[sdiy] Todays weekly Elektor email...

[Noah Vawter]

thanks for the reality check.

I think the spec I was concerned into about is "Output Ripple Voltage."

in the spec sheet for my LM2574, for example, it shows a peak to peek  
ripple voltage of about 24 mV when operated at 5V.
That's more like 0.024V/5.0V = 0.48% of the output is ripply.  Note,  
the ripple signal is an asymmetric  triangle wave with a
fundamental of 52kHz, so that's probably not normally  
audible.....................   and definitely filterable...


On May 13, 2011, at 4:12 PM, Harry Bissell wrote:

> No. Load regulation means that the nominal value of DC stays within  
> 5% from no-load
> to full-load.  Now, how much does your load actually change  
> (probably not much unless
> you are strobing LEDs on and off...  Noise is another thing, that  
> would be an AC signal
> riding the DC voltage.
>
> Good filtering can probably quelch the switcher noise, imho. Unless  
> you're an audiophile :^)
>
> H^) harry
>
>
> ----- Original Message -----
> From: Noah Vawter <nvawter at media.mit.edu>
> To: synthdiy diy <synth-diy at dropmix.xs4all.nl>
> Sent: Fri, 13 May 2011 15:28:45 -0400 (EDT)
> Subject: Re: [sdiy] Todays weekly Elektor email...
>
>
> On May 13, 2011, at 2:57 PM, Tom Arnold wrote:
>> http://www.elektor.com/news/k78xx-switch-mode-lm78xx.1806960.lynkx
>>
>> I've seen these parts before, not this one in particular but in
>> general.
>> Has anyone used them for synth work?  If they are quiet enough they
>> would be
>> a great addition for supply +5 on modules to get rid of those little
>> 7805
>> spaceheaters I normally use...
>
> I would take their "up to 97% efficiency" with a huge grain of salt.
> From what I have seen typical maximum efficiencies for small switchers
> are less than 90%.
>
> Also, the efficiency will vary considerably, e.g. 50%-90% depending
> upon the voltage and current you're running it at,
> relative to input voltage.  It goes down with larger input voltages,
> and up with larger output currents.  Finally, the maximum
> efficiency will be obtained through careful component selection.  In a
> system like their resin-covering, you can't do that anymore.
>
> I definitely agree it's time to upgrade from the 7805 era.
>
> Noise for this device (Load regulation) says less than 5%.  I take
> that to mean @5V:
> output is from 4.75V- 5.25V.  So, noise is 0.5V.  That is
> approximately 1/10th the output....   20 * log10(0.5/5) or about
> -46 dB of noise.  For analog audio work that is out of the question,
> no?   That seems very, very bad.....
> ....Am I doing this right?
>
> FWIW, I've been using a chip called LM2574 for a few years now for
> digital supplies.  It requires minimum of components and comes in DIP.
> It's also very tolerant of layout.
>
> I'm thinking... if waste heat is the problem (I assume you're turning
> ~18V into 5V?) , perhaps you could switch regulate down to
> the drop out of your 7805, e.g. 7.0V, then use a 7805 linear in series
> with that.  That way the 7805 cleans up the noisy switched output
> without wasting huge amounts dissipating higher voltages.  Opinions?
>
> _______________________________________________
> Synth-diy mailing list
> Synth-diy at dropmix.xs4all.nl
> http://dropmix.xs4all.nl/mailman/listinfo/synth-diy
>
> -- 
> Harry Bissell & Nora Abdullah 4eva