Heat Sinks - How hot is too hot?

[Jack.Straw]

I've built this NiMH charger (with lots of help here, thanks!) and it seems to be working. However, I haven't tested very long because i'm worried about the heat. After a couple minutes of charging the heat sink gets so hot that it burns my finger after about 2 seconds. How hot is too hot? Do i need a larger heat sink? Here is a picture of what i'm using now:

**broken link removed**

Here are the details of the project:

Maxim MAX712 Charger IC: Specsheet w/ Schematics

Charging: 6 AA Rayovac Hybrid 2100mAhr NiMH batteries

Rate: C/2 - 1050mA

Power Source: **broken link removed**

Load: LM386 based practice amplifier - The Ruby Schematic

My perfboard design:
**broken link removed**

 

[HiTech]

Heatsinks should become warm to a bit uncomfortable, but never enough to burn or alter skin. Try a considerably larger heatsink, apply thermal transfer paste between the two surfaces, and if need be, a small fan.

 

[MikeMl]

IC chips/transistors are still ok if the metal tab is ~75degC, which will burn your finger. Safest way is to get an IR thermometer, aim it at the tab/heatsink, and measure the temperature right there. If it is too high, get a heatsink with more area, or get more air moving over the heatsink.

 

[Jack.Straw]

Thanks for the responses! I chose that heat sink because it looked like it was made to fit the TO-220 package of the 2N6109 PNP. Do they make larger heat sinks that will still fit the TO-220?

Also, can anyone tell me if this amount of heat is to be expected from a charger like this?

Thanks for your time,
-Scott

 

[MikeMl]

Thanks for the responses! I chose that heat sink because it looked like it was made to fit the TO-220 package of the 2N6109 PNP. Do they make larger heat sinks that will still fit the TO-220?

Also, can anyone tell me if this amount of heat is to be expected from a charger like this?

Yes, and use your Multimeter to measure the transistor's Collector voltage Vc and its Emitter voltage Ve. Measure the current (I) into the regulator along the positive lead.

The heating is proportional to power dissipated. To a first approximation, the power dissipated by the transisor is (Vc-Ve)*I. If you get 2 or more Watts. then that little finned heat sink needs to be larger

 

[Jack.Straw]

Yes, and use your Multimeter to measure the transistor's Collector voltage Vc and its Emitter voltage Ve. Measure the current (I) into the regulator along the positive lead.

The heating is proportional to power dissipated. To a first approximation, the power dissipated by the transisor is (Vc-Ve)*I. If you get 2 or more Watts. then that little finned heat sink needs to be larger

Well, without being home to measure, I took this from the MAX712 datasheet:

Calculate worse case power dissipation:
PDpnp = (max wall-cube voltage under load - minimum battery voltage) * charge current in amps

The batteries at full charge are 7.2 volts, but i don't know what the "minimum" voltage is. Anyone know? If I were to take a wild guess and say 6 volts, then

(12v - 6v) * 1.05A = 6.3 watts

... much higher than 2 watts. How big would you say the heat sink should be?

 

[MikeMl]

Keep making it bigger until it no longer burns your finger

 

[crutschow]

You want a heatsink with about 8C/W or less thermal resistance to air. That would give a 50C rise in heatsink temperature above ambient.

Alternately you could add a power resistor in series with the transistor to reduce the voltage (and thus dissipated power) in the transistor.

 

[mneary]

You have connected your amplifier load so the current it draws is subtracted from the chip's current sense. The amplifier(-) should be connected to the battery(-).

 

[Jack.Straw]

You have connected your amplifier load so the current it draws is subtracted from the chip's current sense. The amplifier(-) should be connected to the battery(-).

Are you sure? I'm certainly new at this, but the MAX712 schematic shows the load(-) on the ground side of rSense:

**broken link removed**


Am i wrong?

 

[mneary]

I hesitate to call it "wrong", but I would almost always connect the load to the battery, to:

(1) protect the regulator.
(2) provide a more direct connection (lower impedance) of the load to the battery.

 

[MikeMl]

I hesitate to call it "wrong", but I would almost always connect the load to the battery, to:

(1) protect the regulator.
(2) provide a more direct connection (lower impedance) of the load to the battery.

Yes, but this charge controller is trying to measure charging current into the battery separated from how much the load is drawing. If you connected the load to the top end of the sensing resisor, the charge controller cannot separate the two currents...

 

[mneary]

That's where I depart from the charger designer. This OP is concerned about the transistor temperature. Connected as shown protects only the battery and leaves the transistor unprotected.

When the battery is charging the load gets Vbat + Vsense, which I also don't like. If a pack is being charged at C/2, its terminals will be well above that. This connection pushes an excess 0.25V to the load.

 

[tvtech]

A quick reply to the original post.

Wet your fingers and touch the heatsink. If you hear a sizzling sound it is running waaay too hot. Seriously.
Tis an old trick I learnt years ago.

 

[Dragon Tamer]

This sounds a lot like a problem that I've been having with my MAX712. For my design I am using a 2955 transistor that was cooled with a very small TO-220 temporary heat sink. The temp was easily getting up to 175ºF. I solved the problem by making a temp controlled fan t cool it off, which I must say has worked very well. I also plan on making a larger heat sink for the finnal build. If you want I can give you the schematic for the fan circuit.

 

[Mikebits]

My Denon 5.1 sound system delivers 100W per channel and it has no fans. If your having to use fans there is a fundamental design problem.

 

[Dragon Tamer]

The fans in my system are a precautionary device, they are normally off, and only come on if the temp of the PNP exceeds the limit in relation to the ambient air.

 

[audioguru]

It is obvious that the heatsink is way too small.

 

[Dragon Tamer]

Right now it is, but like I said, space on the bread board was minimal so some compromises needed to be made.

 

[Mikebits]

Or your input voltage to output voltage drop is too large.