[time-nuts] Super Regulator links

[Dave Brown]

----- Original Message ----- 
From: "Bruce Griffiths" <bruce.griffiths at xtra.co.nz>
To: "Discussion of precise time and frequency measurement" 
<time-nuts at febo.com>
Sent: Friday, December 14, 2007 4:36 PM
Subject: Re: [time-nuts] Super Regulator links


> John Miles wrote:
>> What are some of your favorite low-noise regulators?  When I made 
>> the remark
>> about the uA723, I was specifically thinking of its performance 
>> compared to
>> the LT1762.  With Cref=5 uF, the uA723's output noise voltage is 
>> rated at
>> 2.5 uV from 100 Hz to 10 kHz.  The LT1762, which seems to be among 
>> Linear's
>> quietest parts, is rated at 20 uV from 100 Hz to 100 kHz, with 
>> external
>> bypassing that places most of the noise at lower frequencies. 
>> Snake oil or
>> not, that's 18 dB less noise from the 723.
>>
>>
> This is perhaps a result of using the apparently inherently noisier
> bandgap style references which offer significant advantages (lower
> minimum input voltage, lower dropout but not necessarily lower 
> noise) in
> lower voltage systems.
>> There are various hacks like Wenzel's that can clean up after a 
>> noisy
>> regulator
> That particular circuit is only effective over the 100Hz to 10kHz 
> range.
> The brute force darlington buffered RC filter favoured by NIST is 
> far
> more effective when the  accompanying disadvantages are acceptable.
>> , but if there are quieter fully-integrated solutions out there I'd
>> like to hear about them.  Posted back to the list in case there are 
>> other
>> views on the subject...
>>
>> -- john, KE5FX
>>
>
> I havent yet come across any IC that seems to approach the 
> performance
> achievable with a 723.
> The next step up without adding too much complexity is perhaps to 
> use a
> 723 with a low pass filtered LM329 reference powered from its 
> regulated
> output.
> To improve the ripple rejection you can add an LM317 tracking 
> preregulator.
> Beyond that a well designed discrete regulator is the only way to
> improve performance.
> Off course if you are building hundreds of circuits and dont want to
> take the risk of occasionally striking noisy parts then you either 
> need
> to prescreen parts or buy noisier parts with guaranteed noise specs.
>
> Lots of so called low noise regulators pop up from time to time but 
> when
> you actually read the datasheet and calculate the output noise for a 
> 12V
> supply and not the 1V or so supply to which the noise specs apply 
> they
> all seem to fall short of the 723's performance.
>
> Bruce
>

If the prime energy source is a car battery - or indeed any battery 
based system -  then what other options are there than to use a 
switcher - at least in the first instance- when generating +5, and 
plus/minus 12, etc. for precision timing equipment?
In this case, I assume the use of  separate linear regulators based on 
the 723, would be the best way to clean up the dc supply rails?
DaveB, NZ