Hi,
Not really. I never said speakers can be purely reactive.
(But some drivers, e.g. piezo's and electrostatic panels can
get pretty close at high frequencies if they are allowed to.)
You might test an 8ohm amplifier with phase angles up to 45 degrees,
or more rigourously up to 60 degrees, but in my book its near the
same as testing into say 4R for an 8 ohm amplifier. Current levels
will be different but the thermal dissipation required is similar.
rgds, sreten.
Watch here where I have demonstrate about variable impedance of a speaker under and above resonant freq.
The kewtech multimeter is set for alternative curent seried with the speaker, to measure curent on speaker
The modification of impedance regarding res freq is observable between sec 58 and min 1:40, also from min 2:00 to min 2:30
Resonant freq of speaker is around 38Hz, and 4 ohm nominal (3,6 ohm DC).
Part 4 - YouTube
sreten said:
Yes, but you said purely resistive load. If it is not a speaker, then who cares? Amp is used to drive speakers. Calculate real demands!
OK, but why?
The modification of impedance regarding under and above resonant freq is observable between sec 58 and min 1:40, also from min 2:00 to min 2:30.
Part 4 - YouTube
So impedance of a speaker is variable regarding freq.
You're not making any sense.
Impedance is resistance defined in the frequency domain so naturally impedance varies with frequency. It's the very definition of impedance.
If you measure carefully you'll also note that a speakers impedance varies with current coming into it as resistance increases as the voice coil heats up. This is called power compression.
A pure resistance is independent of freq.
Impedance is the complex (matrhematical term) combination of resistance (dissipates power), and reactance (does not dissipate power), which consists of capacitance and/or inductance.
You mean an ideal resistor. Pure resistance is a value, it can be also a special point on an impedance curve, minimum or maximum value. (Minimum or maximum impedance is always resistive, if it is finite.)
you don`t know that.
you can build class d with as big psu as you want but it`s limited
by protection circuits which is not the case with AB.
Oh, back to the old discussion about leaving the protections out?
Any good amplifier should survive an intermittent short on the output, it's one of the most usual causes of blown output stages ever. This requires current limiting. In practice, most hi-fi amplifiers won't pass that test, even the ones that include elementary short circuit protections (and the manufacturer that says 80A or 120A for a few bipolars has become a joke over time). In the beginning, I got into amplifiers mostly by repairing class AB, I still fix one from time to time.
Any good amplifier should survive an intermittent short on the output, it's one of the most usual causes of blown output stages ever. This requires current limiting. In practice, most hi-fi amplifiers won't pass that test, even the ones that include elementary short circuit protections (and the manufacturer that says 80A or 120A for a few bipolars has become a joke over time). In the beginning, I got into amplifiers mostly by repairing class AB, I still fix one from time to time.
Last edited:
No, no no and no.
As already mentioned class-d is a power converter which can deliver much more current into loads of low resistance than its power supply can deliver. Class AB can't do that because class AB is nothing but a controlled voltage-divider.
This is of course theory and how much it will be practically is depending on the actual implementation.
This statement is too general to be true. It depends on the practical implementation again. An AB amp must be veeeeeeery beefy in terms of output devices in order to be able to survive a very low load resistance or even a short circuit. Keep in mind that you can leave the SOA of the usual bipolar transistors very quickly when they have to conduct high currents while having some tens of volts across them.
Therefore most AB amps also have some sort of short-circuit protection (or the more refined version called SOA protection).
There is actually one problem (or better: challenge) with class-d amps when driving high currents at low voltages: The PSU pumping effect is increased. But it is possible to take care of this as well.
Regards
Charles
The main challenge is fast hard switching.
Typically you'd like to have around 10ns of di/dt switching phase.
Hard switching as little as 10A in 10ns results theoretically in as much as 10V drop over 10nH (that's around 1cm of both leads, internal device connections and PCB traces combined). That's a sort of calculations one used to do in a secondary school.
From this point of view it starts to make sense to parallel devices or perfectly running half-bridges parallel in multi-phase (aside from other advantages)
Typically you'd like to have around 10ns of di/dt switching phase.
Hard switching as little as 10A in 10ns results theoretically in as much as 10V drop over 10nH (that's around 1cm of both leads, internal device connections and PCB traces combined). That's a sort of calculations one used to do in a secondary school.
From this point of view it starts to make sense to parallel devices or perfectly running half-bridges parallel in multi-phase (aside from other advantages)
Don't forgot "the same rating", phillips's nxp 100W got in protection while overloaded, for 100W class AB, it need 2pair of 2sa1916-c2922. with only 5A fuse protection not blow in similar overload.
Because classD itself is reactive, add an additional big supply caps before amp current limiter, may will solve it.
Last edited:
Do you use a specific layout practice to ensure symmetric sharing of di/dt and consequently equal switching losses in all devices?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.