I tried driving the inputs with 2v and gain up a bit. Didn't cause it to start up.
I put shorts across all 4 output, NO input signal. Did not start up.
After a couple times of hitting the remote, the amp does turn fully on and channels work. If I drain the bootstrap caps, it will then again not fire up for a couple attempts.
I'm not sure yet how the clock gen ties into the outputs. I'd guess it's gated or some other enable is being held low unless the bootstrap caps are charged and that section comes up faster.
I had similar thoughts on one channel being the culprit.. I tried just discharging one or two bootstrap cap without seeing anything repeatable.
I put shorts across all 4 output, NO input signal. Did not start up.
After a couple times of hitting the remote, the amp does turn fully on and channels work. If I drain the bootstrap caps, it will then again not fire up for a couple attempts.
I'm not sure yet how the clock gen ties into the outputs. I'd guess it's gated or some other enable is being held low unless the bootstrap caps are charged and that section comes up faster.
I had similar thoughts on one channel being the culprit.. I tried just discharging one or two bootstrap cap without seeing anything repeatable.
I sent Focal a message seeing if there is maybe a known issue with a mod that fixes it.
I'm not holding my breath on getting much info 😉.
I'm not holding my breath on getting much info 😉.
Last edited:
This is a self oscillating design with a master clock to sync all channels under low power conditions. Once the signal level increases the clock is disabled from syncing the 4 channels. The sync is to prevent "squealing" in the speakers owing to the fact that the channels are so close together and there was no space to put rail and Vb filters in this design. Focal demanded that the amp be as small as possible, so I had to leave parts out.
I have no idea what "bias" capacitors are as there is no bias circuit in these amplifiers.
The rail voltage starts at about +/-26v and will ramp up when the signal is high enough to prevent clipping.
I have repaired these amps for nearly 10 years and have not seen this issue which you describe.
Steve Mantz
Zed Audio Corp
smm@zedaudiocorp.com
I have no idea what "bias" capacitors are as there is no bias circuit in these amplifiers.
The rail voltage starts at about +/-26v and will ramp up when the signal is high enough to prevent clipping.
I have repaired these amps for nearly 10 years and have not seen this issue which you describe.
Steve Mantz
Zed Audio Corp
smm@zedaudiocorp.com
The clock has nothing to do with start up. Each channel has a start up series diode Dx01 and Rx23 where "x" is 3 in channel 1 and 4 in ch 2 etc.
Perry I think you may have had it with the -rail+12v supply.
I was probing some more there and it's rise time did seem a bit inconsistent.
It only has a couple ceramic caps on the reg. I reflowed them both well (not that either had a cracked joint) and so far the amp seems to be behaving.
It's my understanding that reflow is good for ceramics and can help realign the structure etc..
I'll likely end up replacing them and do more digging later. But seems like progress.
Thanks Perry and Steve.
I was probing some more there and it's rise time did seem a bit inconsistent.
It only has a couple ceramic caps on the reg. I reflowed them both well (not that either had a cracked joint) and so far the amp seems to be behaving.
It's my understanding that reflow is good for ceramics and can help realign the structure etc..
I'll likely end up replacing them and do more digging later. But seems like progress.
Thanks Perry and Steve.
I've never heard that for ceramic capacitors.
SMD ceramics have a higher failure rate than through hole ceramics. Their failure (cracking that you can't see until solder is molten) isn't generally recognized unless it's a critical application. I saw quite a few in timing circuits.
Replacing them would be good if you know the value. SMD ceramics aren't generally marked. MOER may be able to help with that as well.
Before you let it leave the shop, I'd use freezing, heating, flexing the board and moving various components to rule out other possible intermittent problems.
SMD ceramics have a higher failure rate than through hole ceramics. Their failure (cracking that you can't see until solder is molten) isn't generally recognized unless it's a critical application. I saw quite a few in timing circuits.
Replacing them would be good if you know the value. SMD ceramics aren't generally marked. MOER may be able to help with that as well.
Before you let it leave the shop, I'd use freezing, heating, flexing the board and moving various components to rule out other possible intermittent problems.
Sounds good, it could certainly have been a micro crack in the solder.
Yeah I plan on beating on it for awhile here.
Thanks
Yeah I plan on beating on it for awhile here.
Thanks
Seems that -vcc+12v supply goes to some control circuitry down the line and is variable.
It's about -20v to the power amps when working and about -28v when not.
There is some mechanical element to the fail but I haven't pinned it down. Can put pressure in the middle of PCB to get things working sometimes.
It's about -20v to the power amps when working and about -28v when not.
There is some mechanical element to the fail but I haven't pinned it down. Can put pressure in the middle of PCB to get things working sometimes.
Is it variable when measured directly across the 12v supply? It may vary with reference to ground but I'd expect it to be a steady 12v when using the negative rail as the reference.
What is it using as a regulator? 7812? Something more complex?
What is it using as a regulator? 7812? Something more complex?
It's a 7812 yes, I replaced it too.
No, sorry. It's not variable across the supply. Wish it was that simple.
The low side driver circuitry has a voltage on many parts at around -20v when working and at -28v when not.
I suppose not directly related to the -vcc+12v supply maybe.. but something in that vicinity.
No, sorry. It's not variable across the supply. Wish it was that simple.
The low side driver circuitry has a voltage on many parts at around -20v when working and at -28v when not.
I suppose not directly related to the -vcc+12v supply maybe.. but something in that vicinity.
It was stated that the clock had nothing to do with the startup but, as with most all clocked amps, the first clock cycle drives the low-side FETs on and charges the high-side bootstrap supply.
Is charging the bootstrap capacitors an issue (measure directly across them)? What is the voltage you read when the amp doesn't start?
The clock is making it to the outputs sometimes because you had the same frequency on all outputs earlier. Can you determine why it's not getting through the rest of the time?
Is charging the bootstrap capacitors an issue (measure directly across them)? What is the voltage you read when the amp doesn't start?
The clock is making it to the outputs sometimes because you had the same frequency on all outputs earlier. Can you determine why it's not getting through the rest of the time?
When malfunctioning, the bootstrap caps charge to about 13v where the zeners clamp. 13v zeners in this design. They sit at around 10-11v when amp is working.
I believe the low side driver supply is the issue. Is around -28v when malfunctioning and -20v when working.
There is some control circuitry there that I don't have my head around yet.
I believe the low side driver supply is the issue. Is around -28v when malfunctioning and -20v when working.
There is some control circuitry there that I don't have my head around yet.
All what I have read is wrong.
The floating +12v supply is driven from an independent winding on the transformer with a half wave diode to a 220mfd 25v cap. The 7812 regulator has 0.1mfd 50v ceramics on it's input and output and this circuit has been used on every class D I have designed whether my discrete design or using a chip like the IRS20957 or Si8231.
Because this amplifier starts up in it's low supply condition only means that the 7812 floats on a lower supply than when the amplifier is driven to it's high supply value of +/-44v.
The class D has no knowledge of the main rails, all it cares about is that the floating supply has >10v and that the bootstrap capacitor is pre charged by12k + diode.
The clock has zero to do with the start up procedure.
This is NOT a clocked amplifier, each channel is a self oscillator.
"Seems that -vcc+12v supply goes to some control circuitry down the line and is variable."
There is no control circuitry on the floating 12v supply.
The floating +12v supply is driven from an independent winding on the transformer with a half wave diode to a 220mfd 25v cap. The 7812 regulator has 0.1mfd 50v ceramics on it's input and output and this circuit has been used on every class D I have designed whether my discrete design or using a chip like the IRS20957 or Si8231.
Because this amplifier starts up in it's low supply condition only means that the 7812 floats on a lower supply than when the amplifier is driven to it's high supply value of +/-44v.
The class D has no knowledge of the main rails, all it cares about is that the floating supply has >10v and that the bootstrap capacitor is pre charged by12k + diode.
The clock has zero to do with the start up procedure.
This is NOT a clocked amplifier, each channel is a self oscillator.
"Seems that -vcc+12v supply goes to some control circuitry down the line and is variable."
There is no control circuitry on the floating 12v supply.
Hi Steve,
How are the low side driver voltages controlled?
They seem to vary on this amp.
Thanks
How are the low side driver voltages controlled?
They seem to vary on this amp.
Thanks
You stated that the amp wouldn't start if you discharged the 4 bootstrap (bias) caps. How could they remain discharged if they were being charged by the 12k resistors?
As in pretty much all class D designs which use N-Channel totem pole outputs, the low side is driven from the floating 12v supply which sits 12v above -Vcc.
The low side voltage CANNOT vary unless the 7812 regulator is bad. The voltage between -Vcc and the output of the 7812 is fixed at 12v no matter the rail voltage.
Understand that I design all my class D amplifiers to run off a lower rail under low power conditions and when the signal demand is such that the sensing circuits detect the drive signal is about to clip, the power supply is instructed to switch to it's high supply value.
The reason for this is that the efficiency with program material is higher than with fixed rail designs (Class D amps are NOT 90+% efficient unless driven with a sinewave into a pure resistive load and this does not take into account any inefficiencies of the power supply).
At low power where most listening is done, the amplifier operates as a a 4 x 50w 4 ohm device. There is no need for a high rail voltage under these conditions.
This is why class G/H and full tracking supplies were born.
The noise of class D amplifier sis lower with lower rails and when one wants to listen to the noise in their tweeter with no signal present, a low rail supply lowers the noise floor.
Class D amps are NOT as quiet as class B amplifiers so I give them a little help.
The low side voltage CANNOT vary unless the 7812 regulator is bad. The voltage between -Vcc and the output of the 7812 is fixed at 12v no matter the rail voltage.
Understand that I design all my class D amplifiers to run off a lower rail under low power conditions and when the signal demand is such that the sensing circuits detect the drive signal is about to clip, the power supply is instructed to switch to it's high supply value.
The reason for this is that the efficiency with program material is higher than with fixed rail designs (Class D amps are NOT 90+% efficient unless driven with a sinewave into a pure resistive load and this does not take into account any inefficiencies of the power supply).
At low power where most listening is done, the amplifier operates as a a 4 x 50w 4 ohm device. There is no need for a high rail voltage under these conditions.
This is why class G/H and full tracking supplies were born.
The noise of class D amplifier sis lower with lower rails and when one wants to listen to the noise in their tweeter with no signal present, a low rail supply lowers the noise floor.
Class D amps are NOT as quiet as class B amplifiers so I give them a little help.