Quick heatsink question

[Dan2]

hey guys

i have a 2 ch amp at home which i inverted one side so i could bridge the channels. now its obviously not designed to run in bridged mode - although it does very well, except it does get a little warm - (ie when you play it hard for a while you start to hear a crackle through the speaker) The RMS output is rated at 30W/ch.

the first thing i noticed is that it has a very thin heatsink on it - especially compared to a car amp i have with about 20w more output power. So is it feasable to bolt another (bigger) heatsink on the output transistors and run it on the 4ohm load??

i have a nice thick peice of aluminium with "grooves" in it which i want to use, but i don't know if its better to have a thick piece of alu or a thin piece as a heatsink. my dad reckons the thick piece will hold the heat close to the transistors.

 

[Kevin_Murray]

A thicker heatsink is better since it will conduct heat away from the transistors better. The heat will always go from high temp area(transistors) to low temp area (fins). This allows the heat to transfer to the air better. A heatsink with high mass also better protects the transistors from momentary high power.

 

[wg_ski]

Bigger/better heatsink would make the amp live longer with the heavy load. But it won't solve the 'crackling' problem. It's probably the amp either going into hard current limit or the outputs being starved. For the amount of trouble it would be to upgrade the amp components, you could just build a bigger one.

 

[AndrewT]

Hi Dan,
what is the rated impedance for that 30W output?

The rule for bridging is:- double the power into double the impedance.

I seriously doubt that amplifier was specified for driving a 2ohm speaker. So it won't drive a 4ohm speaker when bridged.

 

[TheMG]

AndrewT is right on the money.

On a bridged amplifier with a 4 ohm load, each channel will "see" a 2 ohm load. In order to drive a 4 ohm load bridged, the amp has to be capable of driving 2 ohm loads on each channel.

If your amp is not 2 ohm rated, then it is not unlikely that it is entering overcurrent protection/limiting, and that you would actually get MORE power running it into an 8 ohm load (bridged).

 

[Dan2]

the amp was a VERY cheap buy ( it has crossover distortion and the loudness control is terrible - no bass without it and too much with it) so if i do blow it im not really that concerned. this is what i want to do with it.

run a car sub off it - i know it doesn't really have the power but its the best i got. i already have the sub.

run a car radio for the other speakers and make a reasonably good sound system for parties.

the amp is very happy with 4 ohm bridged - only gets a bit too hot for my liking. last night i thought about forced cooling - a computer PSU fan hooked up to blow across the heatsink. And then i will set the gain of the amp so it stays cool enough and use it like that.

i was looking at building an amp for this application, but i cannot afford a transformer for it so i was going to use the transformer from this amp (its a really nice toroid) which would mean putting this amp away in a box on top of my cupboard. so thats why i just want to use this amp and save the time and money of building another amp.

Also the power supply caps are 3300uF each - you think i could get some bigger ones for a bit of extra power??

 

[jnb]

Thermal inertia, as Kevin_Murray was talking about.

Imagine two heatsinks, identical except that one has a thicker base plate. If the outputs are getting instantaneously hotter and colder with the music, this thick base plate will tend to hold the amp at the average temp. It will tend to absorb the hotter excursions and maintain the warmth during the colder times, making efficient use of the fins. Keeping the transistors at a steady temperature is also better for their life expectancy.

You still need to ensure that the fins have enough surface exposure to be able to dissipate the amps overall heat.

 

[rjb]

Lets keep it simple.
The thick aluminium will conduct the heat away from the transistors better, to where it can disipate. However the rate the heat then disipates into the air is then controlled largely by the surface area of the heat sink.
The best heatsinks have a thick base where they bolt to the transistors, with thin fins cast on the top, where the air can circulate. The thick base gets the heat evenly to each fin, and each fin gets the heat evenly into the air.
This arrangement is to make most economical use of the aluminium.

If instead of using a proper heatsink, a simple plate is used, it must be thick enough to allow the heat to flow away to the cooler areas, and large enough in area to allow the heat to flow into the surrounding air.

This is expensive for a manufacturer in terms of both size, and quantity of aluminium. But if you have a source of free aluminium offcuts, as I do, that is not of concern.

The other advantage of a thick (rather than thin) plate, is that playing music, most amplifiers produce a varying amount of heat, depending of the volume of the passage played. The larger thermal mass of the thick plate absorbs these peaks better, to maintain a more constant temperature.

The cooler the transistors are kept, the better, so larger heatsinks are always a good idea, (but with diminishing returns as usual).

Regarding larger capacitors in the power supply, these will not give more power on average, but will give more power on the peaks of sound. With no sound, they charge up to a certain voltage level. With sound, the current drawn from them will cause the voltage, (and hence the amplifiers power) to drop. This causes distortion if the drop is too great. If the capacitors are too small, the bass is particularly affected. The lower the speaker impedance the higher the average current required, so usually these caps are designed around a 4 ohm load. In some cheap amps these are made deliberately small so the power output is reduced with low impedance loads to protect other parts of the circuit.
If the capacitors are already big enough, adding more does not do any good.

Running low impedance speakers draws much more current from the amp, so the demands on both the heatsink and the power supply capacitors are increased. However the demands on the output stage transistors also increases,with the higher currents required, and they may not be able to cope. Don't forget the heat has to get out of the transistor chip into the heatsink. Many amplifiers do not double their rated power when the speaker impedance is halved, similarly bridging may not double the available output power.

 

[jnb]

Also the power supply caps are 3300uF each - you think i could get some bigger ones for a bit of extra power??

Originally posted by rjb
If the capacitors are already big enough, adding more does not do any good

The point at which this occurs for 4 ohms at 20Hz is 20,000uF per rail. This may be the theoretical ideal but it is, however, a great deal of stored energy and some may consider it overkill.

 

[Dan2]

thanks for all the info guys

 

[AndrewT]

The point at which this occurs for 4 ohms at 20Hz is 20,000uF per rail. This may be the theoretical ideal but it is, however, a great deal of stored energy and some may consider it overkill.

overkill? no.
I use +-40mF/channel for 4ohm speakers or +-20mF/channel for 8ohm speakers.
I arrive at this using the input filter as my starting point.

I want the high pass roll off of the power amplifier to defined by the passive filter that results from using a DC blocking capacitor at the input.
Even for small speakers I have found that the F-3db rolloff should be below 2Hz. This requires the filter to have an RC time constant of 80mS to 90mS. i.e. using Cinblock= 2u2F and Zin=39k, RC=86mS, F-3db=1.9Hz

The NFB rolloff must be at least half an octave below this demanding RC>=1.4*86mS>=120mS. i.e. using Rfb=1k0 and Cnfbblock>=120uF//1uFpp. use 150uF if 120 is not available.

The conventional PSU now needs it's F-3db rolloff to be another half octave below the NFB rolloff. If 120uF were adopted giving RC=120mS, then PSU RC>=1.4*120>=170mS.
For 40hm speakers this demands Csmoothing to be >=170mS/4ohm>=+-42mF/channel.

Yes, the passband of the amplifier determines the NFB and PSU rolloff frequencies as well as the DC blocking capacitor value.

Many on this Forum do not agree with this philosophy. Try it and judge for your self.

Note.
if the source also has a DC blocking capacitor, that must be taken into the input filter calculation. eg Source DCblock=10uF then increase Zin to 50k to maintain a similar RC value.