RE: insuring the stability of power delivery
Brian and all,
see my comments below.
Thanks
Yair.
===================================================================
[Brian Lynch] I agree that during startup (Inrush in the PSE; when
the
input drops to zero when the PD switch closes and then increases as
the PD bulk charges) that this is not an oscillation.
The problems I saw with the technique were that to insure startup:
1) The PD cap size needs to be limited
2) The DC/DC must be held off until the cap is charged
3) The startup/charging time needs to be limited
4) The PD needs energy storage to hold on the switch when
the input goes away.
Yair Darshan: The problems you saw when the inrush current is in the
PSE are actually not a real problems and I try to explain why:
1) The PD cap size needs to be limited :
Correct, the input cap needs to be limited.
There is no need for higher cap than 470uF for more than 95% of the
applications that we see using the Power over MDI technology. Standard
switching power supply working at 100Khz and supporting 12-15W can work well
with 470uF at their input with large margin.
This statement is tested and verified in the field many times.
Other applications that needs more than 470uF will be supported by
the PD inrush current limiter.
The point is that there is no reason to increase the PD cost for for
95% of the application while we have the hardware to do this function in the
PSE for protection, current sensing switching the port to off etc.
2) The DC/DC must be held off until the cap is charged
Correct, however, you need to do it anyway if you want to increase
your design reliability by preventing the power supply from turning on at
low voltage and drawing much more current that it needs. I=P/Vin, P is
constant....
You need this function to ensure startup and working on the stable
region as I showed in earlier presentations.
Most of the commercial PWM controllers have the UVLO function any
way for the above reasons.
3) The startup/charging time needs to be limited
Why it is a problem? In light of May discussion it seems that we
want to decrease the time for single port insertion and this is one of the
components that sets this time.
If for special application we want to have more time, its fine, add
the inrush current at the PD, however it is not the problem
of 95% of the applications.
In standard DC/DC supply for 12-15W with 470uF max at its input the
total charging time is around 50mSec and the power supply startup time is
10-30mSec thus the PD can be fully functional within 100mSec max.
4) The PD needs energy storage to hold on the switch when
the
Correct, add 0.1uF between the gate and source of the Isolating
mosfet switch and utilize the high impedance of Mosfet gate
to have long memory effect. (I have circuit that is working and
tested in the lab, I can send it to you if you want)
(The inrush current limiter that you want to add to the PD contains
much more components than the needed to solve the storage time issue)
Brian:
My thoughts are that even though we can put limits on these
parameters in
the spec, they are four parameters that are not needed if inrush
limiting
is put in the PD.
We have limits all the time, and limits on the design are needed
even if the inrush current limiter is in the PD.
And the reasons are:
1) The PD cap size needs to be limited :
The inrush current limit set point is needed to be calculated
according to the PD power supply input capacitor along with the the PD power
supply normal operating current right after startup. Otherwise you will have
startup problem.
You can argue this statement by saying, that in your application or
model you haven't noticed that problem, and it could be true
for a specified sets of parameters... but not for any set of
parameters.
The bottom line is that you move the definitions and limitations to
the PD box instead of to the System definitions.
I prefer the system definitions and not complicating the PD in order
to make it simple and low cost.
2) The DC/DC must be held off until the cap is charged
In many power topologies and/or applications, you need to do it
regardless if the inrush limiter is in the PD.
3) The startup/charging time needs to be limited
If you need long startup for some reason you can add the hardware to
the PD to get it. However it is not a drawback for 95% of the applications.
Brian:
The only conceptual issue I see with inrush limiting in the PSE is
that if the PD is not capacitive at its input, there could be
oscillation
on the cable. A non capacitive input is unlikely, but possible if
the PD
does not have a DC/DC converter input stage.
95% of the applications are with capacitive input since they have
DC/DC.
Why we should care for the 5%.
In any case we cant spec. for 100% of the applications (the known
today and the future) since it will not be a cost effective solutions.
===================================================================
{Brian Lynch] In my simulations, I could put in components which
would make the system fail to start. Only by putting in extra
constraints on component values could I guarantee startup.
Isn't it a normal design procedure that any design will fail with
some set of numbers and will work perfectly with other set of numbers? I am
sure that if you put the inrush current limit in the PD, it will fail for a
some set of numbers.
From the experience that we gathered from the field, for the case
that the inrush current was in the PSE, the PD designers find that it is
easy to design the PD power supply if they know the energy source parameters
during startup and during normal operating mode and the only limitations
that they have on the PSE and PD where:
PSE side:
1. Max. available power during normal operating mode.
2. Voltage range
3. Max. average current at the lowest line possible.
4. Peak current and time from the PSE during startup.
PD side:
4. Max. PD input cap
5. Turn on and turn off voltages
Very simple and it works always.
I agree that for very large PD input capacitor the Inrush current
limit should be in the PD, and we already agree to this concept in St Louis.
We need to close the value of this capacitor in which it consider to be
large capacitor.
======================================================================
> -----Original Message-----
> From: Lynch, Brian [SMTP:brian_lynch@xxxxxx]
> Sent: ?, ???? 01, 2001 4:51 PM
> To: 'Yair Darshan'; 'Dave Dwelley'; Lynch, Brian; 'Rick Brooks';
> stds-802-3-pwrviamdi@xxxxxxxx
> Subject: RE: insuring the stability of power delivery
>
> All,
>
> See below
>
> Brian
>
> >-----Original Message-----
> >From: Yair Darshan [mailto:YairD@xxxxxxxxxxxxxx]
> >Sent: Thursday, May 31, 2001 10:18 PM
> >To: 'Dave Dwelley'; Lynch, Brian; 'Yair Darshan'; 'Rick Brooks';
> >stds-802-3-pwrviamdi@xxxxxxxx
> >Subject: RE: insuring the stability of power delivery
> >
> >
> >Dave,
> >See my comments below.
> >Yair.
> >
> >
> >> -----Original Message-----
> >> From: Dave Dwelley [SMTP:ddwelley@xxxxxxxxxx]
> >> Sent: ?, ??? 31, 2001 11:52 PM
> >> To: Lynch, Brian; 'Yair Darshan'; 'Rick Brooks';
> >> stds-802-3-pwrviamdi@xxxxxxxx
> >> Subject: RE: insuring the stability of power delivery
> >>
> >> This is a new wrinkle! I'm not a distributed power supply
> >designer, but
> >> I'll take a stab at this anyway. Shout me down if I don't
> >know what I'm
> >> talking about.
> >>
> >> At 11:02 AM 5/31/2001 -0400, Lynch, Brian wrote:
> >> >In our work with power systems, we have found that to insure
> >> >stability of a system, the closed loop output impedance of the
> >> >sourcing supply must be lower than the closed loop input impedance
> >> >if the load power supply over the frequency band and voltages
> >> >of interest. This concept is also discussed and taught at various
> >> >universities. VPI and MIT are two I have direct knowledge of.
> >>
> >> I see this as two problems - startup and steady state.
> >>
> >> In startup, the output impedance of the PSE power supply is
> >what it is,
> >> and
> >> the input impedance of the PD appears in series with
> >whatever is doing the
> >>
> >> inrush limiting, be it in the PSE or the PD. The PSE power
> >supply isn't
> >> likely to oscillate in this case. The line voltage is
> >another matter -
> >> presumably if the limiter was in the PSE, the line voltage
> >could oscillate
> >>
> >> (by Brian's theory);
> > [Yair Darshan] I do not agree to this assumption. If
> >the limiter is
> >in the PSE, the PD input voltage can go to zero for some time
> >and then ramp
> >again.
> > This is not an oscillation it is a normal short circuit
> >condition
> >when a voltage source is applied to a discharged capacitor.
> > With other simple implementation of the isolating switch, you
> >prevent this condition. In any case this condition is not a problem.
> > The case I present here was tested and confirmed in lab and also
> >with simulations. I think (and we need to check with Brian) that he had
> >implementation problem and not a conceptual problem.
>
> [Brian Lynch] I agree that during startup (Inrush in the PSE; when the
> input drops to zero when the PD switch closes and then increases as
> the PD bulk charges) that this is not an oscillation. The problems I saw
> with the technique were that to insure startup:
> 1) The PD cap size needs to be limited
> 2) The DC/DC must be held off until the cap is charged
> 3) The startup/charging time needs to be limited
> 4) The PD needs energy storage to hold on the switch when the
> input goes away.
>
> My thoughts are that even though we can put limits on these parameters in
> the spec, they are four parameters that are not needed if inrush limiting
> is put in the PD.
>
> The only conceptual issue I see with inrush limiting in the PSE is
> that if the PD is not capacitive at its input, there could be oscillation
> on the cable. A non capacitive input is unlikely, but possible if the PD
> does not have a DC/DC converter input stage.
> ===================================================================
>
> >> if it was in the PD, the line would stay quiet, but
> >> the node inside the PD between the inrush-limiting pass
> >device and the
> >> input of the PD supply could oscillate. The PD designer can
> >beat this by
> >> holding the switcher inactive until the input cap has
> >reached its final
> >> value, or by balancing impedances properly.
> > [Yair Darshan] Again, in my opinion you are discussing
> >a problem
> >that is not exist. Can any body send a detailed circuit
> >showing a problem ?
> {Brian Lynch] In my simulations, I could put in components which
> would make the system fail to start. Only by putting in extra
> constraints on component values could I guarantee startup.
> ======================================================================
> >