Conical sring or compression washers are recommended by some authors for holding TO-247 and the like to heatsinks. However, this doesn't seem to be something one finds at the local hardware store. Anyone know a source for these (other than industrial source that have a minimum order of 100k pieces per month?)
As long as I'm on this subject, how about some form of spring clip that could be mounted on a heatsink and hold down a row of four to eight TO-247's.
I'm trying to see if I can find a convenient DIY solution better than just a bolt and a flat washer.
As long as I'm on this subject, how about some form of spring clip that could be mounted on a heatsink and hold down a row of four to eight TO-247's.
I'm trying to see if I can find a convenient DIY solution better than just a bolt and a flat washer.
The idea is good untill you look at the math. The thermal expansion of 4mm (approx. thickness of a TO-247). You will be operating within less than 70C (room temp to max temp), so thermal expansion is no issue as the thermal expansion of the device is very close to the range of thermal expansion seen for metals.
Counting out all that, disc springs are meant to work as springs, not as washers, so when loaded hard enough to mount the device properly, Im quite sure(guesstimating here) that it have turned into nothing but a flat washer. Ive been working with disc springs in different applications, they sure are springs.....not washers, as you quite easily overload them.
Magura
EDIT: I just read on Mc Master that they recommend them as washers, according to my experience that is highly optimistic, as they loose tension or crack over time, when under max load...and over time dosnt mean 10 years.
Counting out all that, disc springs are meant to work as springs, not as washers, so when loaded hard enough to mount the device properly, Im quite sure(guesstimating here) that it have turned into nothing but a flat washer. Ive been working with disc springs in different applications, they sure are springs.....not washers, as you quite easily overload them.
Magura
EDIT: I just read on Mc Master that they recommend them as washers, according to my experience that is highly optimistic, as they loose tension or crack over time, when under max load...and over time dosnt mean 10 years.
Magura, was your experience with "disc springs" used as washers or with "belleville washers" designed for use as such? This raises some concern, as I've used belleville washers (e.g. McMaster-Carr # 90127A007) to mount TO247s in a number of amps. One is now several years old, but I've seen no signs of failure yet...
yeah, this is the first I've heard of "belleville washers" and the description sounds like they are siutable. Not just to secure the fastener but als to distribute the presure over a wider area of the TO-247. I looked on the web and they seem to be a standard industrial or construction component for heavy loads. I would like to hear if this is the same or different from what Magura is refering to.
These are the most reliable way to mount power semiconductors. Just go to the On Semi website and look for an app note called "Power Semiconductor Mounting Considerations" or something like that.
Regarding your question about spring clips, these don't apply nearly enough force for TO-247s. They only work for small packages dissipating a fraction of a watt.
Regarding your question about spring clips, these don't apply nearly enough force for TO-247s. They only work for small packages dissipating a fraction of a watt.
Charles Hansen said:
Totally, totally, in agreement..
Bellvilles will last forever if used properly.
Over time, the plastic of the TO package will distort or creep. A regular washer will unload after a coupla years, while a bellville will retain a larger portion of it's initial load, depending on the spring constant.
I used a stack of five bellvilles to maintain pressure on a diode, and used 50 mils as the deflection required at 12 thousand pounds operational force. I build 880 devices, and they all have survived cooldown to liquid helium now for the last 5 years. They are extremely reliable...they had to be..downtime for the machine is approx. a million dollars a day, with 5 days minimum to warm up a sextant, repair, and cooldown.. so the owners frown on broken diodes.
Bellvilles can be stacked in series or parallel...if they all cup together, the force per unit deflection goes up (they are stiffer), and if you antiparallel them, the spring constant goes down..gets softer.
They are also very good when you use polymer insulators or nylon isolation hardware for metal packages..how many of us had to re-torque the semi's on a heatsink after 10 years?
Cheers, John
and I've neverheard of any failing!
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.