[sdiy] SSM2164 last time buys

[Oscar Salas]

Hello Steve,

> 1. Put in the resistor but add a largish decoupling cap at
> the supply pin to reduce the impedance at the frequencies of
> interest (yuk)

Well as I asked before to Neil, how the capacitor can reduce the DC impedance on there?

> 2. Replace the resistor with a simple constant-current
> source set above the maximum normal operating current but
> below that which causes damage. The source would operate in
> saturation most of the time (because the current drawn by
> the 2164 would be less than the set current of the source)
> but, under fault conditions, would limit the supply current
> and protect the 2164.

This works!, I just breadboard it with a lmm334 that I had. The device provides 10mA. Will be more adequate something as 20mA but for the tests but was OK. 

> 3. Give each 2164 considered to be vulnerable its own
> supply voltage regulator, choosing regs with a low enough
> current limit to protect the chip.
> I have seen option 3 used in pro audio equipment where
> balanced line drivers are prone to latch up and release the
> smoke from the output device (SSM2142 or similar) - rather
> than power all the outputs from a single pair of 1A
> 7815/7915s, they were divided into groups each fed by
> 100mA-rated 78L15/79L15s.

Interesting. This is a good option, maybe more difficult to implement than the other possibilities.(in a modular synthesizer)

> In this application I would go for option 2 though.

Yes it is good! only the resistor will be better if it doesn't affect the 2164 performance because its simplicity. The LM334 needs a resistor.
I sent a mail asking both AD and coolaudio, lets see the answer.
Meanwhile, could we determine it? what tests would be appropriate? 

Anyway I did a test measuring current on positive supply.
V2164 in class A operation. 
Input signal +15V in four VCAs
Then I measured the current with the following settings in the control voltage input of the four VCAs (a)0V (b)+1V (c) +2V and as reference (e)-15 but without connect negative supply.

The purpose of this test is see what is the difference of current of the standard operation(without protection) and the other protective solutions. I don’t know if it reveals important thing .What do you think?
All amounts are mA:
For standard operation..a=8.18 b=5.24 c=5.14
with 1K resistor........a=8.02 b=5.18 c=5.09 e=13.2
with 510 resistor.......a=8.11 b=5.21 c=5.12 e=24.9
with 330 resistor.......a=8.14 b=5.22 c=5.13 e=36.7
with 100 resistor.......a=8.18 b=5.24 c=5.14 e=94.1
Lm334 providing 10mA....a=8.17 b=5.23 c=5.14 e=9.8

--- On Tue, 3/1/11, Steve Lenham <steve at bendentech.co.uk> wrote:

> From: Steve Lenham <steve at bendentech.co.uk>
> Subject: Re: [sdiy] SSM2164 last time buys
> To: synth-diy at dropmix.xs4all.nl
> Date: Tuesday, March 1, 2011, 11:46 AM
> I would instinctively shy away from
> putting a non-negligible value of resistance in the supply
> to any IC - it converts the naturally varying supply current
> into voltage modulation, which may or may not cause problems
> or feed through to outputs. But I think you are on the right
> lines...other options would be:
> 
> 1. Put in the resistor but add a largish decoupling cap at
> the supply pin to reduce the impedance at the frequencies of
> interest (yuk)
> 
> 2. Replace the resistor with a simple constant-current
> source set above the maximum normal operating current but
> below that which causes damage. The source would operate in
> saturation most of the time (because the current drawn by
> the 2164 would be less than the set current of the source)
> but, under fault conditions, would limit the supply current
> and protect the 2164.
> 
> 3. Give each 2164 considered to be vulnerable its own
> supply voltage regulator, choosing regs with a low enough
> current limit to protect the chip.
> 
> I have seen option 3 used in pro audio equipment where
> balanced line drivers are prone to latch up and release the
> smoke from the output device (SSM2142 or similar) - rather
> than power all the outputs from a single pair of 1A
> 7815/7915s, they were divided into groups each fed by
> 100mA-rated 78L15/79L15s.
> 
> In this application I would go for option 2 though.
> 
> Cheers,
> 
> Steve L.
> Benden Sound Technology
> www.bendensound.co.uk
> 
> 
> On 01/03/2011 00:51, Oscar Salas wrote:
> > Hey Tim thanks,
> > Yes I have the same question, it would be great the
> solution just with one resistor. If not the other solution
> would be the resettable fuse. Lets see more opinions!
> > 
> > --- On Tue, 3/1/11, Tim Ressel<madhun2001 at yahoo.com> 
> wrote:
> > 
> >> From: Tim Ressel<madhun2001 at yahoo.com>
> 
> >> Excellent work Oscar! This will be a
> >> relief to many I'm sure. The only question
> >> I have is does the current limiting resistor
> affect normal
> >> operation?
> >> 
> >> --TimR
> >> 
> >> 
> >> ----- Original Message ----
> >> From: Oscar Salas<osaiber at yahoo.es>
> >> To: Synth DIY<synth-diy at dropmix.xs4all.nl>;
> >> 
> >> As a summary I would say:
> >> 
> >> A) With a resistor in series with the positive
> supply the
> >> chip is protected
> >> restricting the current. I would say that a value
> about
> >> 510Ω or 1K would do the
> >> job.
> >> 
> >> Question: There is any inconvenient in use a
> resistor?
> >> could it affect the
> >> performance of the VCAs
> >> 
> >> B) I didn't test it but another solution would be
> a
> >> resettable fuse but in this
> >> case stay within 10mA or 50mA per chip. Maybe
> better one
> >> fuse per chip.
> _______________________________________________
> Synth-diy mailing list
> Synth-diy at dropmix.xs4all.nl
> http://dropmix.xs4all.nl/mailman/listinfo/synth-diy
>