Can SOT23 transistor dissipate 175mW when on minimal size pads on the PCB?

[Flyback]

Hello,
Is the following PNP transistor going to overheat in this circuit (attached)?
The PNP transistor is put there so as to provide a low output on its collector if the load ever goes open.
However, the PNP is just a CPH3105 type, and is in a SOT23 package, and is on a minimal footprint on the PCB, with no cooling copper added to any of its pads.
This PNP is dissipating 175mW (Ieb * Veb)

CPH3105 PNP transistor datasheet:
https://www.onsemi.com/pub_link/Collateral/EN6084-D.PDF

 

[alec_t]

The specs only relate to the setup where the pnp is on a 600 mm^2 substrate, so it's difficult to say.
Can you reduce the value of R2 to, say, 5 Ohm? That still allows switching but drastically reduces the pnp power consumption.

 

[ronv]

Most are rated at 350 mw on FR4. But if you want to be sure you should buy one with a spec for your board.

 

[JimB]

I agree that he datasheet does not give sufficient information the determine if the transistor will become too hot.

However the statement:

This PNP is dissipating 175mW (Ieb * Veb)

does not seem to be quite right.
The base-emitter current multiplied by the base -emitter voltage?
The base-emitter voltage is going to be of the order of 0.7volts, inferring a base emitter current of 250mA.
Sounds a bit on the high side?

On Edit
Looking again at your circuit, if the load is 175mA, do you really want R2 to be 100 Ohm?

JimB

 

[alec_t]

175mW is right. Nearly all the 175mA load goes through the base-emitter junction, which drops nearly 1V (according to LTspice).

 

[Flyback]

Page 2 of the CPH3105 datasheet says that Vbe(sat) is from 0.88V to 1.2V....so I am not sure where you are getting 0.7V from?
Also, the load is drawing 175mA...so I cannot understand where you are getting 250mA from? Have I missed something here?

 

[JimB]

Nearly all the 175mA load goes through the base-emitter junction

Yes, I agree.
But, why not use a lower value resistor?
Does the transistor need to be thrashed so hard?

JimB

 

[Flyback]

Most are rated at 350 mw on FR4

Page 3 of the below says its 336degC/Watt for a SOT23 on a 2 layer board (this board is 2 layer).
So 350mW would give a temperature rise of 0.35 * 336 = 117degC, which is surely way too much temperature rise.

SOT23-3 dissipation (page 3)
https://ww1.microchip.com/downloads/en/AppNotes/00792a.pdf

..also please note, that this SOT23 (CPH3105) is mounted on a minimal footprint, with no extra cooling copper attached to any of its pads.

 

[JimB]

Page 2 of the CPH3105 datasheet says that Vbe(sat) is from 0.88V to 1.2V....so I am not sure where you are getting 0.7V from?

Just a quick assumption from the average silicon transistor.

My point is, why are you thrashing the transistor with such a large base current?

JimB

 

[Flyback]

But, why not use a lower value resistor?
Does the transistor need to be thrashed so hard?

I know what you are saying, but may I ask, "why" do you believe the transistor is being thrashed hard, when its datasheet says its Base current can be up to 600mA?

I am not sure why the designer made it 100R (its an 0603 package)....but this circuit has gone out in small volumes and is now in cars.....we have to decide if the CPH3105 will overheat.

(by the way, the actual load is a LDU0830S350 SMPS LED driver , driving two leds at 350mA....hence its input current from 12Vin, is 175mA, as shown)

...By the way, the full circuit is actually as below, two of the PNP's, one for each load....its right and left fog lights or reverse lights.....if either load goes open, the collector of Q2 goes low , indicating the fault.

 

[JimB]

I know what you are saying, but may I ask, "why" do you believe the transistor is being thrashed hard, when its datasheet says its Base current can be up to 600mA?

So what?
The base current is well with in limits, but the power dissipation is in question and that power dissipation is due to the high base current.

I am not sure why the designer made it 100R

Why dont you ask him?

....but this circuit has gone out in small volumes and is now in cars.....we have to decide if the CPH3105 will overheat.

A bit late in the product development cycle isn't it?

So once again you are using ETO to do your design reviews for you.
Are there any competant electrical\electronic guys where you are?

JimB

 

[Flyback]

Why dont you ask him?

..he doesn't work here any more.

If it were my decision, I would *not* use this circuit, but we are a million miles behind schedule, and the boss is asking us to define exactly what is the risk of this circuit...as a respin of the PCB will delay us and infuriate our customer.
I mean, its one of those "weird" circuits that I would never ever have done....so I am not sure how to exactly quantify the exact risk....does any reader know?
There are two electrical people in the company..... the senior engineer and me...the senior engineer just wants to keep shipping it as it is, since he reckons we'll get away with it, after all, a few prototypes worked fine on the bench, or so he said, I only just joined this company a few days ago.

I actually believe that the operation of this PNP transistor, with Ib a lot greater than Ic, is a weird operation, and probably none of the datasheet parameters apply..do you agree?

I don't think any single engineer can be expected to understand PNP operation when its hooked up in such a weird way.....I didn't think power current was supposed to go through the BE diode?....I thought the BE diode was for small control current only?.....I suspect that the datasheet value of 600mA for the BE current refers only to a short transient current of 600mA? (since the CPH3105 datasheet does say its for 'switching' operation)...maybe they are thinking the CPH3105 would normally be used to turn a mosfet ON , and the 600mA would only flow at turn ON time...for a short interval...who knows?, the datasheet doesn't really say.

I'll own up now and say I have never before seen a BJT being used with the power current going through the BE junction...i thought BE junctions were delicate little blossoms that we should be careful with?

 

[crutschow]

According the this, the free air thermal resistance of a is SOT23-3 is 400°C/W and 263°C/W when mounted with a standard land pattern on a glass epoxy board. This means you could dissipate 175mW up to an ambient temperature of 150°C - (.175 * 400) = 80°C in free air so, even without the added power dissipation from the board mounting, you should be fine.

My interpretation is that the base-emitter junction can carry 600mA continuously since it does not say pulse.

The use of the transistor in this manner is not typical but it makes sense for the purpose of detecting whether the load is drawing current or not. It's a simple and effective circuit.

 

[crutschow]

Deleted

 

[Flyback]

My interpretation is that the base-emitter junction can carry 600mA continuously since it does not say pulse.

yes but its a very brief datasheet.
Surely they don't say "continuously", because surely they simply don't expect anybody to put the power current through the BE junction?

I thought BE junctions were very thin, and thus the resistance of the semiconductor going through the BE junction would be large?

Also, surely you don't say that 150 degC is an acceptable junction temperature for a BJT?....I know its the "maximum allowable", but really , one would want to limit junction temperature to 110degC absolute max?....and preferably lower, to increase lifetime?

 

[crutschow]

yes but its a very brief datasheet.
Surely they don't say "continuously", because surely they simply don't expect anybody to put the power current through the BE junction?

I thought BE junctions were very thin, and thus the resistance of the semiconductor going through the BE junction would be large?

Also, surely you don't say that 150 degC is an acceptable junction temperature for a BJT?....I know its the "maximum allowable", but really , one would want to limit junction temperature to 110degC absolute max?....and preferably lower, to increase lifetime?

What they expect and what they state are two different things. I go by what they state.

This is a small power transistor which can carry 3A of collector current so the base-emitter junction is apparently quite rugged.

If you want to use 100°C for reliability, that's okay. If you crank through the numbers you should still be fine at 175mW. (I'll leave that calculation as an exercise for the reader ).

 

[Flyback]

deleted

 

[Flyback]

If you want to use 100°C for reliability, that's okay. If you crank through the numbers you should still be fine at 175mW

Maximum Internal ambient (inside lamp enclosure) is 70degC.......then 0.175* 263 = 46 degrees......meaning that he junction will be at 116 degrees...this is surely too high?

When I worked at a big SMPS place, the chief engineer used to tell us junction temperature of a transistor should be 109degC absolute maximum.

 

[ronv]

My daddy always taught me to not fight if I knew I was going to loose or it was going to be a draw.
Since it only does work when the system fails, you can be pretty sure it will outlast the rest of the system.
But seriously, excluding early life failures it probably doesn't matter if it fails 1 in a million years or one in 10,000. If you have built a significant number and have any failures you can plot the failure rate using standard methods. When we built 30 million disk drives a year it was still difficult to see a stressed transistor and we did a lot of failure analysis.

 

[JimB]

Since it only does work when the system fails, you can be pretty sure it will outlast the rest of the system.

Ah, but...

If I understand correctly, the load of 175mA is the normal load for this circuit, so the transistor is dissipating 175mW in B-E junction in normal operation.

What we have not been told is the condition which this transistor has to detect.
I assume that it detects no load, ie 0mA.
If that is the case, I think that replacing the 100R resistor with 10R would considerable reduce the transistor dissipation, so ending the debate as to whether it gets too hot.
However, this will increase the dissipation in the resistor. What is the power rating of the resistor?
Maybe a compromise is needed between power in the resistor and power in the transistor.

JimB