[SOLVED] 12v 30A Power Supply - Water Cooled Heat Sinks

[abuhafss]

Hi

I am building 12v 30A DC power supply shown at

**broken link removed**

I am using two separate water cooled heat sinks ----- one for the 6 transistors and the other for the bridge rectifier & voltage regulator LM7812. Both the heat sinks are connected in series i.e. the cooling water from the 1st heat sink goes into the 2nd heat sink and then to the radiator.

The circuit did not worked and found LM7812 to be the culprit. I wonder if the water circulating in the heat sinks is making any sort of conduction between GROUND (from LM7812) and COLLECTORS of the transistors, which might have ruined LM7812 OR simply LM7812 was actually faulty. I will get the new LM7812 soon but before installing it I would like to have some valuable comments.

 

[chuckey]

I don't understand how the circuit works, but the input voltage to the 7812 looks very high. 24 X 1.4 = 33.6 - 1.6~ 32V. I would be very careful about using water,because it will provide a conductive path and any voltage drop it experiences will cause corrosion, especially if you use copper and aluminium components.
Frank

 

[betwixt]

It isn't an efficient design but as Chukey said, you must be careful to use distilled water and ensure nothing can dissolve into it or you will get corrosion and electrolytic effects.

Chuckey, it's operation is quite simple: the 7812 provides all the output curent until enough voltage is dropped across R7 to start forward biasing the power transistors. They then carry any additional current. It has the advantage that the voltage across R7 is proportional to the output voltage when under load so the regulator controls the transistors as well and it's thermal shutdown also shuts the transistors off. However, in this schematic, R7 at 100R is FAR too high in value, it should drop Vbe (say 0.7V) at about 75% of the 7812's maximum current. If we say that is 750mA, it's value should be V/I = about 0.9 Ohms. In fact f the DC from the rectifiers is ~34V it is only possible for the 7812 to conduct a maximum of 0.34A so the 1A fuse can never blow.

Brian.

 

[IanP]

However, in this schematic, R7 at 100R is FAR too high in value, it should drop Vbe (say 0.7V) at about 75% of the 7812's maximum current. If we say that is 750mA, it's value should be V/I = about 0.9 Ohms.

I think R7=100ohms is all right as most of the 7812's current comes from (E-B)s of the 6 PNP transistors and the IC has to sink (In-Out) all that current that can be quite high:
30A / gain factor of PNPs
R7 only sets the threshold at which the transistors are "turned on" and it can be quite low - means bigger value of R7 ..
So, the design is not as good as it looks ..
:wink:
IanP

 

[abuhafss]

you must be careful to use distilled water and ensure nothing can dissolve into it or you will get corrosion and electrolytic effects.
Brian.

I have used a steatite insulator between the new 7812 and the heatsink. The possible conductive path issue is resolved but unfortunately the circuit is still not working.

The voltage at INPUT and OUTPUT of 7812 is same 31V.
The 7812 is working fine.
All transistors are also OK, when disconnected and checked.

 

[betwixt]

The one thing the 7812 can't stabilize is leakage around it. Possibly you have a leakage path through the transistors, in the silicon, not because of a construction defect. All transistors leak slightly.
Try drawing a small current from it and see if the voltage drops. I would suggest 10mA so you would have to wire a resistor of 1200 Ohms across the output. If the voltage drops significantly, it indicates leakage and there isn't much you can do about it except keep the resistor in place and possibly adjust it's value until you get 12V out.

Brian.

 

[picgak]

Hi abuhafss,
The efficiency of the circuit is very poor. Do not use water for cooling, Use transformer oil for the purpose it is a very good insulator as well.
regards ani

 

[chuckey]

abuhafss, your idea of using liquid cooling of the heatsinks is only halfway there. Liquid only transfers the heat away from a source, the heat still has to be dissipated somewhere - bigger heatsink +fan?. Also there hot liquid needs to be moved from the source to the dissipation heat sink either a pump or make the dissipation heatsink physically higher and use a large bore pipe and rely on thermo syphon to get the liquid to circulate. When using liquids, the system needs to be sealed against spillage incase the power supply gets tilted, which means that you have to have a bellows or similar to encompass the expansion/contraction of the liquid with a change in its temperature. Have I put you off yet?, I used to work on water cooled electronic equipment, water cooling is not some thing you undertake lightly!!
Frank