利用PNP实现3.3V转2.5V的问题

tottionly

REALTEK公司推荐的2.5V供电原理图如上图所示，其中VCTRL是由百兆PHY芯片给出的控制信号，得到2.5V电源后又给百兆PHY芯片供电。
请问此电路图的原理是什么？此时PNP管工作在什么状态？VCTRL应该给出什么控制电平？（未接PNP时我测得VCTRL是低电平，接上PNP后变为2.6V左右）

xinchen

VCTRL的电压只要比VCC低一个开启电压管子就处于开启状态。这种接法和让基极和集电极接在一起的工作原理应该基本相同。

guang3000

工作时候VBE正偏，VBC稍微反偏，PNP管工作在有源区。看着PNP管像个LDO的输出管, VCTRL是输出管的控制电压, 否则怎么保证PNP的集电极刚好是2.5V啊

lovexxnu

3.3-0.7=2.6v
这么接应该有一定的保护隔离作用

fhchen2002

二楼 & 三楼的看法正確

This PNP is part of the control loop.
Usually a discrete PNP component has a small-signal current gain about 50-150.
It works in active region.

VCTL provides an analog control voltage around 3.3 V - 0.7 V (One Vbe drop from 3.3-V VCC).
In case the loop finds VCC_2.5V too low, VCTL will be driven lower (i.e., more |Vbe|, more Ib), thereby creates more collector current (IC) to charge up VCC_2.5V node, and vice versa.

lgy747

5楼3楼完全正确，2楼第2句错误

adefan

It is usually a single transistor linear regulator, and should have a feedback signal to the controller.

hbx310

我感觉可以直接降0,7，因为那里有个PN节

tottionly

REALTEK公司推荐的2.5V供电原理图如上图所示，其中VCTRL是由百兆PHY芯片给出的控制信号，得到2.5V电源后又 ...
tottionly 发表于 2011-5-11 13:50

                               
登录/注册后可看大图

今天又做了一个实验，发现如果将集电极开路，则Vc=3.3V 。原理图中的VCC_2.5V给PHY芯片供电，如果将2.5V与PHY芯片断开，另外在2.5V输出上加电阻负载，发现2.5V并不能稳压，变动范围较大。
请问此电路的2.5V稳压性能跟所接负载有关吗？

fhchen2002

回复 9# tottionly

Yes, the regulator's stability is highly dependent on its load.
I cannot imagine what takes place when the collector of the PNP is disconnected from the power supply pin of the PHY.
My speculation is that the PNP would find its collector current load suddenly disappears (from, for instance, several milli-amperes) to zero,   So the collector terminal voltage would jump up to VCC - VCE(saturation) ~= 3.3 V.  VCTL would go up only if the regulator loop thinks VCC_2.5V is too high.
On the other hand, if VCC_2.5V already lost its supply source (which once was from the collector of the PNP), how could it be charged too high?

Regarding the reason why simple resistive load would cause instability:
Question: what is the resistance of the resistor you tried?
1 KOhm?
You may want to shunt an additional capacitor with the resistor and try its stability again.  You may want to try 1 nF to 10 nF.
I anticipate the on-chip decoupling capacitor has the capacitance in the aforementioned range (between VCC_2.5V to ground).