Re: "Goosing" with a higher voltage

On Apr 28, 9:54 am, John Popelish <jpopel...@xxxxxxxx> wrote:
Joerg wrote:
John Popelish wrote:
eromlignod wrote:
Hi guys:

I have an application where I'm driving a device with 5 volts DC.
From time to time it is necessary to give it a short pulse of 24 Vdc
then revert back to 5V.

I have a power supply that has both 5V and 24V outputs.  Can I just
pulse the 24V right onto the 5V line (with a power transistor or
relay) as long as I include a diode on the 5V output to avoid back
current?  Or is this foolish?

It sounds like it could work.  I would use a Schottky diode to
minimize the voltage loss when the 5 volt supply was providing the
load with current.  I even that loss is too much, you might use a
MOSFET as a synchronous rectifier (with the channel conducting in
parallel with the body diode when the 5 volt supply was driving).  But
then, you have to produce a gate drive signal that is synchronized
(non overlapping) with the disabling of the 24 volt drive control.

Be careful. If there are any bypass caps downstream the voltage on them
might still be higher than 5V when the 5V sync-rectifier comes back on
-> phssst ... *POOF*.

Yep, that was one of the reasons for the non overlapping
warning.  I would probably turn the MOSFET on only after
the body diode was carrying the load during the 24 volt
switch off, and turn the MOSFET off before turning the 24
volt switch back on.  That assumes the system could handle
those brief sags below 5 volts.

--
Regards,

John Popelish- Hide quoted text -

- Show quoted text -

I'd say Don should use OR-control, that way nothing can go wrong.
He already has 24 volts to drive a high-side mosfet.


24v
|
+-----+----,
| | |
| | |
|\ | 6k2 |
,------|+\| | |
| | \ | |
| LM393| >----+ |
| | / | o|
| ,--|-/| | |--power
| | |/ | | o| switch
10k 10k gnd | |
| | | |
| '-----------|----+
| | |
| ,------+---+ |
| | | | |
| 18v | |_ |
| zener 10k __ |
| | | || |
| 5v-+------+---''---+--load
| | s d
| | n-channel mosfet
| \ SPP80N03
| /
'--->\1k
/
\
|
|
50k
|
|
gnd

I didn't draw a .1 uF bypass cap for the comparator, but it should
probably have one.
The mosfet needs oriented source to the 5 volt supply and the drain to
the load.
The 1k pot nulls out any negative offset in the comparator, to make
sure it trips when the source voltage exceeds the drain voltage,
preventing the mosfet from ever conducting back into the 5 volt
supply. If the comparator has positive offset, it will oscillate at
light loads under a few tenths of an amp or so, which probably won't
hurt anything. But if one prefers, using an op amp instead of a
comparator would get around that. I like the comparator 'cause it's
fast, cheap and robust.
.

Relevant Pages

  • Re: Synchronous rectifier
    ... comparator sense the Vds of a MOSFET to control the turning on and off ... So the comparator is a zero-crossing detector. ... But in very light load/zero load, ...
    (sci.electronics.design)
  • Re: "Goosing" with a higher voltage
    ... load with current. ... MOSFET as a synchronous rectifier (with the channel conducting in ... parallel with the body diode when the 5 volt supply was driving). ... The 1k pot nulls out any negative offset in the comparator, ...
    (sci.electronics.basics)
  • Re: Current Transformer to 0-10 Volts
    ... resistor. ... If you used your original 10:1 ratio, the ten volt drop would reflect ... Then you need to convert the output to whatever the PLC ... additional transformer load. ...
    (sci.electronics.design)
  • Re: Drivers with more source current than LM5112?
    ... If you are looking at the last graph on page 6, ... current delivered or drawn from a 5 volt source, ... sink, source currents. ... and P-channel single-chip soic-8 MOSFET ICs are plentiful. ...
    (sci.electronics.design)
  • Re: 12V Electronic Tranny
    ... use with domestic 12 volt lighting. ... The effect employs a 12 volt, 100 watt dichroic lamp plus an 80mm, 12 volt fan blowing across PSU and lamp. ... Initially, I imagined that the 60 VA rating would be inadequate for the 100 watt lamp load, but on test it proved to work fine with very little heating of any of the parts - even with no fan cooling. ... I would be concerned about an SOA or peak power abuse of the transistors though. ...
    (sci.electronics.design)