mounting components on the same heatsink / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

mounting components on the same heatsink

Move Thread

LAN_403

StevenCaton

Tue Nov 17 2009, 06:16AM

Registered Member #1845 Joined: Fri Dec 05 2008, 05:38AM
Location: California
Posts: 211

I am going to build a full wave voltage doubler. It will have two chassis mount diodes. I plan on mounting both the diodes on the same heatsink.

Under what circumstances should one not mount components on the same heatsink? I vaguely remeber someone suggesting that putting mosfets on the same heatsink could be risky a while ago.

Do I have anything to worry about with the diodes?

Thanks
Steven

Sulaiman

Tue Nov 17 2009, 12:57PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3141

The only consideration will be electrical insulation - which shouldn't really be a problem.
Diodes/rectifiers sharing a heatsink is very common.

doctor electrons

Tue Nov 17 2009, 08:50PM

Registered Member #2390 Joined: Sat Sept 26 2009, 02:04PM
Location: Milwaukee Wisconsin
Posts: 381

Yep! If the surface that mounts to the heatsink is in anyway part of the circuit (usually with transistors) then you don't want them
tied to the same heatsink. Other than that the only consideration would be dissipation of heat. Too much thermal load can render the heatsink useless.
I do not know of any other reasons.

Myke

Tue Nov 17 2009, 10:21PM

Registered Member #540 Joined: Mon Feb 19 2007, 07:49PM
Location: MIT
Posts: 969

Sometimes you want temperature sharing so that the component's values don't drift too far apart from each other.
An example of this is when you have two diodes put in parallel on the same heat sink. You want the voltage drop across each of them to stay within a certain range otherwise one of the diodes will conduct a lot more than the other which will most likely lead to its destruction.
You really only put diodes in parallel if the single diode costs more than the two diodes or if the dimensions of the higher current diode is a problem. There may be other reasons though.

Proud Mary

Tue Nov 17 2009, 10:21PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

Hello Steve, I'm a very conservative thinker, and don't think you can go too far with precauations and safety features.

Heat sinks cost only pennies, so I wouldn't hesitate to use seperate sinks.

You are not a commercial manufacturer whose design decisions will be guided by the saving of a single cent.

It is an appalling mistake to imagine that the manufacturers of commerical PC boards have done the best possible and are
therefore to be trusted - on the contrary, they have been given a price, and their circuit diagrams flow from there.

You - the amateur - does not have these constraints, and can afford to use 'best practice' as your guide in every design decision.

It's unsound to copy the techniques of mass manufacture - PCB and surface mountage - when you are creating only a single instance of a circuit.

Many here will be able to able to help you with such a simple matter as doubler design - available in any first year textbook - but if you still find yourself in heavy weather, the Minotaur might well come out of its Labyrinth to see you safely on your way,

Stella X

teravolt

Wed Nov 18 2009, 04:31AM

Registered Member #195 Joined: Fri Feb 17 2006, 08:27PM
Location: Berkeley, ca.
Posts: 1111

if you have to-247 or to-220 packages just use mica or silicone insulators. Mica is good for about a Kv. if you are using stud diodes mica and mounting hardware is a little less common so I use 2 heat syncs. the best for high power stuff is diodes with a insulated mounting base hope that helps NB

Dr. Slack

Wed Nov 18 2009, 08:27AM

Registered Member #72 Joined: Thu Feb 09 2006, 08:29AM
Location: UK St. Albans
Posts: 1659

There are several issues here, which carries the most weight depends very much on the specific situation.

Heatsinks conduct two things from device to device, heat, and electricity.

Heat -

Matching may require them to be on the same sink, but this doesn't appear to be an issue here.
If thermal instability requires them to be on different ones, then the design would be very marginal, and to be avoided for that reason alone.

Electricity -

There is a very good reason to strive to have all heatsink hardware grounded - it simplifies the heatsink mounting to the case, and it is safer if you are poking around inside your kit debugging it. Once all the sink surfaces are at 0 volts, it matters little whether they are physically the same or different sinks. Having them as the same sink allows thermal diversity to be used if the devices generate heat at different times. A large sink can even form part of the case.

It may be that the potential from device to sink is too high to acheive an insulated mounting. It would make sense here to put all devices of the same potential, uninsulated, to the same sink, and use a different heatsink for each potential. While it's possible to insulate some devices if they are at a different potential to a high potential heat sink, it's ugly and untidy, and to be avoided on the grounds of art alone.

StevenCaton

Thu Nov 19 2009, 06:27AM

Registered Member #1845 Joined: Fri Dec 05 2008, 05:38AM
Location: California
Posts: 211

Thanks for the responses guys. Well, I hate to rain on the parade, but I've figured out a way to build a voltage doubler (with 117VAC input using one bridge rectifier. One bridge = one heatsink.

However, I now understand that mounting components on the same heatsink should usually be avoided, or, heat pads should be incorporated to maintain insulation.

Proud Mary

Thu Nov 19 2009, 10:45AM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

Hi there Steve,

you didn't say what your required output current was - from which we can roughly estimate the best size of capacitor for your doubler.

A few little tweaks which you probably know all about, but which might help the newcomer:

It's not a bad idea to put some small Y-class condensers in parallel with your diodes. This will effectively slow down the rise time, and so reduce square-wave type noise (60Hz and its harmonics) and also give some surge protection to them. By Y-Class I mean the type of condensers certified for direct connection across the mains - but I am not that familiar with American 'electro-taxonomy' (I just made that one up!

) so there may be a different certification scheme in the US. The main features are an ability to withstand a couple kilovolts for one minute, (lower for you at half the mains voltage) and being flame-proof etc.

Don't forget to put a slow-blow fuse of the right size immediately before the diodes.

I'm probably one of the last creatures left on Earth who routinely uses LF choke-input HT smoothing, so I shan't recommend it here.
But do put some generously rated wirewound resistance of low value (say 1R ) upstream of the condensers, to stop
catastrophic inrush currents (i.e. at switch-on the uncharged capacitor looks like a dead short, and heavy - perhaps destructive - currents will flow momentarily.)

It's usually a good idea to put some type of surge suppressor - a MOV or TVS diode - across the lines to clip off mains-borne voltage transients that might destroy the doubler, or whatever application it is feeding. If you are a conservative thinker like me, you may also want to add one or more spark gap arrestors to keep out egregious over-voltages.

With your 110V US mains input, I would suggest you use aluminium electrolytics with a minimum 375V working voltage.

If you put pea-neons with a suitable resistor in parallel with your fuses, they will light up only when the fuse has gone, which can be very helpful in more complex circuits where dozens of fuses may be used.

A nice touch would be to add an AC ammeter on the output side, so you can see at a glance how much current your application is using.

Put the whole caboodle in a metal case with a microswitch on the inspection panel so that the power is automatically disconnected when the case be opened.

If you are not going to put an isolating transformer upstream of the voltage doubler, you are entering the hazardous world of
the "live chassis technique" so even grub screws and collets on the front panel must be made of plastic. Personally, I think this is very bad practice, but the 'live chassis technique' was introduced by the manufacturers of cheap TVs and mains radios to avoid the cost of a transformer.

As a further refinement you might like to add, at a later date, a slow-starter relay, so that the HT builds over a hundred milliseconds or so, which will very likely increase the Mean Time Between Failures of your application by several hundred percent.

Lastly, don't forget to use a DPST switch, so that both live and neutral lines are disconnected in the OFF state. And a red HT ON dial lamp as the cherry on top of the cake.

Good Luck!

Edit: After thought: Golly, I can't think what came over me, but I forgot to mention the importance of suitable, very generously rated bleeder resistors across the output. These should draw about 10% of the total current and are there for two reasons, namely, to stop the HT rising too much in the case of low or absent loads, and to discharge the capacitors for safety's sake with HT OFF

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob