Speaking in general terms, and not on this kit in particular:

Yes, driving the FET on a boost switcher, directly from a uC IO pin, is a 
bad idea. Not only, are they not fully driven ON, but you can kill the uC. 
I've actually done it. I think Nick warned us about this, many, many years 
ago. There is capacitive coupling from the drain to the gate. If the uC 
gets a "reset", while that FET is ON, or more destructively, when It just 
turned OFF, the HV pulse can couple over to the gate. Normally, it sees a 
relatively low impedance load, from the circuit driving the FET. But at "uC 
reset", all IO pins revert to inputs, presenting a high impedance. In 
short, use a FET driver, either one "off the shelf", or made up of both 
sourcing and sinking transistors, that always present a low impedance. Like 
I said I killed a couple of those 8-pin Tiny12 uCs, driving an IRL640 FET, 
directly. Its a logic level FET. But they died (the uC) by me just 
resetting the circuit.

But this particular clock kit, does not do this. It has a separate MC34063 
based power supply. I refer to the circuit used, as a MKII design. The FET 
is not driven directly from the MC34063, but thru a push-pull PNP/NPN 
emitter follower pair of transistors. There is also the MK1.5 design, which 
uses the MC3406s's own transistor to pull the signal up, and an added PNP 
to drag it back down. Both the MKII and MK1.5 have active drive, both 
sourcing and sinking. There is also the MKI design, with passive pull down 
(a simple resistor), but I'd avoid that one for all but the lowest amount 
of power delivery. Though the MKII, and MK1.5 are old designs, using a 
really old chip, they are adequate for most nixie clocks applications. They 
should only get modestly warm delivering 25mA or less. I've tested some to 
45mA (at 170V), and they can get quite "toasty", at just around 65C, at the 
coil.

Here's the manual of the kit in question:
http://www.pvelectronics.co.uk/kits/LTC1040/LTC1040_v4.pdf

The schematic, is on the last two pages of the 30 page document.

As stated by many earlier, there is probably nothing wrong with the unit, 
and what is observed, is the just the normal expected amount of heat. Maybe 
the OP could slap a thermoelectric cooler on the FET. I'd just leave it 
alone.

On Monday, October 13, 2014 5:34:05 AM UTC-7, joenixie wrote:
>
> Now for my two cents. I have not looked at the schematic due to time 
> issues, but I know I ran into issues driving a standard mosfet with a logic 
> level gate signal. It got hotter than blue blazes! If the gate signal only 
> goes to 5 volts, you must use a mosfet that will turn on fully at this 
> level otherwise, it acts like a power resister.
>
> -joe
>
>
>

-- 
You received this message because you are subscribed to the Google Groups 
"neonixie-l" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send an email to [email protected].
To view this discussion on the web, visit 
https://groups.google.com/d/msgid/neonixie-l/59cb4dd3-3700-4d9e-aa4f-07e8711d08ac%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to