMS output power is higher than 50-60 W."
what I understand from the text above is tda7294 only suitable for 50-60W RMS
"From fig. 20, where the maximum power is around 200 W, we get an average of 20 W, in this condition, for a class AB amplifier the average power dissipation is equal to 65 W. The typical junction-to-case thermal resistance of the TDA7294 is 1 o C/W (max= 1.5 o C/W). To avoid that, in worst case conditions, the chip temperature exceedes 150 o C, the thermal resistance of the heatsink must be 0.038 o C/W (@ max ambient temperature of 50 o C). As the above value is pratically unreachable; a high efficiency system is needed in those cases where the continuous RMS output power is higher than 50-60 W."
what I understand from the text above is tda7294 only suitable for 50-60W RMS
How hot is the chip??
This chip claims both SOA and thermal protection, no data given. But typically a protection system will cut-out in some way; there are many variations. It may current-limit, flat-line. It may current limit with a limit that drops with temperature. National had good details of their SpiKe protection; ST is less revealing.
I'm sure you are asking too much of one tiny grain of Silicon.
You're right, the case feels much (!) hotter than the heatsink.
Still I don't know how the chip can tell my bench supply to lower it's output voltages without exceeding the PSU's current capability.
Best regards!
Still I don't know how the chip can tell my bench supply to lower it's output voltages without exceeding the PSU's current capability.
Best regards!
It doesn't. You choose the voltage of the PSU depending on the speakers you want to use.
I would expect the cause of the "strange" behaviour is the PSU is reaching its current limit.
I would expect the cause of the "strange" behaviour is the PSU is reaching its current limit.
I am wondering if the power supplies are not SMPSU types. If that is the case, you may have ‘rail pumping’ aka ‘bus pumping’. This is normally associated with class D amps, but SMPSU can also suffer from this issue in some cases when used to power a linear amplifier.
In that case, wire one of the amps in anti phase (I.e. input signal inverted wrt the other) and retry your test. If the problem is resolved, it’s probably a rail/bus pumping issue and will not show up with a standard linear power supply.
In that case, wire one of the amps in anti phase (I.e. input signal inverted wrt the other) and retry your test. If the problem is resolved, it’s probably a rail/bus pumping issue and will not show up with a standard linear power supply.
Kay Kirinha said in the first post that PSU is bench power supply with dual voltage output and curent limiting, equiped with measuring instruments...
Yes, and it is 20 years old, and it is quite an old school linear PSU that economizes by relay switching the PT secondaries' taps.
Best regards!
Best regards!
In the first post you stated that the output is 45Vpp with a 4 ohm load. This means that the output current is 5.625A peak. Could this be activating the current limiter of the PSU?
- Status
- Not open for further replies.