Is PNP Transistor going to overheat in this circuit?

[godfreyl]

There seems to be some confusion here. I'm not sure why you think any of this is weird or bizarre.

> You asked what Vbe would be with 175mA of base current and negligible collector current.
> WimRFP replied that Vbe would probably be less than 1V, and explained his reasoning.

Is there some part of his explanation you don't understand, or is there some part of it you don't agree with?

 

[treez]

I saw the graphs that WimRFP looked at to get those figures.....the graphs just show it for nominal Vbe......as the front page shows, and as for any bjt, the actual vbe spread over tolerance is significantly more than this....so indeed, vbe could be >1v when ic is only a few microamps and ib is 175mA...when the full tolerance range is considered. ...over temperature and batch, age, etc etc.

Any junction voltage in any bjt is very variable with tolerance , as I am sure you know......I mean how many times do people replace a comparator and reference with a bjt.?......very few, since vbe is too variable over temperature and batch etc.

WimRFP certainly put forward that the BE junction is thin, and thus would have large bulk semiconductor resistance.

 

[godfreyl]

I agree about variation between samples etc, so obviously nobody can tell you exactly what the voltage drop will be. Presumably you were just asking for an estimate.

Anyway page 2 says that for Ib=100mA and Ic=2A, Vbe is typically 0.88V but could be up to 1.2V maximum (at 25 degrees C).

 

[FvM]

Strictly speaking, there's no Vbe(sat) max specification for Ib=175 mA. From the 1.2V@100 mA we could guess numbers around 1.5V. Technically, the base resistance of typically 1.4 ohms (from CPH3105 Spice model) and it's process variations are responsible for Vbe(sat) range. In so far, the circuits suggested in post #3 and post #5 can guarantee a lower voltage drop bound.

I assume that the allowable voltage drop for the monitor circuit would be rated over the full battery voltage range and maximum variation of load voltage. I just guess that values around 1 V may be unwanted. Power dissipation is probably the lesser problem.

I also presume that you don't want the monitor circuit signaling "good" state with a few mA load current as the original circuit does.

 

[treez]

FvM
Power dissipation is probably the lesser problem.

....though you say Vbe can be 1.5v, and from that I calculate power dissipation of 262mW (since ib = 0.175A) .......for a SOT23 on minimal pads.......isn't that too much dissipation?
They say 263 degC/Watt for a sot23 on minimal pads.......so the 262mW gives a temperature rise of 68 degrees, and at the maxmum internal ambient of 60 degrees, that gives a junction temperature of 126 degC............surely that is too much?
I worked in an smps company once, and the chief engineer always told us "no more than 109 degC" for a transistor junction.

If you are saying that 262mW is not too much dissipation for a sot23 on minimal pads, then where do you draw the line for a sot23 on minimal pads.?

I also think WimRFP's comments about inrush into the 22uF cap could blow up the BE junction of the CPH3105 PNP transistor...do you agree?
The capacitor is as follows...
**broken link removed**
the 22uF cap is EEHZC1V220R, as above datasheet.

......inrush is a "step" input of current, and steps have steep sides, which means high frequencies..which means the 22uF electrolytic capacitor will present a low impedance to this inrush, therefore this inrush will be very high current and will blow up the PNP transistor, do you agree?

 

[WimRFP]

@treez: The way you are acting now is not constructive anymore.

You asked us to comment on whether or not the transistor will survive. I think given the info you provided, I gave a solid answer: If you don't want to spend time testing and measuring to assess reliability, use another circuit. You may reread BJT theory with regards to the collector and base current in the base region. This will anwer your high resistance issue.

 

[treez]

@treez: The way you are acting now is not constructive anymore.

.....yes , agreed, but this is because the circuit I am working with, all would agree, is not a "constructive" circuit, its a circuit I would never have designed if it were me, but now our boss is saying lets leave it in there because he doesn't want changes. It would be nice to get an absolute "bill of ill health" for this circuit to take back to the boss, but I believe the conclusion here is that it is generally a bad circuit, but might just be a survivor, despite having a higher chance of failure than other , improved methods. Basically, the designer has hopped it and left us holding the baby.
If all would agree that Vbe could possibly be up to 1.5V, (as FvM said), then that would be a pretty good "death knell" for this circuit.
Any further thoughts on the inrush through the BE junction?...as WimRFP originally considered before he was sidetracked by other considerations.
Nobody got maximum power dissipation allowable in a sot23 on minimum footprints?....think my 263degC per watt is realistic?...(from a rather vague application note).
The boss wont want extensive tests...maybe just power up and put a finger on the sot23 case.

 

[WimRFP]

It is very likely not the steady state dissipation (provided everything is well designed from a thermal perspective (automotive temp. range?)), but it are the transients that may occur in real world, or during EMC testing (negative transients included). You know analysis of transients takes time.

I know people (managers) don't like testing/evaluation (as testing equals money). Try to avoid that people "force" you to say it is fine without further analysis. When it goes wrong, they will blame you (likely losing your job and angry customers).

 

[treez]

Thanks,
regarding the allowable dissipation of an sot23 device...
Comparing it with SMD resistors is helpful..
A sot23 (3mmx1.3mm) body size is slightly smaller than a 1206 resistor (3.2mmx1.6mm).
A SOT23's pad contact area is, however, a mere 1.2mmx0.4mm.
A 1206's pad contact area is 1.6mm x1mm....much more than a sot23.

...so we can say that a sot23 wouldn't be expected to have a power rating as much as a 1206 resistor (250mW)

In fact, the pad area for a 0805 resistor (1.25mm x 0.8mm) is more than for a sot23.
An 0805 resistor, does however have a smaller body area(2mm x 1.25mm) than a sot23.

So I would say that a sot23 could dissipate no more than an 0805 resistor, i.e. 125mW.

At my last company, a Consultant Engineer, who runs his own electronics consultancy, has a range of lighting electronics products on sale to the commercial lighting world, and has completed a number of electronics projects for the military, plus used to design the neon lights for picadilly circus, plus used to design audio amplifiers for commercial sale, told me never to dissipate more than 100mW in a SOT23 package. (He was referring to my use of SOT23 BJTs in a battery charge circuit., for one of his sister companies)

Is there concurrence with the 100mW maximum for a sot23?

Is there at least concurrence that a sot 23 is at most, equivalent to a 0805 resistor in terms of dissipation allowance.?
Its worth noting that resistors don't ever come in SOT23 packages....and this suggests that the SOT23 package is not as good at dissipating heat as a SMD chip resistor package.?

 

[WimRFP]

Referring to your post #29, I mentioned in #28: "provided everything is well designed from a thermal perspective".

The device can have 0.9 W dissipation (Tj=150, data from datasheet) when mounted on about a square inch of PCB material. When dissipating 260 mW (0.175A, 1.5V worst case vbe?) PCB temperature can be more, and you can specify lower Tj (as you mentioned in #25)

The actual Tj depends on things you know, but we don't (maximum ambient temperature, temperature rise in housing [due to other heat generation], allowable space for copper [heat sinking], difference between FR4 and ceramic thermal properties, etc). If you don't know, you need to measure/calculate it (your boss don't like..).

Regarding maximum dissipation in SOT23. Your guideline may be true for small signal devices, but development goes on. There are more and more sot-23 (3 pin) devices that can handle more dissipation on moderate size landing patterns. So if your boss shows one example, he may disapprove your calculations.

I'm very sorry I can't provide you a yes or a know, but I don't like the circuit and I wouldn't use it without further analysis. I need to say that I am somewhat biased as I designed several hot swap circuits for heavy capacitive loads where I used degenerated current mirrors (as also mentioned by FvM).

- - - Updated - - -

regarding #25, that is my worry about transient current. A relative high Rbb' (base bulk resistance) in combination with transients in the A range gives large dissipation in the base region.

Also don't forget what happens when you pull-down the input (so a voltage step from for example 12V to 1V). This will fry the transistor as the BE juncion breaks down (around 7..8V). Because of this large voltage drop (compared to the forward voltage drop) the BE dissipation is high.

 

[FvM]

The power is mainly dissipated through the collector pin, so the rating depends on the thermal connection of the respective pad. But either if 1/4 or 1 W maximum rating are achieved in your design, it's no good design praxis to run the transistor at high power rating when a dfferent circuit can go with 10 percent the power dissipation.

 

[godfreyl]

The discussion so far has had a rather narrow focus. It may be useful to step back and look at the bigger picture.

(In case anyone is interested, the actual load is 2 LEDs being driven by a LDU0830S350 module)

The module mentioned is a 350mA constant current LED driver, which raises a couple of interesting points:

a) Do you actually have to worry about current surges through the capacitor caused by sudden changes of input voltage? I would think not, given that the input is a current source, not a voltage source.

b) Since the circuit shown in post 1 claims to draw 175mA, I presume two of those are connected in parallel to the output of one LDU0830S350 module, which delivers 350mA. Presumably (again), the load connected at the output of those two circuits will be strings of LEDs.

Now, have any steps been taken to ensure equal current sharing between those two circuits? Without current sharing resistors somewhere, one circuit may be consuming e.g. 250mA, while the other consumes only 100mA. Furthermore, if one LED string goes open circuit, the full 350mA would flow through the other circuit. OTOH, that's probably the reason for this circuit's existance in the first place - to switch off the LED driver if either LED string fails.

Anyway, it seems to me that everyone's assumption that the input is 12V and the load is 175mA (as shown in post 1) is completely wrong.

 

[treez]

My apologies, i failed to elaborate there......the LDU0830S350 delivers 350mA to 2 LEDs in series....the 2 leds in series is about 6V......so the input to the LDU0830S350 is about 175mA, as the input voltage is about 12V.

OTOH, that's probably the reason for this circuit's existance in the first place - to switch off the LED driver if either LED string fails.

absolutely correct......this circuit is a retrofit for what were incandessant bulbs for the reverse lights...........if one fails, then all must be turned off so that zero current flows to the module, and then the control unit in the car detects zero current , and flags to the driver that the lights are gone out.

Don't ask me why they do it like this......it seems a little silly to put all the lights out if just one fails...but that's how they want it.

...in the rear lights, there are over 50 LEDs in loads of strings of 3...and if one single LED fails.....yep , you guessed it, the whole rear light is put out so that the zero current failure will be triggered.

 

[godfreyl]

OK, so the output of the circuit in post 1 goes to the input of the LDU0830S350?

 

[treez]

yes indeed.

the below is the full circuit....(I put resistors instead of the LDU0830S350's as it takes too long to simulate if I put in a model for the LDU0830S350)

...The ltpsice sim is also here

 

[FvM]

My apologies, i failed to elaborate there......the LDU0830S350 delivers 350mA to 2 LEDs in series....the 2 leds in series is about 6V......so the input to the LDU0830S350 is about 175mA, as the input voltage is about 12V.

In other words, according to buck converter V/I input characteristic, the claimed 175 mA must be read as "< 350 mA" considering possible operation conditions. If converter efficiency and voltage drop of your monitor circuit are taken into account, the input current at 12V supply already calculates as about 200 mA.

 

[WimRFP]

@Godfreyl: You are right regarding the narrow focus.

The original question was a simple one (the capacitor wasn't present in the original posting), and then you get a simple answer. Now we know it will be an automotive application with unknown number of units (likely high volume), and a load description that is other then 175 mA (likely more as mentioned by FvM).

You need to be an interrogator to figure out what a customer wants and/or needs before drawing a single transistor on piece of paper. A small detail that seems unimportant for the customer can have large impact on the actual design. When doing electronics design, large part of the work is spent on transient/ behavior (including "simple" on/off cycling) and foreseeable misuse.

 

[treez]

large part of the work is spent on transient/ behavior (including "simple" on/off cycling) and foreseeable misuse.

..yes, and I took your advice there , because I have posted another thread about serious overvoltages which appear when car lights are tunred on/off due to filter inductor current being broken.

Well, I don't know if it will be high volume, maybe yes, maybe they will use it in other cars too....but at the moment, we are only doing this for one car type......and yes, you would not believe me in a million years if I told you which car this circuitry is now being used in.......only in small volumes at the moment...they hope volumes will be high....but the price tag on this car...put it this way...nobody here will guess just how much.

 

[WimRFP]

We have now more then 35 postings in this thread, and still you don't have a satisfying answer. There are people replying to your posting that have long experience in electronic design and they don't say that your circuit is fine. I hope your employer will understand that carefull evaluation is required. Relying on "we have 10 circuits running for a year now" doesn't mean you can put 100k circuit around the planet (very likely you knew that allready).

Regarding transients, you may do a search on automotive transient load dump. This will give you usefull links regarding transient phenomena you can expect. Some transients may "shock" you as they are nasty.

I forgot to mention there is also a long duration test that simulates depletion of the car battery overnight. That may cause unexpected problems also (think of increased input current for a buck converter).