Question and Answer Forum

Why two transistors?

This forum post has been removed

Number of replies: 16
The content of this forum post has been removed and can no longer be accessed.
In reply to Deleted user

Re: Why two transistors?

by Mark Hughes -
Hi Peter!
That's a great question....for Dr. Bob :) He's working on writing up why he designed things the way he did. Just give us a little time on this one.
Thanks!
Mark

33 words

In reply to Deleted user

Re: Why two transistors?

by Jesse Robinson -
I had an answer written up, but I think I misunderstood the question. Are you talking about just Q204 entirely, or just Q204 for the battery voltage but Q201 and Q202 are still present on the USB voltage?

If you were questioning Q203 with Q201/Q202/Q204 still present, I am wondering that myself.

54 words

In reply to Jesse Robinson

Re: Why two transistors?

by Mark Hughes -
Let's get a discussion going -- I think I sent Bob three or four emails about this circuit when I first saw it. If I have questions, and others have questions, let's get them out there and start a dialog :)

39 words

In reply to Deleted user

This forum post has been removed

The content of this forum post has been removed and can no longer be accessed.
In reply to Deleted user

This forum post has been removed

The content of this forum post has been removed and can no longer be accessed.
In reply to Deleted user

This forum post has been removed

The content of this forum post has been removed and can no longer be accessed.
In reply to Deleted user

Re: Why two transistors?

by Jesse Robinson -
There's a switch to disconnect the battery when use isn't desired.

11 words

In reply to Deleted user

Re: Why two transistors?

by Jesse Robinson -
I think it looks good, only concern would be the transient time between USB plugin and regulator shut down. The USB should come up quick and the Enable should go low between 0.8 and 1.5V on the USB ramp but can't find any info on the timing of the enable/shutdown in the datasheet. The other circuit was kind of break before make which is nice for that kind of thing, but this is much less complicated.

77 words

In reply to Jesse Robinson

This forum post has been removed

The content of this forum post has been removed and can no longer be accessed.
In reply to Deleted user

This forum post has been removed

The content of this forum post has been removed and can no longer be accessed.
In reply to Deleted user

Re: Why two transistors?

by Mark Hughes -
You know -- it might be interesting if we could get a simulation of the circuit block -- I know I wouldn't mind seeing from the 5V0_RegOut net to the 5V0 net. I don't mind doing it, but I'm a bit tied up writing content at the moment. Do we have any takers out there? I might be able to scrounge up a badge for you.

65 words

Attachment PowerSelectionBlock.jpg
In reply to Mark Hughes

Re: Why two transistors?

by Jesse Robinson -
I'm using TINA and not sure how to import that PSPICE model into TINA, so I found a transistor with similar RDS and VGS from TI's catalog to sim (CSD25213W10). I've attached pictures and the simulation, I used a 50 Ohm load to simulate a 100mA draw.

I used a piecewise linear voltage generator to simulate the supplies going up and down
Signal Definitions
ABATREG - current from the VBATREG supply
AUSB - current from the VUSBY supply
GATEBAT - the gate voltage for the USB control transistors Q201 and Q202
GATEBAT - the gate voltate for the battery supply control transistors Q203 and Q204
VBATREG - the regulated voltage from the battery
VOut - the output voltage from the power switch circuit
VUSB - the voltage from the USB supply

I started with VUSB and VBATREG off, the output was off.  Then I turned on VUSB, this allowed VOUT to come up and VUSB to supply the load.  With VUSB up I brought up VBATREG, this allowed the USB transistors to shutoff and for VBATREG to take over the load as seen in the currents.  There is a dip in the voltage to the output as VBATREG comes up and shuts off Q201 and allows the turn on of Q203 and Q204.  Then I turn VUSB off and there is no change on the load nor the output voltage.  I turn VUSB back on and there are no changes.  I turn off VBATREG and VUSB again takes over with a small dip.

The dip could be enough to cause the micro to reset.  I added a 22uF capacitor to the output of this circuit and it did cut the voltage dip down significantly, but it will be load dependent as it has to hold up the load while the switch over happens.  I don't think this is a big deal as the standard use case will probably not be switching the battery connection on and off while connected to USB, I imagine USB will primarily be for development.

I also simulated without Q203 because I'm not sure it's doing anything in this circuit and it didn't seem to affect the results other than a smaller voltage dip on the output during switchover.  There could be some edge case with supplies not at 5V where it affects things I'm missing.

*edit the order is not controllable of those inline pictures and the name isn't visible that I could see.  The sim without Q203 precedes the standard sim

401 words

Attachment SchematicStandard.png
Attachment SimNoQ203.png
Attachment SimStandard.png
In reply to Jesse Robinson

This forum post has been removed

The content of this forum post has been removed and can no longer be accessed.
In reply to Deleted user

This forum post has been removed

The content of this forum post has been removed and can no longer be accessed.
In reply to Jesse Robinson

Re: Why two transistors?

by Mark Hughes -
Well isn't that something. I never would have anticipated that big a dip in VOut. Looks like it drops down to 2.5VDC or so. And that dip decreases with the removal of Q203? So then Q203 must be for some other purpose -- perhaps there are some other factors we need to consider.  Bob did mention that this circuit block is in use in several Microchip evaluation boards.
Here are two off the top of my head.
  • Is this circuit block from something else, and therefor solving a problem we might not have in this circuit.  For example, assume that the 5V LDO is no longer present and this is some off-board connector like the USB connection.  Was the designer trying to avoid some short-circuit / reverse current flow event.
  • Is there some inductance introduced during layout that we have not accounted for in this simulation?  Real-world and Ideal sims are two different things, and fast switching times + reactance = problems.
  • Something else entirely.

This is turning out to be very interesting.

Thanks Jesse!

171 words

In reply to Mark Hughes

Re: Why two transistors?

by Jesse Robinson -
I imagine it has to do with stopping different supplies feeding each other, our case is pretty ideal with both supplies both being near 5V so it might not apply, but doesn't hurt to have it. I was just wondering for curiosity, if its taken from a different design which may have had a need for it, that sufficiently answers my question, reusing circuit blocks is very common.

As for the voltage dip, I didn't include the 3.3V LDO and its output capacitance and I used relatively long transition times (1ms) in my simulation from low to high to make it easier to see, it's likely it won't be an issue that affects our design

115 words