[time-nuts] huntron tracker advice & troubleshooting without schematic advise

[Magnus Danielson]

Hi Jeroen,

Jeroen Bastemeijer skrev:
> Hi Magnus,
> 
> I totally agree with you. But doing a quick repair is more an art than a 
> science ;-) By having some feeling for common faults it is easier to 
> locate them.

Certainly, and I totally agree. Repairing electronics requires 
experience, and you need to gain experience in many diverse fields.
Experience include typical error behaviour of components, designs and 
also knowing typical design types helps. Also learning to "cheat" in a 
safe way can also be rewarding.

> The Huntron tracker was of use for equipment which was almost impossible 
> to service when under power. For example power-electronics with high 
> voltages and nice little radios which consisted of a stack of boards, 
> here it was impossible to probe around. The only solution was soldeing a 
> wire to a component, assemble the whole thing, switch on measure... for 
> testing another point, the whole procedure had to be repeated. Ok, for 
> some equipment we used extension cables, but it was not always possible 
> to use them.

These days the 3,3 V is becoming more and more a high voltage power rail 
from which you can switch down to suitable voltages. 5 V is becoming 
scary high and 12 V or +/- 15 V is just unheard of. Sigh.

> Some failures are very difficult to track: Firmware related problems. 
> The only solution to that, was finding a working unit and copy the 
> firmware to the faulty machine (e.g. EEPROM copying)....

Making backup images of EPROMs etc. is a good thing. While the actual 
EPROM may not be faulty, even single bit errors may be the cause of much 
greif.

> But as I indicated in the beginning: Repairing is more like an art than 
> a science. Over the years a whole bunch of tricks and strategies was 
> developed to locate the faults and fix them. Repairing needs a solid 
> background in electronics and a lot of experience.

Certainly. I find myself being pulled into the lab every now and then to 
aid a debug-hunt. It seems that I am looking a little more out of the 
box than they do... pulling a spec in and point on the error or just 
turn the timebase knob to zoom in on the spurious signal...

> But: When you manage to repair something, it usually gives a lot of 
> satisfaction! ;-)

Certainly. Also, the rarer or more expensive it is, the greater 
satisfaction. Also beating someone to it can be gratifying.

A common mistake is to invent too complex fault causes and try to proove 
those. It is usually a much simpler cause, but the engineer keep 
inventing complex causes.

Cheers,
Magnus