I haven't had any hands on experience with the problem you describe, but
I have studied the theoretical problem in conjunction with a system we
have under development.

All loads, of course, will have a given current step size and transition
time.  With the new low-voltage technology, this has resulted in a
severe problem for power subsystem designers.  The largest and fastest
load steps are probably seen with Intel's microprocessors, various types
of DRAM and bus terminations like GTL+.  Intel has an application note
out describing the problem and solution for their processors and GTL+.
AT&T (now Lucent) produced an excellent paper a few years back that
described the problem more generically and they developed some products
to solve the problem of fast current transients.  

With regard to 5V technology, the problem could still, obviously exist,
but it would be very mild compared to the huge, monstrous transients
that you get with 3.3V and 1.2V IC's, for example.  My guess is that
your 5V system will probably run OK even with the 1V pulses.  However,
the fact that they are at such a high frequency indicates that they
would be easy to filter out with the appropriate use of capacitors.  A
potential problem with capacitors, however, is that they can destabilize
the feedback loop of the converter.  So, you would probably have to
install these on a trial and error basis.  Both total capacitance and
ESR (equivalent series resistance) are factors in destabilization.  With
a long cable, my guess, is that the resistance and inductance of the
cable would tend to cancel out the capacitance and ESR problem.

In a worst-case scenario, you might be able to switch vendors on the
DC/DC converter.  Astec America has a standard product with fast
transient response capabilities for 3.3V.  They could undoubtedly modify
a 5V converter to work with your load also.  Astec's phone number is
760-930-4713.

Max Kelson
[email protected]
Evans & Sutherland Computer Corp.

> -----Original Message-----
> From: Bob Tims [SMTP:[email protected]]
> Sent: Tuesday, January 06, 1998 8:21 AM
> To:   [email protected]
> Subject:      Noise in Power Lead
> 
> 
>   Happy New Year All,
> 
> I have an interesting noise problem brought to me. 
> The noise is a 500-900 mV pulse seen at 50 MHz on a 
> power lead from a DC to DC converter (48V to 5V) to a 
> VME cage. 
> There were two scenarios: first, both the power supply
> and the cage are earth-grounded, and the pulse is seen 
> at both the power supply and the cage. Second, the cage 
> is left floating, and the pulse is only seen at the 
> power supply.
> The noise seems to be a ring, and the leads between the 
> supply and cage are fairly long.
> Has anybody seen something like this before, or does anybody 
> have any suggestions or comments?
> We have several ideas, but not much time.
> The frequency and the fact that the pulse reaches almost a volt 
> concerns us greatly.
> Thanks in advance,
> 
> Bob Tims
> Standards and Compliance
> [email protected] 
> 
> The opinions expressed in this correspondence in no way reflect
> the opinions of Ericsson Inc.     
> 
>   
>   RCIC - http://www.rcic.com
>   Regulatory Compliance Information Center
>   
> 

Reply via email to