[sdiy] expo pair "heat shield?"

[Neil Harper]

On 10/11/22 14:25, Mattias Rickardsson wrote:

> If there are very nearby objects with strong pulsating heat, which
> affects one transistor more than the other, then a heat shield might do
> some good. There probably aren't! (-:
>
> (I assume that "shield" means metal foil or some other reflective
> material, not just any material that is in the way but absorbs incoming
> heat)
>
> What's most important is to keep the pair at identical temperatures,
> which is done by having substantially less thermal resistance between
> the transistors compared to between the pair and their environment. But
> it's important that they are not totally isolated - if the contact
> between pair and environment is very weak, such as using a shield that
> reflects heat from the transistors back on themselves, then there will
> be increased self-heating effects with different temperatures depending
> on the currents through the transistors. Not optimal. I guess the
> component legs would give enough contact in most normal cases though.
>
> Making some kind of blob around the pair in order to increase thermal
> inertia... Well, might be somewhat useful, but also potentially give
> some of the self-heating problems mentioned above. Keeping them at one
> constant temperature is the second important thing.
>
> Don't forget to optimize the inter-transistor contact first. Thermal
> paste between them and something like shrink tubing around them should
> do. :-)
>
> /mr



thanks matthias, that's some good advice..

i'm actually using a matched-pair on a single tiny SOT363 IC that I
managed to hand solder. the tempco resistor just goes right overtop of
it, in direct contact. so far I haven't added any additional thermal
paste between the junction or done anything else.

I thought maybe adding the hot-snot would isolate it from actual
convection currents inside the eurorack enclosure.

but i suppose the most important thing is just that entire enclosure get
up to an "operating temperature" and remain there, and that the
transistor pair and resistor also get up to this same temperature. I
suppose if it's 'insulated' in any, it is still destined to eventually
reach that operating temperature, but might just take longer to get
there.. which isn't any benefit. and good call on insulation possibly
keeping any self-generated heat in, instead of letting it dissipate..



--
/// Neil Harper
/// Every Wave is New Until it Breaks