Chip/Die identification – Pentawatt/TO-220-5 package

[Jo]

Yes that's true, i did some quick measurements as a sanity check and it turns out to be 56 and 59 µA.

My measurements did puzzle me a bit, but then I thought it may be due to my limited knowledge.

[Jo]

Maybe a look behind the curtain helps, therefore I gave the die another bath.

Before:



After a bath in HCl, you know where I got the idea from. 



I thought that the thin oxide layer with an air gap behind was the reason for the bad picture quality, therefore I gave the die a bath in an ultrasonic cleaner.



And finally I tried putting those pictures together into one.



Hope you enjoy these pictures* and if they show something helpful about the chip that would be even better.


*I know they are not even close to quality of the beautiful pictures Noopy provides us with.

[magic]

I'm starting to doubt that these power transistors are vertical NPNs because in the standard bipolar process there is a diode from the lowest supply rail (ground) to each NPN collector.

Perhaps they are lateral PNP. It's not entirely unheard of, there are LDOs which use a large PNP pass transistor to allow output within millivolts of the input. The same could be the case in a buck converter or a smart switch.

Based on the appearance of the left transistor, I think pin 1 would be the more likely candidate for VCC in such case. Furthermore, there are traces going from pin 1 to several places in the control circuitry (a few looking like PNP emitters) and only one trace going from pin 2 to the control circuitry that I can see.

There is still the mystery of that structure between pin 1 and pin 3, but that could possibly be some sort of clamp to protect from negative input voltage or whatever. Once the datasheet is located, it would be explained.

Maybe it's time to go YOLO?
Your measurements suggest it should be safe to ground pin 3 and apply 5V power to pin 1. Maybe with current limit to a few mA or with 1kΩ resistor.
See what's the output on pin 2. Load pin 2 with 10kΩ to ground and measure again.
Pull pin 5 (input, feedback, enable, ??) with 10kΩ to ground, to 5V and to pin 2 (feedback). See what happens to the output in each case.

[Jo]

Thanks for the suggestions, I guess you are right, there is not much else left to do?

I tried as you suggested using a 1k resistor for current limiting and a 10k as a pull-up/-down.

At first there was nothing happening and pin 2 was low, but when I pulled pin 5 high the IC pulled the voltage on pin 1 down to 2,3V and almost the same voltage was on pin 2. A 10k pull-down resister on pin 2 didn’t change anything.
Then I tried changing the current limiting resistor to 680 Ohm but the voltages kept the same.
Afterwards I tested the breakdown voltages to see if I killed the chip and the only one that changed was “D7” it was now 8,2V but as it wasn’t stable when I first tested it I’m not to worried.

Not really sure what to make of this, I tested a slightly different setup. 5V with 1k current limiting resistor on pin 2, ground still on pin 3. Now there were 4,6V on pin 1, with a 10k pull-down it changed to 4,0V and the supply to 4,6V (that was expected), with a 1k pull-down it changed to 0V and the supply back to 5V.*
I removed the pull-down resistors and put a 10k pull-up on pin 5 and again the supply voltage dropped to 2,3V. When I changed it to feedback from pin 1 to pin 5 it did exactly the same.

I still think 5 is an input, but I’m not sure what to make of the behavior of 1 and 2.

Do you got any ideas for the next step or should i just remove the current limiting resistor and see if the magic smoke escapes?


*typo corrected

[magic]

I forgot to say, if you pull the input up, pull it to the supply pin after the limiting resistor rather than to 5V because the latter may take the input outside absolute maximum ratings and/or trigger some unintended behavior.

It is possible that VCC sags down to 2.3V because the chip needs more supply current than the resistor permits and 2.3V is the point where it begins to shut down and limit its current consumption. In such case it will stay at 2.3V until it is allowed to consume all the current it wants, then VCC will rise to 5V-I·R as expected.

I don't understand the results of the second attempt. I gather you applied power to pin 2, was that 4.6V measured on pin 2 or pin 1? Where was the pull-down applied, which pin was 4.0V and which was 4.6V?

[Jo]

I forgot to say, if you pull the input up, pull it to the supply pin after the limiting resistor rather than to 5V because the latter may take the input outside absolute maximum ratings and/or trigger some unintended behavior.

At first I did it that way, but later I forgot about it... 

It is possible that VCC sags down to 2.3V because the chip needs more supply current than the resistor permits and 2.3V is the point where it begins to shut down and limit its current consumption. In such case it will stay at 2.3V until it is allowed to consume all the current it wants, then VCC will rise to 5V-I·R as expected.

That is quite possible, I guess I didn't go enough YOLO  then. 

I don't understand the results of the second attempt. I gather you applied power to pin 2, was that 4.6V measured on pin 2 or pin 1? Where was the pull-down applied, which pin was 4.0V and which was 4.6V?

I agree it's a bit confusing (and there is a typo) I try it a different way.

2. Setup:
Pin 1: NC
Pin 2: Power limited with 1k resistor
Pin 3: Ground
Pin 5: NC
Measurements:
Pin 1: 4,6V
Pin 2: 5,0V
Pin 5: <0,1V

3. Setup:
Pin 1: 10k pull-down resistor
Pin 2: Power limited with 1k resistor
Pin 3: Ground
Pin 5: NC
Measurements:
Pin 1: 4,0V
Pin 2: 4,6V (that was expected)
Pin 5: <0,1V

4. Setup:
Pin 1: 1k pull-down resistor
Pin 2: Power limited with 1k resistor
Pin 3: Ground
Pin 5: NC
Measurements:
Pin 1: 0,0V
Pin 2: 5,0V
Pin 5: <0,1V

When the pull-down is removed it takes a few seconds until the voltage goes back to the values in Setup 2. I think this behavior is interesting but not quite sure what it means for the pin out of the IC.


5. Setup:
Pin 1: NC
Pin 2: Power limited with 1k resistor
Pin 3: Ground
Pin 5: 10k pull-up resistor (after 1k resistor)
Measurements:
Pin 1: 2,3V
Pin 2: 2,3V
Pin 5: 2,2V

6. Setup:
Pin 1: Feedback to pin 5 with a 10k resistor
Pin 2: Power limited with 1k resistor
Pin 3: Ground
Pin 5: see pin 1
Measurements:
Pin 1: 2,3V
Pin 2: 2,3V
Pin 5: 2,2V

[magic]

2. Setup:
Pin 1: NC
Pin 2: Power limited with 1k resistor
Pin 3: Ground
Pin 5: NC
Measurements:
Pin 1: 4,6V
Pin 2: 5,0V
Pin 5: <0,1V

Are you sure it's not 5V-50mV on pin 2?

I don't think we can rule out other possibilities with certainty, but...

The results so far appear consistent with a PNP transistor being between pins 1,2. It seems that the chip tries to sink significant base current (a few mA?) from that PNP when pin 5 is raised. A PNP is quite symmetric, so it can conduct in both directions and there is no difference if power is applied to pin 1 or pin 2 as long as (by any random chance, leakage, whatever) the chip enters a state where base current is drawn form the PNP for a moment and both pins get equalized.

I think pin 1 is the proper supply so the chip works more or less reasonably in this configuration and behaves like a high side switch or something of that sort.

When you apply power to pin 2, current flows from the collector to the base and apparently from there it somehow finds a way to pin 1 (with 6.7~8.2V breakdown voltage) or, if pin 3 is grounded, to other circuitry and from there also to pin 1, enabling the chip to be turned on by pin 5. If pin 1 is loaded with a resistor, there is one diode drop lost along the path from pin 2. And I suspect that once pin 1 is pulled below the magic threshold of 2.3V, something inside turns off and that's why pin 2 goes open circuit again.

I still don't know what it is. There are a few contact pads for factory testing, it's not clear if a dumb switch would need them by they are common on voltage regulators. There seems to be a small capacitor (somewhat irregular, flat metal area near the right end of the big transistor/diode/whatever at the bottom).

Can it drive 1kΩ load when it is tuned on?

[Jo]

Are you sure it's not 5V-50mV on pin 2?

Good one
Maybe there is a reason why I used such a low resolution. 

But just for the fun of it I took a measurement of the voltage drop across the resistor: 31,8 mV and the resistor value is 997 so the current should be 31,9 µA .

So it seems like it is not a constant current device.

Joking aside, your explanation makes sense to me and gave me a bit more confidence moving forward.

And I suspect that once pin 1 is pulled below the magic threshold of 2.3V, something inside turns off and that's why pin 2 goes open circuit again.

I thought it might be some kind of protection circuit.

Can it drive 1kΩ load when it is tuned on?

Yes it can:

7. Setup:
Pin 1: Power limited with 1k resistor
Pin 2: 1k pull-down
Pin 3: Ground
Pin 5: 10k pull-up
Measurements:
Pin 1: 2,3V
Pin 2: 2,1V
Pin 5: 2,2V

8. Setup:
Pin 1: Power limited with 1k potentiometer and a resistor (also called an amperemeter ) in series
Pin 2: NC
Pin 3: Ground
Pin 5: 10k pull-up
Measurements:



9. Setup:
Pin 1: Power
Pin 2: 1k pull-down
Pin 3: Ground
Pin 5: 10k pull-up
Measurements:
Pin 1: 5,0V
Pin 2: 5,0V
Pin 5: 4,9V

10. Setup:
Pin 1: Power
Pin 2: 180  pull-down
Pin 3: Ground
Pin 5: 10k pull-up
Measurements:
Pin 1: 5,00V
Pin 2: 4,96V
Pin 5: 4,86V

And finally I checked the breakdown voltages, they haven’t changed, therefore I’m somewhat confident that I haven’t fried the IC.

[magic]

That's good, I guess there is still chance that we are not entirely wrong about everything

How does the output current depend on pin 5 voltage and does it turn on at something like 1.25V or Vout-1.25V? That would suggest an adjustable voltage regulator. If not, I don't know, maybe a fixed regulator with shutdown (at least 5V) or just a stupid high side switch.

The other structure (at the bottom) appears to be nothing more than a clamp catching power supply going 6V below zero and/or above some threshold (apparently higher than 12V that was tested). I don't get it, is it meant for automotive use?

I can't find any reasonable datasheet. Buck regulators like L200 have different pinout, smart switches are all the VN series, I think 100% CMOS (but I could be wrong), linear LDOs have outputs on pin 5.

edit
I forgot to add previously, one reason I think that pin 1 is VCC is because there is 6.7V breakdown going from pin 2 to pin 1 (and even less when ground is connected). So if the chip were to take power on pin 2 and output on pin 1, it would fail with more than a few volts of supply. Erm, screw that, it wouldn't work at all with ground connected because of the leakage to pin 1 capable of pulling 1kΩ to 4V and possibly more.

edit edit
Well, unless the "leakage" is part of intended operation
But the breakdown is very fishy, I wouldn't want an IC working at more than 5V to behave like that when it loses ground connection.

[Jo]

How does the output current depend on pin 5 voltage and does it turn on at something like 1.25V or Vout-1.25V? That would suggest an adjustable voltage regulator. If not, I don't know, maybe a fixed regulator with shutdown (at least 5V) or just a stupid high side switch.

I think for something stupid there is a lot of (to much?) silicon "wasted" for stuff that isn't powetransistor.

Your guess with was very good:

11. Setup:
Pin 1: Power
Pin 2: NC
Pin 3: Ground
Pin 5: 10k pull-up with variable voltage
Measurements:
Pin 1: 4,91V
Pin 2: 0,0V
Pin 5: <= 1,15V
and:
Pin 1: 4,91V
Pin 2: 4,90V
Pin 5: >= 1,25V

12. Setup:
Pin 1: Power
Pin 2: 180 Ohm pull-down
Pin 3: Ground
Pin 5: 10k pull-up with variable voltage
Measurements:
Pin 1: 4,91V
Pin 2: 0,0V
Pin 5: <= 1,15V
and:
Pin 1: 4,91V
Pin 2: 4,87V
Pin 5: >= 1,25V

So the voltage and therefore the current isn’t changing, at least in the range I tested.

Then I looked again at the devices manufactured by STM in a Pentawatt package, these were the ones I found:

TVS-Thyristors  (STIL0*)	I don’t think that’s possible.
Audio Amplifiers (TDA20*)	I don’t think that’s likely considering the test results and the Pinout of the ones I took a look at.
Power OPAMP (L165)	I don’t think that’s likely considering the test results and the Pinout of the ones I took a look at.
SMPS IC (VIPer*)	It’s in the name, it uses “VIPower M0Technology” therefore not possible.
smart power SSR (VN*)	AFAIK they all use “VIPower Technology” therefore not possible.
Ignition coil driver (VB027*)	AFAIK they all use “VIPower Technology” therefore not possible.
adjustable voltage regulator (L200)	Apart from everything else,  just look at the pictures taken by Noopy.
Voltage regulators (L*)	While I think this could be a possibility, I didn’t find any device with a matching pin out.
Alternator voltage regulator (L9918)	Maybe? I think that might be possible, but I have to do a bit more research and right now I’m not sure how to test it.

This is a link to a very short datasheet for the L9918:
https://www.st.com/content/ccc/resource/technical/document/data_brief/group2/c5/c5/11/fe/4a/4b/4e/88/DM00626620/files/DM00626620.pdf/jcr:content/translations/en.DM00626620.pdf

[magic]

That doesn't look like an adjustable regulator. Such regulators generally turn on when feedback voltage is less than 1.25V above ground or more than 1.25V below the output. This chip turns on at more than 1.2V above ground, so I'm not sure how you would wire a feedback network to make it work as a regulator. I think pin 5 is simply a turn-on switch, perhaps reacting to two diode drops above ground or something like that.

It looks quite like those LF series regulators but pins 2 and 5 are swapped and polarity of "inhibit" signal is reversed
https://media.digikey.com/pdf/Data%20Sheets/ST%20Microelectronics%20PDFS/LF00_Rev_17.pdf

Maybe it will output constant voltage when you load it with some 1kΩ and 10µF to ground and increase supply voltage...

Quite surely it's not L9918 because pin 4 is shorted to pin 3.

[Jo]

It looks quite like those LF series regulators but pins 2 and 5 are swapped and polarity of "inhibit" signal is reversed
https://media.digikey.com/pdf/Data%20Sheets/ST%20Microelectronics%20PDFS/LF00_Rev_17.pdf

The similarities are convincing but, even though there is the possibility that these parts are from a rejected batch, I don’t think that these are likely faults.?

Maybe it will output constant voltage when you load it with some 1kΩ and 10µF to ground and increase supply voltage...

I just tested it. I guess I embraced a bit of YOLO attitude now. 

11. Setup:
Pin 1: Power limited with 180 Ohm resistor (voltage variable)
Pin 2: 1k pull-down with 10µF
Pin 3: Ground
Pin 5: 10k pull-up
Measurements:
Pin 1: <2,3V
Pin 2: 0,0V
Pin 5:
and:
Pin 1: 2,3V < U < 13V
Pin 2: ~U
Pin 5(U =13V): 12,8V

Yes, I chickened out at 13V but I think it's unlikely that the voltage is higher than that?
Another thing that doesn't fit with the LF00 is the "Quiescent current" without load.

Quite surely it's not L9918 because pin 4 is shorted to pin 3.

I have to agree with this, but what is your opinion on the basic functionality?
Even though I haven't found one in a pentawatt package, I wouldn't rule out that there was a version without the LIN-communication.

Other wise we would have ruled out all the parts on my list, but there my be more?

Any suggestions where I can find a better list of obsolete parts from STM using the pentawatt package?

[magic]

No way that a faulty part would have a bunch of pins swapped around or reversed logic input action
I thought it could be something similar.

But hey, you ordered a high side switch, you know how to use it as a high side switch now, so what's the problem?
The Chinese guy delivered.

I really don't know what it is. These guys are some fanatics who collect all sorts of documentation about old computers and electronic components, but all they have on STM and SGS (the S in STM) is computer chips and really old stuff. I looked at the SGS "computer and industrial ICs" databook from 1988 for inspiration but found none. Not entirely surprising, given the 1992 date on the die.

I think at least my guess about PNP between pins 1 and 2 is correct because only a saturated PNP or P-MOSFET could keep the pins isolated when off and equalize them perfectly when on, regardless of which pin is used for supply and which for load. And a PNP better matches the horrendous current consumption. And the lack of diodes, though MOSFETs are four terminal devices and maybe one could be made without a body-source short on an IC; off the top of my head, I'm not sure if it would be doable and work well.

[Jo]

But hey, you ordered a high side switch, you know how to use it as a high side switch now, so what's the problem?
The Chinese guy delivered.

It's just that the guy who did the PCB layout didn't prepare for that, how shortsighted of him. 

I really don't know what it is. These guys are some fanatics who collect all sorts of documentation about old computers and electronic components, but all they have on STM and SGS (the S in STM) is computer chips and really old stuff. I looked at the SGS "computer and industrial ICs" databook from 1988 for inspiration but found none. Not entirely surprising, given the 1992 date on the die.

Thanks for the link, what a wonderful website, the wealth of information there is really awesome.
Using their SGS Documents I found a few more chips that were available in a Pentawatt-package. After a bit more digging I think the L9351 and as a less likely alternative the L9350 is not just a possible, but a likely solution to this search of the Chinese mystery chip.
To me the L9351 and L9350 look like the predecessors to the VN-series of high-side smart SSRs.

[magic]

Yep, automotive PNP smart switches. I don't know where you found them but they are perfect candidates

That's not the worst outcome you could get playing Chinese roulette; another guy received open collector comparators relabeled as audio opamps

[Jo]

 Big thanks for all your help during this project and also thanks to everyone else who contributed to this Topic, I learned a lot. Isn’t it amazing how much entertainment you can get from being betrayed by a ebay-seller.

That's not the worst outcome you could get playing Chinese roulette; another guy received open collector comparators relabeled as audio opamps

I think if there were resistors available in a pentaawatt/TO220-5 package they wouldn’t resist sending me those labeled as whatever I need...

If you wonder why, back in February, I couldn’t get the chip to do anything, I think that that chip was already damaged in previous testing.

Yep, automotive PNP smart switches. I don't know where you found them but they are perfect candidates

I found them in the ‘87 shortform catalog looking for devices using the pentawatt package and narrowed it down using their datasheets:

L9350:
https://www.datasheetarchive.com/pdf/download.php?id=5f461c8cdfa9b4f8ff702198357cfbc16cedb2&type=M&term=L9350

L9531:
https://www.datasheetarchive.com/pdf/download.php?id=bc8d4fdcb7e45a227605df6cb46a9e03d20f71&type=M&term=L9351

Here’s a quote from the L9351 datasheet “Very low standby current (30mA max.) “ I think some things have changed since then.

I tested the over current and over temp protection on my chip and indeed it’s working. The current limit is 1,2A at 12V and when the over temp is reached it switches off, as soon as it is cooled down a bit it switches back on.
Is there anything else I should test?