Help with this amp? A patchwork product of simulation

[AndrewT]

Hi,
with that hole you have options.
You can bolt the To126/220 package to top side or bottom side.
On topside an insulating spacer and a small heatsink under the package for To220, or heatsink on top for a reversed To126.

In any case, a similar hole for every medium power device is very useful.

[MJL21193]

Originally posted by AndrewT
Hi,
with that hole you have options.

Hi,
I have it for heatsink to sit on top, facing up. That way I don't need to change the pin routing and worry about side clearance for the sink.
I saw the space there, so might as well use it up.

EDIT: Andrew, do you think the 2 ounce copper on the top would provide enough cooling for this VAS without a heat sink?

[AndrewT]

I wouldn't use the PCB for cooling. It might short out.
What is your estimate for the PCB dissipation? 100C/W or 1kC/W?

But you like to run big currents and that probably needs a real sink. 20C/W does not take much space.

[Nico Ras]

Originally posted by AndrewT
I wouldn't use the PCB for cooling. It might short out.
What is your estimate for the PCB dissipation? 100C/W or 1kC/W?

But you like to run big currents and that probably needs a real sink. 20C/W does not take much space.

The clip on types are between 22 and 20K/w and pretty small, you may even use a piece of 1.0 mm aluminium sheet, about 15 x 25 mm will do.

[Salas]

Originally posted by Nico Ras


Or speaker designers just lost the nack for making efficient descent sounding speakers.

You like that?

To the contrary.

[MJL21193]

For that VAS laying flat on the circuit board with it's back up, I will do like I show in this sketch.
I have quite a few of these small finned heatsinks salvaged from old computer power supplies.

With this I can really make them cook!

[Don S]

John, have you done any more testing for specifications? What is the actual -3DB points of the amp. Have you made distortion tests with the PC sound card?

Don S

[MJL21193]

Originally posted by Don S
John, have you done any more testing for specifications? What is the actual -3DB points of the amp. Have you made distortion tests with the PC sound card?

Don S

Hi Don,
I went out today and bought a new sound card and I've just finished installing it. I am now checking over my Wallen jig, which will save me from destroying this sound card (like I did with the last one).

As soon as I get everything hooked up and running right, I'll try to get a measurement with RMAA.
Judging from the squarewave response and my own listening impressions, I'll be very surprised if the distortion number is a nasty one.

BTW, what's the -3db points?

[Don S]

John, while I believe that too much distortion is a bad thing, that in itself is not the whole story of an amp.

Every amp will have a HF rolloff, generally specified as the 3DB down point (-3DB@), 70Khz, 100Khz, 200Khz. Well you get the idea.

If the amp will not amplify DC, then it will have a LF rolloff. In your design C7 and C13 both provide LF rolloff. 6DB per octave for each capacitor, or filter if you like. C7 prevents DC and LF from getting in and amplified and C13 in the feedback circuit provides 100% feedback of all DC in the circuit's amplification, and can be crucial to good DC offset. You will have a LF rolloff at some frequency of 12DB/Octave, 6DB for each cap or filter. This is what causes your squarewaves to start tilting as frequency is decreased.

Don S

[MJL21193]

Originally posted by Don S
John, while I believe that too much distortion is a bad thing, that in itself is not the whole story of an amp.

Every amp will have a HF rolloff, generally specified as the 3DB down point (-3DB@), 70Khz, 100Khz, 200Khz. Well you get the idea.

Hi Don,
I thought that's what you meant, but I assumed it was predetermined by those filters.


Well, I have results from the RMAA. It took a while to get everything set up. I was going to us my Wallin jig first, but reconsidered. I thought it would put the output impedance too high. In the end, I made a voltage divider, using several 1/4 watt resistors in parallel for 1K and 100R.

I spent a lot of time checking and rechecking the output before I connected to the new sound card. Finally got things right and ran two tests.
Here are the results: