I have built numerous 2092 based amplifers and they have worked fine.
I have had the occasional 2092 click very loudly on power off.
Recently I have built up a couple of pcb's and they both give out a loud siren type noise as the power supply dies. Changing the 2092 makes no difference, I tried 6 and the all made a siren noise of power down.
My pcb is the same as the last batch that were ok so I am not sure what I am doing wrong, I have just used the basic datasheet circuit. The only difference is the batch of 2092's are different.
The 2092 is supposed to inhibit the outputs during power down so I should not be getting any signals to the mosfets.
Any input would be appreciated.
I have tried increasing chip reset capacitor and makes no difference.
I tried a bit more decoupling and this made no difference.
I have had the occasional 2092 click very loudly on power off.
Recently I have built up a couple of pcb's and they both give out a loud siren type noise as the power supply dies. Changing the 2092 makes no difference, I tried 6 and the all made a siren noise of power down.
My pcb is the same as the last batch that were ok so I am not sure what I am doing wrong, I have just used the basic datasheet circuit. The only difference is the batch of 2092's are different.
The 2092 is supposed to inhibit the outputs during power down so I should not be getting any signals to the mosfets.
Any input would be appreciated.
I have tried increasing chip reset capacitor and makes no difference.
I tried a bit more decoupling and this made no difference.
Last edited:
IRS2092 noise
Pin 1 must NOT be driven from your main rails which of course hold their voltage for too long after turn off.
Drive this from a smaller supply which discharges immediately on turn off.
CSD pin must be shut off to ground through a small value (100R) resistor in series with a 1N4148 diode.
The cathode of the diode must be held low on turn off and again a simple NPN pull down device can be used.
Steve Mantz
Zed Audio
Pin 1 must NOT be driven from your main rails which of course hold their voltage for too long after turn off.
Drive this from a smaller supply which discharges immediately on turn off.
CSD pin must be shut off to ground through a small value (100R) resistor in series with a 1N4148 diode.
The cathode of the diode must be held low on turn off and again a simple NPN pull down device can be used.
Steve Mantz
Zed Audio
Siren sound from a self oscillating amplifier is usually an indicator of wrong oscillation frequency. When supply voltage is too low, gain becomes low too and another oscillation mode (180 deg crossing) at a lower frequency may appear.
btw: At high power the opposite may happen: A 180 deg crossing at >1Mhz may appear (parasitics not dealt with properly) making the amplifier oscillate at a too high frequency, producing extra heat, and ultimately destroying it.
btw: At high power the opposite may happen: A 180 deg crossing at >1Mhz may appear (parasitics not dealt with properly) making the amplifier oscillate at a too high frequency, producing extra heat, and ultimately destroying it.
I did some experimentation tonight and replaced the 2092's from Future Electronics with a couple from RS components and the rasping is still there.
I looked on the scope and it happens as the supply voltage falls below a certain level.
The only other chnage I made from last time was some different TC4420 gate drivers.
So for interest I took them out and put in a link, the noise went away.
So it looks like something to do with the gate drivers as the supply voltage falls.
I will buy some more gate drivers elsewhere and see if that fixes the problem.
I looked on the scope and it happens as the supply voltage falls below a certain level.
The only other chnage I made from last time was some different TC4420 gate drivers.
So for interest I took them out and put in a link, the noise went away.
So it looks like something to do with the gate drivers as the supply voltage falls.
I will buy some more gate drivers elsewhere and see if that fixes the problem.
Check that the bootstrap capacitor and resistors provide enough power for TC4420, when output is high this IC draws far more current than the IR drivers, see datasheet.
Check that that TC4420 inputs stay low (<.8V) when the IRS2092 shuts down, TC442x chips tend to need pull down resistors at the inputs, they don't include weak pull downs like IR drivers. Be careful, oscilloscope probe may already pull the input high or low if it's floating. The IRS2092 like other IR drivers will probably allow outputs to float to +2V when it's off, not enough to turn on a MOSFET but enough for a TC4420.
I made this mistake in a PFC a few months ago, with Microchip TC4427. It would blow switching transistors randomly at startup, because the outputs of the modern logic gates that were driving the TC4427 were in high impedance state due to undervoltage protection, until +5V finished coming up.
Check that that TC4420 inputs stay low (<.8V) when the IRS2092 shuts down, TC442x chips tend to need pull down resistors at the inputs, they don't include weak pull downs like IR drivers. Be careful, oscilloscope probe may already pull the input high or low if it's floating. The IRS2092 like other IR drivers will probably allow outputs to float to +2V when it's off, not enough to turn on a MOSFET but enough for a TC4420.
I made this mistake in a PFC a few months ago, with Microchip TC4427. It would blow switching transistors randomly at startup, because the outputs of the modern logic gates that were driving the TC4427 were in high impedance state due to undervoltage protection, until +5V finished coming up.
I have used a 100uf capacitor on the lower mosfet power supply and 47uf on the upper supply. The amp runs fine even at very high power. I thin 10uf is suggested as a starting point.
I tried 1k pull downs on the 4420 inputs but it made no difference.
I had a read through 4420 spec and it can have problems with slowly changing edges on the input causing double spikes on the output. It could be as the osc runs down the edges become slower.
I have ordered a different make of 4420 and also a couple of mcp1407's to try.
I bought in a batch of 4420's from Future electronics for this batch of amps and that is when the problem started.
I just use PNP/NPN buffers for class D because the floating driver IC is already doing all level shifting, UVLO, etc. and they don't use extra bootstrap current. For power supplies sometimes I use ICs like TC442x to drive power devices from logic gates (to do a fast 3.3V/5V --> 15V level shifting, which would require a dozen parts otherwise). I don't like to use ICs unless it's really advantageous.
To close this discussion I used some BD243/244 transistors as gate buffers and the rasping noise has gone away.
I recently upgraded my amps to take 2 * 20,000uf and the rasping came back.
IR couldnt help and just suggested adding a power low detect circuit to reset the 2092.
SO I came up with a little PIC circuit that monitors the voltage into the VCC regulator and resets the chip if it falls below 15 volts. Works a treat.
IR couldnt help and just suggested adding a power low detect circuit to reset the 2092.
SO I came up with a little PIC circuit that monitors the voltage into the VCC regulator and resets the chip if it falls below 15 volts. Works a treat.
The 2092 is sensitive to slowly falling pwoer supplies.
I tried an amp tonight with no gate drivers and irfb4019's.
Once the supply capacitance gets up to 20,000uf the power down clicks and noises start to occurr.
I tried an amp tonight with no gate drivers and irfb4019's.
Once the supply capacitance gets up to 20,000uf the power down clicks and noises start to occurr.
- Status
- Not open for further replies.