Interesting, except that all the PL380 blurb clearly shows ISOTOP packaged devices (also quote -- "a single pair of large-die FETs"), not the APT power module.
And this is an *enormously* heavy-duty module (110A 1250W) which I can't believe QSC would use, it would be way over-rated. Mind you APT do even bigger ones like the APTM50UM13SAG (250A 3300W)...
If you were going to go the module route, you can get a complete half-bridge (with series-parallel diodes) in a single module like the APTM50AM38STG which will be cheaper and switch faster than two separate modules.
Mind you, I bet "cheaper" is only relative, I suspect these modules cost several hundred dollars each...
Ian
And this is an *enormously* heavy-duty module (110A 1250W) which I can't believe QSC would use, it would be way over-rated. Mind you APT do even bigger ones like the APTM50UM13SAG (250A 3300W)...
If you were going to go the module route, you can get a complete half-bridge (with series-parallel diodes) in a single module like the APTM50AM38STG which will be cheaper and switch faster than two separate modules.
Mind you, I bet "cheaper" is only relative, I suspect these modules cost several hundred dollars each...
Ian
Eva said:
Indeed, which is why body diodes with Trr>100ns are *bad* news for power MOSFETs in class D amps (switching current more than doubled by reverse recovery = switching power loss more than quadrupled)...
So if you want really high efficiency (say >95%) you not only need to switch fast but also either need to use 100V MOSFETS (which have Trr<50ns) or series/parallel hyperfast diodes to keep the current out of the MOSFET body diodes -- IGBTs just aren't fast enough, even the IR Warp2 series.
Or maybe some semi-resonant switching scheme to make sure the FET current is always forward at turnoff, but this also increases power dissipation in the FET which is about to turn off as well as needing some way to recover the energy stored in the inductor.
I've been looking at all this because I've been investigating using Zetex's DDFA chipset for high-power multichannel PA amps (2kw/4ohms), the problem is it switches at 850kHz so switching loss is a *big* issue...
Ian
Ian
Eva said:
Only Bobbin is covered with this extra turn
Winding order is S
+ET:SThe extra turn has no external connection, its end is terminated and insulated with tape, but it covers the whole bobbin just like a shield.
The core has another copper foil wrapped around it covering the bobbin and core externally, but its connected to Live negative BUS of SMPS.
#Workhorse i had a similar idea some time ago, maybe its same:
to cancel the switching spikes, transfered by the stray-cap. from the coil, you could use a switching signal with inverted phase, and couple this with some stray-cap. also...
so most transfered energy from stray-cap. is canceled, without needing a big Y - cap.
maybe this "free" extra turn does exactly this...
to cancel the switching spikes, transfered by the stray-cap. from the coil, you could use a switching signal with inverted phase, and couple this with some stray-cap. also...
so most transfered energy from stray-cap. is canceled, without needing a big Y - cap.
maybe this "free" extra turn does exactly this...
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.