In general, the question of why high-power amplifiers of 4 kW-20 kW, all built on the same base, have been worrying for a long time.
To these amplifiers I attribute:
Scar Audio 4500
GROUND ZERO GZPA 1.10K.HCX and its clone
ALPHARD AUDIO EXTREME GR12000.1D
B2M1
American Bass and others using drivers with microcircuits Ir21844 4pcs on one board
Theory One:
First of all, I would like to know how the operating frequency can affect the correct operation of the amplifier?
How to determine the correct operating frequency for these amplifiers?
For example:
-We have a driver running at 150kHz carrier frequency,
-There is a four-phase output stage and 32 transistors on it. In each phase we have 8transistors (4 transistors in each arm).
I often repair amplifiers of similar classes and quite often there is a different frequency - 120 kHz, 130 kHz, 135 kHz, and on Sundaun Audio amplifiers there is generally a working frequency of 155 kHz.
I would like to understand how to choose the working frequency correctly, what is the difference between 120 kHz and 150 kHz? Understandably, the higher the frequency, the lower the heating, but in addition to the high frequency, we basically do not need it as this is a monoblock and we only remove the low frequency at the output.
With a low working frequency, the error of self-healing of the control stage of controlling the output stage of the risk is less or I am wrong.
The second theory:
The second theory is that when heated, the transistors change their characteristics and possibly the transistors' gates become heavier and therefore the IR21844 chips simply do not withstand and burn, I also think that with such a load as 4 transistors on the LO and 4 transistors on the HI line of IC without buffer (auxiliary transistors) for the microcircuit is rather heavy and it also heats up and also changes some of its characteristics, and also especially noted that the ceramic capacitances and resistors of the driver located on the motherboard when heated are also quite s They change their characteristics and this can also affect the failure of the amplifier.
What can be done for more stable work?
How to choose the working frequency correctly, provided that the Sandown Audio amplifiers operate at a working frequency of 155 kHz more stable than other amplifiers ..
I wrote a few days ago in one of the posts about the SD line and I would like to add, maybe try in the sd line to replace the condetator with a larger capacity or put a resistor for more. This will give us a kind of delay in switching on the amplifier, because maybe when a person starts to listen very loudly the music from his car’s on-board power supply drops out below 12v and the amplifier turns off at some point and turns on again getting a big jump and at that moment it burns?
Any thoughts on anyone else?
To these amplifiers I attribute:
Scar Audio 4500
GROUND ZERO GZPA 1.10K.HCX and its clone
ALPHARD AUDIO EXTREME GR12000.1D
B2M1
American Bass and others using drivers with microcircuits Ir21844 4pcs on one board
Theory One:
First of all, I would like to know how the operating frequency can affect the correct operation of the amplifier?
How to determine the correct operating frequency for these amplifiers?
For example:
-We have a driver running at 150kHz carrier frequency,
-There is a four-phase output stage and 32 transistors on it. In each phase we have 8transistors (4 transistors in each arm).
I often repair amplifiers of similar classes and quite often there is a different frequency - 120 kHz, 130 kHz, 135 kHz, and on Sundaun Audio amplifiers there is generally a working frequency of 155 kHz.
I would like to understand how to choose the working frequency correctly, what is the difference between 120 kHz and 150 kHz? Understandably, the higher the frequency, the lower the heating, but in addition to the high frequency, we basically do not need it as this is a monoblock and we only remove the low frequency at the output.
With a low working frequency, the error of self-healing of the control stage of controlling the output stage of the risk is less or I am wrong.
The second theory:
The second theory is that when heated, the transistors change their characteristics and possibly the transistors' gates become heavier and therefore the IR21844 chips simply do not withstand and burn, I also think that with such a load as 4 transistors on the LO and 4 transistors on the HI line of IC without buffer (auxiliary transistors) for the microcircuit is rather heavy and it also heats up and also changes some of its characteristics, and also especially noted that the ceramic capacitances and resistors of the driver located on the motherboard when heated are also quite s They change their characteristics and this can also affect the failure of the amplifier.
What can be done for more stable work?
How to choose the working frequency correctly, provided that the Sandown Audio amplifiers operate at a working frequency of 155 kHz more stable than other amplifiers ..
I wrote a few days ago in one of the posts about the SD line and I would like to add, maybe try in the sd line to replace the condetator with a larger capacity or put a resistor for more. This will give us a kind of delay in switching on the amplifier, because maybe when a person starts to listen very loudly the music from his car’s on-board power supply drops out below 12v and the amplifier turns off at some point and turns on again getting a big jump and at that moment it burns?
Any thoughts on anyone else?
The frequency isn't the same for any self-oscillating amp and the frequency often changes greatly at the amp is driven harder. I don't check them often but I remember some being about 130kHz at idle and dropping to about 80kHz at higher power.
The components in the feedback are the main determining factor for the frequency.
Include the terms 'spread spectrum' and 'sigma delta' when searching for class D information for more on this.
The Class D forum on this site could probably give much finer/definitive explanation on this but if you mention that it's for car audio, they will likely move the thread back to here.
The components in the feedback are the main determining factor for the frequency.
Include the terms 'spread spectrum' and 'sigma delta' when searching for class D information for more on this.
The Class D forum on this site could probably give much finer/definitive explanation on this but if you mention that it's for car audio, they will likely move the thread back to here.
That is the power supply 380v = 190v per rail which is huge power.
all those big caps are reservoir caps for power peaks, the 8 torroids are for power conversion. This is a switching regulated supply which like class D amps uses a frequency of 75 to 150 Khz to drive the torroids.
Most Class D amps, when they blow , destroy the high and low side driver chips as well. These will need to be replaced.
all those big caps are reservoir caps for power peaks, the 8 torroids are for power conversion. This is a switching regulated supply which like class D amps uses a frequency of 75 to 150 Khz to drive the torroids.
Most Class D amps, when they blow , destroy the high and low side driver chips as well. These will need to be replaced.
Last edited:
- dear friend, I have no questions to repair, I have a direct question - What should be done in order for the amplifiers to become more reliable ...
Thanks Perry!
Thanks to you, I learned how to repair amplifiers of any complexity. I had a great desire, I studied a lot of literature and unfortunately not everything and not always can be found in books. Often everything comes with experience, and experience comes from practice. I am an intern and knowing your knowledge I turned to you for more help. What would you recommend for improving the amplifiers, so that they work more reliably? Does it make sense to experiment with DeadTime? As I understand it, this is the delay of the amplifier on.
I met an EDGE 10000 amplifier and its amplifier turns on 10 seconds after turning on the power supply ..
TNks
The deadtime is the time for the Fets to be switch on and off. The time between high side go off on low side switch on. If this time to short, you will possibly make a shortbreak and the current in the Fets rises until the blow. Small amount should be ok. If the deadtime is too long, you lose effectivity in a SMPS. If the deadtime in an Audiobank is to high, THD starts rising. If its too low, the Fets burning.
The main reason for class d amps failing is incorrect deadtime.
If it is too small or none existent you get shoot through (short circuit) which takes high currents.
In my experience when they blow they take out mosfets and driver circuit and class d IC too.
What if we take and increase this line of inclusion by increasing the nominal value of the resistor from 4.7K to 10K or of the capacitor from 100uf to 220uf?
After switching on AMP we have a charge of capacitors and it takes time. By increasing the element ratings we get a longer run of the amplifier or am I wrong? would that make the launch stable ?
Thanks
Last edited:
Looking at an amp I have here, it appears that the switching from SD mode to enabled is determined by the capacitor on pin 2 of the 21844 and the internal pullup resistor (maybe, actually a constant current source) in the 21844. I can't say that the circuit above can't make a difference but I think it's unlikely to be critical.
The ICs use a fairly high value SMD capacitor between the SD pin and the negative rail. Increasing it would make switching from SD to enabled, slower.
I tried removing the cap from SD to the rail and switching the SD line very quickly (by raking a wire onto the negative rail). It survived a bit of this but the IC failed twice. The failure was simply a failure to switch on. I don't know if this failure could be a problem in larger amps that could cause catastrophic failure. With the cap in place and within tolerance, I don't think you could make the IC fail because the switching time would be much slower.
The ICs use a fairly high value SMD capacitor between the SD pin and the negative rail. Increasing it would make switching from SD to enabled, slower.
I tried removing the cap from SD to the rail and switching the SD line very quickly (by raking a wire onto the negative rail). It survived a bit of this but the IC failed twice. The failure was simply a failure to switch on. I don't know if this failure could be a problem in larger amps that could cause catastrophic failure. With the cap in place and within tolerance, I don't think you could make the IC fail because the switching time would be much slower.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.