diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Class D (http://www.diyaudio.com/forums/class-d/)
-   -   typical unwanted amp byproducts (http://www.diyaudio.com/forums/class-d/115036-typical-unwanted-amp-byproducts.html)

gearheadgene 8th January 2008 02:04 AM

typical unwanted amp byproducts
 
I've got my first class-D amp going. It's a triangle wave, carrier based design. So far so good.

The FET's are running very cool at idle (although it's in my unheated garage - you haven't lived until you debug in 30 degrees F :eek: ) Anyway, the FET's are cool to the touch, but the torroid is warm. How warm? I don't have a thermo probe handy, but it's rather warm to the touch but not buring. Is that typical?

Also, what is the usual residual output voltage? I have 1V p-p sinewave at around 400Khz.

Nothing yet is optimized. This is just my initial observations.

BWRX 8th January 2008 02:57 AM

Nice work gene. So what type of construction method did you use - p2p, breadboard, pcb? Any more details on parts you used, schematic, pictures, etc? :)

Depending on the core type, size, and windings, warm to the touch can be ok. Normally you want the coil to be cool when idling though.

gearheadgene 8th January 2008 03:13 AM

for now, it's a point-to-point layout. For something that started out really neat, it's become a managable mess. I first built a triangle wave generator on it's own board to test out that function. Then I built up the main portion.

I didn't mention this, but it also includes a preamp as well. Sadly, my first attempt at the preamp sucked :bawling: and I wish I did it on a separate board.

For a test, I also built up a battery powered sine-wave generator board.

I'll take some snap shots tomorrow. Not terribly exciting but what the heck!

The learning experience so far is pretty cool. Biggest thing I've learned so far? Layout counts - big time!

Any guess why the coil is hot? I'll dig into it more manana.

What about that 1Vp-p 400 khz output?

BWRX 8th January 2008 04:23 AM

Don't be shy now. Seeing point to point and breadboard projects come to life is what makes this hobby fun!

As for your inductor heating issues: What is the inductance of the coil? Are you using a powdered iron core? What size and permeability core? What gauge wire did you use to wind it?

Chances are the core material is not suitable for high frequency use. Do you have any other cores to try?

1Vpp at 400khz should be fine depending on the values of the output filter.

gearheadgene 8th January 2008 10:22 AM

2 Attachment(s)
good point about the toroid core - I hadn't thought of that. The inductor is made by JW Miller, and it came from digikey. Part number is 2206-v-rc, and this is the data sheet digikey offers: http://www.bourns.com/pdfs/2200_series.pdf

The inductance is 27uH spec'd at 1 KHz (per data sheet).

Here's a snapshot of the board

gearheadgene 8th January 2008 10:29 AM

2 Attachment(s)
this is a closeup of the main board, component side. Most of the main stuff is in the upper 1/2 of the card. The lower right half has the feedback network while the lower right half is my sorry attempt at a preamp :)

You can see the inductor clearly in this picture.

How do you like my heat sink! Originally there were separate aluminum heat sinks that looked more traditional. During debug, I was having trouble with the FET's and though they needed to be thermally connected so I used a small piece of aluminum. For now, they stay cool.

gearheadgene 8th January 2008 10:33 AM

2 Attachment(s)
This is the sine wave generator. Looks a little like star ship enterprise (if you squint a little and swig some sambucca, that is). The circuit can generate any frequency from about 100Hz up past audio frequency. Output is variable too. With batteries, the output is fully floating - which I liked.

gearheadgene 8th January 2008 10:37 AM

2 Attachment(s)
last shot for now is the under belly of the main circuit board. Most of the wiring is here. I used solder wick for buses.

Notice the large caps on the bottom. During debug it appeared that the rail decoupling was a problem so I needed to add more capacitance. No room on top for a good layout, so they wound up on the bottom instead.

BWRX 8th January 2008 01:51 PM

Quote:

Originally posted by gearheadgene
The inductance is 27uH spec'd at 1 KHz (per datasheet).
That may indeed be the reason the core is getting hot. You need to be very wary when manufacturers only specify their cores at very low frequencies. I certainly wouldn't use any inductor that is only specified at 1kHz in a class d amp that switches at 400kHz.

gearheadgene 8th January 2008 03:23 PM

Looking at other inductors from JW Miller, they have "low loss core" styles too. Those are spec'd at much higher frequencies. I've fired off an email to the manufacturer for some recomendations.

Any feelings about using shielded inductors instead of toroidal? The EMI benefit seems worthwhile.


All times are GMT. The time now is 04:28 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2