[sdiy] Twin diode clippers

[Andrew Simper]

Hi Brian,

Thanks for your input. Not sure how often a silicon diodes fails, but
good to know it may just be redundancy. Korg didn't seem to care about
this with their MS20 KLM 307 filter board, they have three diodes in a
row each way without interconnects:

http://machines.hyperreal.org/manufacturers/Korg/MS-synths/schematics/KLM-307.GIF

Cheers,

Andy

On 25 June 2017 at 15:47,  <rsdio at audiobanshee.com> wrote:
> I have seen and used this interconnect with LED indicators. The benefit there is that if an LED burns out, the remaining LED units will still be lit because there will still be a current path through the chain. Normally, a series strings of LEDs would all go dark as soon as a single LED goes back. A pair of parallel strings has a similar problem, because half the LEDs will go dark if one fails. Adding those interconnects.
>
> The circuit you're looking at is different from an LED chain, of course, because the diodes are back to back instead of all in the forward direction. However, I think that it still benefits from continued function in event of a failed diode. It looks like the clipping function would change if one diode went bad, but the circuit will still mostly function. Not sure how common it would be for one of the four diodes to fail, so perhaps redundancy is not actually a design goal - it just seems like a possible explanation.
>
> In other words, try running the SPICE simulation with one diode removed, to simulate a failed diode. That should show a difference in performance with and without the extra interconnect.
>
> Brian
>
>
> On Jun 24, 2017, at 8:00 PM, Andrew Simper <andy at cytomic.com> wrote:
>> In my research into diode clipping I noticed this circuit where there
>> are two diodes in each direction to soften the clipping, but there is
>> an interconnect in the centre:
>> http://www.parasitstudio.se/uploads/2/4/4/9/2449159/5838251_orig.png?487
>>
>> I did a quick LTspice simulation using this schematic to see what the
>> difference is between adding this wire and having no wire and the
>> plots looked identical:
>> www.cytomic.com/files/circuits/diode-clipper-dual.png
>>
>> Does anyone know if there is some reason to add the extra wire with
>> real world diodes?
>>