F-5 output device failure mode

Russellc

Member
Paid Member
2003-03-06 12:59 am
midwest
Recently posted concerning bias, junction temp and longevity of the mosfet output devices.

Checking the 4-40 bolts that hold my IRF devices to the heatsink, I found they indeed needed just a slight twist to snug up again....not too tight though. Wondering if they were loose enough that any damage was done to the mosfets, because of not transferring enough heat.

Frankly, they were just a smidge looser, not much at all, but it got me to wondering about how these devices fail. The only ones I've read about failing
on F-5 have been from miswires, usually switching the wrong device in the wrong place on the board.

Do these devices tend to either "work" or "not work" when they fail, or is it a gradual breakdown with worse and worse sound quality or what?

Thanks to any that know,

Russellc
 

Russellc

Member
Paid Member
2003-03-06 12:59 am
midwest
Use spring washers to keep bolts tight.
PS. And do not use hand held cordless drill for drilling heat sinks. The holes need to be ideally vertical.

I'm not familiar with "spring washers" Frags, do you have a link or pic?

As to drilling the sinks, I found the 10-32 holes for the angle aluminum were the most difficult to keep square, but mattered the least. The 4-40 holes with careful eyeballing were square as can be....I did the first as a trial and it was true as I could measure, of course there are only 4 of those for my F-5 build. The larger 10-32 were more difficult to keep straight and true....but since they just went through holes in the angle material (whose hole can be hogged out to allow fitting and trueing) so it really wasnt as important as clamping a fet square to the sink surface with the 4-40.

russellc
 
I haven't seen many reports of failure of the outputs other than catastrophic failure due to a miscue. If they do fail, it probably wouldn't be gradual and if your device-to-sink connection was fairly diligent, you should be OK.

I know our two builds are pretty close, same heatsinks, same chassis, different outputs but current handling is the same. I've measured output device temps, case temps, sink dissipation, etc. and was startled to see how low the actual values actually were.

One thing I did to insure a good mating surface was to mirror polish the heatsinks, especially in the area where the devices are positioned. FWIW, I positioned them as close to 1/3 up from the bottom as I could, that was mentioned as a good placement. I don't think it is absoluetly critical, the M&M heatsinks dissipate the heat very well.
 

frags

Member
2009-08-20 7:25 pm
I'm not familiar with "spring washers" Frags, do you have a link or pic?
russellc
Spring washer:
[IMGDEAD]http://www.toolstation.com/images/library/stock/webbig/99682.jpg[/IMGDEAD]
You can find them in any home hardware store.

To mount FETS I'm using regular washer (10mm outside diameter) plus a spring washer.
 

Attachments

  • 99682.jpg
    99682.jpg
    26.5 KB · Views: 190
Last edited:

Russellc

Member
Paid Member
2003-03-06 12:59 am
midwest
Spring washer:
[IMGDEAD]http://www.toolstation.com/images/library/stock/webbig/99682.jpg[/IMGDEAD]
You can find them in any home hardware store.

To mount FETS I'm using regular washer (10mm outside diameter) plus a spring washer.

Oh, I see now, I've just never heard "spring" in reference to them. I'll get some, thanks for the tip!

Russellc
 
IMHO, a Belleville washer (or conical waher) first appears to be the ideal solution, in that it exerts pressure uniformly around the perimeter, which is not the case for the humble spring washer. However, if you would tighten a screw with a conical washer till flush (totally flattened), and then release it again, you would find that the conical washer would have been deformed and lost a sigificant proportion of its conicity (and hence the ability to generate preload force).

Or in simple English, a conical washer should never be tightened till flush, but only be tightened using some form or torque control, so that it is given the right preload within its designed working range.

Disc Springs to DIN 2093
(Note stress level vs deflection)

Heavy Duty Bolting Washers


Patrick
 
Just to confirm the earlier comments on the difference between "shakeproof" and "constant load" washers.

The split washer, the serrated washer and similar are to prevent an improperly tightened nut/bolt/set screw from vibrating off. Similar to self locking nuts.

The "Belleville" washer is one type of constant load type fixing. It is designed as an elastic "spring" between the underside of the nut/head and the device to be clamped.
A stack of Bellevilles can be used to arrive at different spring ratings.
Two Bellevilles one inside the other has roughly double the spring rate (force change as distance changes, eg. due to temperature exapansion) and double the holding power (twice the force).
Two Bellevilles facing in opposite directions have half the spring rate and the same holding power. Combinations of opposite and same direction series Bellevilles can achieve many different loads and stiffnesses.
 
the split washer needs very little force to squeeze to flat.
As the fixing slackens off the two cut edges expand out and dig into the surfaces to prevent further unscrewing. By this stage the fixing is "loose" with very little clamping load. The split washer has done it's job. It stops the joint falling apart and the nut/bolt falling out.
 
All,

Just my two cents. Consider using a thread-locking compound, such as Locktite 243. We routinely build high-performance road racing engines that rely upon accurate and reliable fasteners, to endure through many heat cycles, and under the worst condidtions. We would not assemble a $50,000 engine without the use of Locktite on critical fasteners.

I'll be using it on my F5 MOSFET machine screws. If you use it, just be careful--a little goes a long way, and avoid getting it on the heatsink base or the heatconducting substrate of the MOSFET. I have never considered the "thermal impedance" of Locktite, but I'd assumine it's not a good conductor of heat.
 
The split one is commonly called a "split ring lockwasher"... fwiw. At least it is around this part of the NE USA...

There are also internal and external tooth lockwashers...

I've never had a problem with screws holding output devices getting loose when properly torqued and using a "tooth" type lockwasher on the nut side. I also use a small washer to spread the down force on metal tab type transistors, not on older TO-3 type metal packages.

Locking compound would not hurt... I guess... the Belleville washers are nice... but I personally have never needed to use them...

_-_-bear

PS. if only a 4-40 size was the available through hole (a plastic pak that can't be drilled) I'd
add a bar to hold the entire face of the plastic pak device as well... talking F5 sized power devices,
not <5watt plastic paks...
 
Last edited: