Do diodes, switching diodes or zeners go anymore leaky than transistors? - UK Vintage Radio Repair and Restoration Discussion Forum

Stylo N M · 17th Feb 2017, 5:38 pm

Hi,

As the title says, are diodes likely to become anymore leakier than transistors over time, or does it depend on their aplication and what they do?

Thanks, Paul.

ionburn · 17th Feb 2017, 9:02 pm

Without going into details I suspect that it would depend on the type and it's working environment (localised stresses). Some components are notoriously unreliable (eg tin whiskered germanium transistors). I may well be corrected in the wider case, but I have come across more leakage issues with transistors than diodes, especially when considering the total short type of leak.

Stylo N M · 17th Feb 2017, 10:02 pm

Thank you ionburn, yes that's a helpfull bit of imformation, it's always intreaging to hear what others have exspirenced over the years.

Paul.

kalee20 · 18th Feb 2017, 1:17 am

Do they go leaky, or are they always leaky?

Switching diodes do tend to be more leaky than transistors, especially Shottky diodes. Small gold-doped diodes like the ubiquitous 1N4148 are nanoamps leaky - the gold acts as recombination centres which gives a big improvement in reverse recovery time, but has the side effect of making them leaky. By comparison, a junction FET gate-channel can be used as a diode with picoamp reverse leakage.

Zener diodes - low-voltage types - have a rather soggy 'knee' and it's a bit difficult to say whether they are leaking or whether the onset of breakdown is being approached. Higher-voltage types (>15V) are more abrupt, and below the rated voltage have very little leakage.

Radio Wrangler · 18th Feb 2017, 3:46 am

Over-voltage leading to breakdown, even if the current is limited, will lead to progressively worsening leakage, reduced gain and worsening noise figure.

David

Stylo N M · 18th Feb 2017, 2:46 pm

Yes i soppose over voltage would be a major factor, it seems the humble diode other than that, seems to go on forever, unless as you say it gets overloaded.

And of course, the odd faulty one on rare accasions

Paul.

julie_m · 18th Feb 2017, 4:37 pm

Quote:

Originally Posted by Radio Wrangler

Over-voltage leading to breakdown, even if the current is limited, will lead to progressively worsening leakage, reduced gain and worsening noise figure.

What about Zener diodes, which are specifically designed to spend most of their working lives in reverse breakdown?

I've seen LEDs used as self-rectifiers on mains, and even switched mains being fed -- not exactly straight, but via only a suitably-sized resistor -- into the base of a small-signal transistor. This on supposedly long-lasting domestic appliances.

So I guess some devices must be more robust than others .....

Radio Wrangler · 18th Feb 2017, 5:30 pm

Zener diodes have rather extreme doping profiles and large areas to survive their harsh treatment.

The special small area diodes used as microwave noise sources are also operated in their avalanche breakdown mode. Their design too is special in order to stop their noise characteristic changing over time. The tiny diode region is surrounded by a ring of heavily doped material to prevent dopant migration. More info in the HP journal article on the 346A noise source, for example.

David

G6Tanuki · 18th Feb 2017, 6:13 pm

In my experience, "it depends".

As part of his MBA an acquaintance got access to internal statistics and did a costing-focussed analysis of the replaced-under-warranty failures of alternators made by Hitachi and NipponDenso (as fitted to pretty much every Japanese car for some decades). Each alternator has 9 diodes (six high-current ones to rectify the high-current 3-phase output, and three low-current ones to provide a lower 'excitation' current to the rotor/regulator).

Paradoxically, the three low-current diodes were the most-frequent-to-fail, Even so, the failure numbers were minimal compared to brush/slip-ring and bearing failures.

Personal experience: the front-end 'bridge' on SMPS often go bad particularly after thunderstorms [the trip-reclose-trip-reclose-trip-reclose-tripout cycle on a 'problem' overhead line doesn't help].

Then the next-favourite failure is in modern gear - things like TVs and monitors - where you may have several power-diodes run in parallel without current-sharing resistors. OK, if they all come from the same bandolier there's a good chance they all came from the same wafer so their characteristics *should* be pretty close so you should be able to get away without sharing-resistors.

If one diode goes S/C you need to replace all three with new ones *from the same bandolier* to get any hope of a match. Thankfully such diodes are dirt-cheap compared to the days when I advocated replacing all four Motorola bipolar transistors in a 600-Watt HF linear amplifier because one had failed after a 'high SWR' incident and this 'could have' damaged the other three. . . .

Four new transistors cost about the same as the bench-tech's weekly salary.

17th Feb 2017, 5:38 pm	#1
Stylo N M Heptode Join Date: Jul 2012 Location: Derby DE1, Derbyshire, UK. Posts: 626	Do diodes, switching diodes or zeners go anymore leaky than transistors? Hi, As the title says, are diodes likely to become anymore leakier than transistors over time, or does it depend on their aplication and what they do? Thanks, Paul.

17th Feb 2017, 9:02 pm	#2
ionburn Heptode Join Date: Dec 2014 Location: Lincoln, Lincolnshire, UK. Posts: 583	Re: Do diodes, switching diodes or zeners go anymore leaky than transistors? Without going into details I suspect that it would depend on the type and it's working environment (localised stresses). Some components are notoriously unreliable (eg tin whiskered germanium transistors). I may well be corrected in the wider case, but I have come across more leakage issues with transistors than diodes, especially when considering the total short type of leak.

17th Feb 2017, 10:02 pm	#3
Stylo N M Heptode Join Date: Jul 2012 Location: Derby DE1, Derbyshire, UK. Posts: 626	Re: Do diodes, switching diodes or zeners go anymore leaky than transistors? Thank you ionburn, yes that's a helpfull bit of imformation, it's always intreaging to hear what others have exspirenced over the years. Paul.

18th Feb 2017, 1:17 am	#4
kalee20 Dekatron Join Date: Feb 2007 Location: Lynton, N. Devon, UK. Posts: 7,086	Re: Do diodes, switching diodes or zeners go anymore leaky than transistors? Do they go leaky, or are they always leaky? Switching diodes do tend to be more leaky than transistors, especially Shottky diodes. Small gold-doped diodes like the ubiquitous 1N4148 are nanoamps leaky - the gold acts as recombination centres which gives a big improvement in reverse recovery time, but has the side effect of making them leaky. By comparison, a junction FET gate-channel can be used as a diode with picoamp reverse leakage. Zener diodes - low-voltage types - have a rather soggy 'knee' and it's a bit difficult to say whether they are leaking or whether the onset of breakdown is being approached. Higher-voltage types (>15V) are more abrupt, and below the rated voltage have very little leakage.

18th Feb 2017, 3:46 am	#5
Radio Wrangler Moderator Join Date: Mar 2012 Location: Fife, Scotland, UK. Posts: 22,880	Re: Do diodes, switching diodes or zeners go anymore leaky than transistors? Over-voltage leading to breakdown, even if the current is limited, will lead to progressively worsening leakage, reduced gain and worsening noise figure. David __________________ Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done

18th Feb 2017, 2:46 pm	#6
Stylo N M Heptode Join Date: Jul 2012 Location: Derby DE1, Derbyshire, UK. Posts: 626	Re: Do diodes, switching diodes or zeners go anymore leaky than transistors? Yes i soppose over voltage would be a major factor, it seems the humble diode other than that, seems to go on forever, unless as you say it gets overloaded. And of course, the odd faulty one on rare accasions Paul.

18th Feb 2017, 5:30 pm	#8
Radio Wrangler Moderator Join Date: Mar 2012 Location: Fife, Scotland, UK. Posts: 22,880	Re: Do diodes, switching diodes or zeners go anymore leaky than transistors? Zener diodes have rather extreme doping profiles and large areas to survive their harsh treatment. The special small area diodes used as microwave noise sources are also operated in their avalanche breakdown mode. Their design too is special in order to stop their noise characteristic changing over time. The tiny diode region is surrounded by a ring of heavily doped material to prevent dopant migration. More info in the HP journal article on the 346A noise source, for example. David __________________ Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done

18th Feb 2017, 6:13 pm	#9
G6Tanuki Dekatron Join Date: Apr 2012 Location: Wiltshire, UK. Posts: 13,998	Re: Do diodes, switching diodes or zeners go anymore leaky than transistors? In my experience, "it depends". As part of his MBA an acquaintance got access to internal statistics and did a costing-focussed analysis of the replaced-under-warranty failures of alternators made by Hitachi and NipponDenso (as fitted to pretty much every Japanese car for some decades). Each alternator has 9 diodes (six high-current ones to rectify the high-current 3-phase output, and three low-current ones to provide a lower 'excitation' current to the rotor/regulator). Paradoxically, the three low-current diodes were the most-frequent-to-fail, Even so, the failure numbers were minimal compared to brush/slip-ring and bearing failures. Personal experience: the front-end 'bridge' on SMPS often go bad particularly after thunderstorms [the trip-reclose-trip-reclose-trip-reclose-tripout cycle on a 'problem' overhead line doesn't help]. Then the next-favourite failure is in modern gear - things like TVs and monitors - where you may have several power-diodes run in parallel without current-sharing resistors. OK, if they all come from the same bandolier there's a good chance they all came from the same wafer so their characteristics should be pretty close so you should be able to get away without sharing-resistors. If one diode goes S/C you need to replace all three with new ones from the same bandolier to get any hope of a match. Thankfully such diodes are dirt-cheap compared to the days when I advocated replacing all four Motorola bipolar transistors in a 600-Watt HF linear amplifier because one had failed after a 'high SWR' incident and this 'could have' damaged the other three. . . . Four new transistors cost about the same as the bench-tech's weekly salary.