PQFN 5X6mm Packaged N channel Power Mosfet in ClassD ...

Status
Not open for further replies.
4 layers realy should be the min with SMD, just to distribute power.
Anyway I digress a quick bit of "joining the dots" and here is the first pass at routing. I got a few things to sort, mainly copper pours for the power, +/-vcc pours will also provide the heatsinking for the FETS.
Nite for now.
 

Attachments

nice work, yeah i'm no designer, but I would think the same, the 'traces' have to go somewhere. so as discussed earlier in the thread you will be providing pads for attaching heatsinks externally? or is the copper enough?
 
Hi,
all components are added to my Cadstar libraries as parts, so I can instantly provide: Manufacturer, part number, pick and place tape size, terminal finish etc etc, so we can create a full bom, pick n place etc from the design. I will do an assembly drawing (I've just got to tidy up the names), gerbers, excellon drill data etc, basicly the same data pack as I create when doing this at work.
Thermal design will be fun, and any input would be gratefuly appreciated.
there are a number of avenues to explore such as:
http://www.anglia.com/product_guide/heatsinks/414.pdf
The 0.5W and 1W resistors are Vishay RCL series, which combine small size and power handling.
Vishay - Resistors, discrete - RCL e3 - Long Side Termination Thick Film Chip Resistors

To give an idea of FR4 thermal properties a 2x2 inch square of 1oz copper is between 16-23 deg C/W. This design would be better with 2oz copper or higher. I have done some work with these
Teknoflex flexible circuits - Sculptured Circuits
where you can get up to 7oz copper, and would like to look at using them for an audio design as you can eliminate a lot of internal wiring etc plus have flying limbs to attach hot devices to heatsinks.
Again another avenue to examine is the localised power supplies, a lot of boards have multiple voltages and there are some nice solutions that are more efficient than linear (less heat) such as:
Power Management ICs | National Semiconductor ? SIMPLE SWITCHER Voltage Converters, Voltage Regulator, LED Driver, LDO, WEBENCH Power Supply Design, Voltage Reference, MOSFET Gate Driver Controller, Battery Management, DC-DC Voltage Converter

🙂
 
HOLY SH1T BATMAN!!!!! that is the most insane thing i've seen and the surface mount interconnects, sculpted pin headers and pin-flex would have saved me so much time with my dac. I have used a veritable bucketload (well over 100USD) of molex and pin headers to keep interconnections short as possible while keeping the build modular. we have the upgraded PTFE with gold plate over silver immersion on 4 oz copper coming up and there will be no kits for this one, so we only have to cater for ourselves and that makes this a possibility provided they would accept a short run

I think this tech is very exciting man!! I have often wondered about such a thing, but had no idea the tech was that far advanced along that path already. is it widespread? or are these guys pretty pioneering? what is it stereo lithography, CNC or some other kind of process that deposits the material in layers like that?

I would also think they could incorporate the termination resistors (if any are needed still) directly into the PCB as a feature of the board (hell maybe even diodes)

i'm beginning to think you work for someone rather scary in a mountain lair
 
Last edited:
Inspired by this thread - last night my brain digged out a simple cooling concept.
In fact so far I was considering the IRFB4127 for my new amp but since I had a simple cooling idea - I am really considering to go for the 5020, too.
The idea is to use one or two standard CPU cooling systems, with heat pipe and temperature controlled fan. Normally you won't need the fan and it is off.
Only when playing continuously at high level during a party or concert it would turn on - and then its noise doesn't hurt.
The CPU cooling system could be placed on the opposite PCB side.
Heat can be transfered through many thin vias to the cooling side where the CPU cooling system would be placed and isolated by a thin isolating thermal interface.
 
Funnily enough we are investigating heat pipes as well.
The peltier elemnt sound a good idea.
To add to the info got my SMT mag this morning and this was in it:
Surface Mount Technology May-June 2010

Hald day at work tomorrow so I shall be back on the job.
Qusp, flexi circuits are used quite extensively, but are used in millitary/aerospace due to the reliability over wires and seperate connectors (vibration proof), plus for all industries ease of assembly, signal integrity etc.
 
i'm talking more about the pinheaders that extend right out of the PCB, the trace IS the pin header and the flexible pin strips for joining PCBs. flexible circuits are in iphone/ipod etc for all the cables etc. its the copper sculpting stuff thats way cool and I havent seen so much of
 
Top and bottom copper. A couple of things to tidy up. The output capacitor to add.
All routed, only 2 links, solid copper ground plane.
I've routed it so it should be easy to etch at home using 1oz (or 2oz) FR4, the vias can be created using some thin wire and soldered top and bottom.
Will write more later as Im done it (it 10 here in the UK).
 

Attachments

Hi Savu,
i was thinking of doing two versions, a home etch one and one if using a PCB manufacturer. I hope to get back on it today! as my lovely daughters had a bit of a disagreement over face book, with the net result thay the keyboard and mouse do not work now!!
 
Hi
no, I'll add them and the rest, as the GND connection of that cap is quite important, as it forms part of the switching loop so I want to concentrate all the noisy returns to one small area of PCB, this will limit the loop for the switching currents, one of the main criteria for EMC reduction, which can cause some interesting effects when spikes are generated.
 
Hi Savu,
been reading through this lot amonst other stuff:

http://www.irf.com/technical-info/appnotes/an-978.pdf

http://www.irf.com/technical-info/appnotes/an-1135.pdf

Searches "+ir2110 +layout"

As a result I am re-doing the layout, but focusing on getting the current loops as low inductance as possible. Figures 7 & 8 of AN-1135 are a good starting point. The initial layout IMO had too much stray inductance for the drives to the FETS, and I just wasn't happy with it after reading the stuff and having a look on some of the Electronics forums.
I had origionaly layed it out to minimise any jumper connections, this unfortenatly this was wrong, I should have concentrated on the critical switching loops and how they interact with the power supplies, mainly +VCC & -VCC😱.
 
Status
Not open for further replies.