BC857 Destroyed

[James Arthur]

Yes.

None of the relays have the classic inverse parallel diode at all.

Whoever drew that schematic ought to have been shot too. I'm hard pressed to
tell what are connections and what aren't.

Graham

+10v +10v
-+- -+-
| | TR8
|1 '>| BC556
D1 V |-------
o--- -+- C10 /|
o--^ 3| 470uF |
.----o----||----o
||( 3| + |
||( -+- .-. R3
||( ^ D1/2 | | 4.7k
'----o | |
RL1-3 | '-'
| |
|/ TR2, |
---| TR3, ===
|>. TR4
|
===

HTH,
James

 

[Eeyore]

AFAICT the BC556 shouldn't break,

AFAICS they're not using a BC556 but the SMT equivalent with a lower Pd.

Graham

 

[James Arthur]

AFAICS they're not using a BC556 but the SMT equivalent with a lower Pd.

Graham

The confusion notwithstanding, TR8 is switching, just 20mA, briefly,
and only 10v. Dissipation should be minimal. Almost any transistor
should handle that. Even if TR8's left on, C10 time-limits any
pulse's duration.

Worst case I can figure is if C10 is charged and relays RL1-3 are
activated simultaneously. Using Kasper's lowest figure of 1276 ohms
per coil and 20v applied, that's only about 50mA max., worst-case.
Even TR8's skimpy 2mA of base drive should saturate it handily.

So, something's not as advertised. For example, if the relay coils
were lower than expected resistance and TR8 failed to saturate, TR8
could fail from excess dissipation. If R5 were an incorrect too-high
value, same thing. If TR8 had low beta, same thing.

Still, the failure mode sounds like punch-through--missing or mis-
wired diodes maybe?

Regards,
James Arthur

 

[Jim Thompson]

The confusion notwithstanding, TR8 is switching, just 20mA, briefly,
and only 10v. Dissipation should be minimal. Almost any transistor
should handle that. Even if TR8's left on, C10 time-limits any
pulse's duration.

Worst case I can figure is if C10 is charged and relays RL1-3 are
activated simultaneously. Using Kasper's lowest figure of 1276 ohms
per coil and 20v applied, that's only about 50mA max., worst-case.
Even TR8's skimpy 2mA of base drive should saturate it handily.

So, something's not as advertised. For example, if the relay coils
were lower than expected resistance and TR8 failed to saturate, TR8
could fail from excess dissipation. If R5 were an incorrect too-high
value, same thing. If TR8 had low beta, same thing.

Still, the failure mode sounds like punch-through--missing or mis-
wired diodes maybe?

Regards,
James Arthur

Pardon me if I return to fundamentals...

Get a current probe and plot a V-I load line on your scope.

You'll probably find a safe-operating-area (SOA) issue.

...Jim Thompson

 

[Tam/WB2TT]

D1 prevents it.

Cheers,
James Arthur

I printed out the schematic, but still have trouble telling what are
connections and what are crossover. Anyhow, I am talking about what happens
when TR8 and RL2 turn off. The problem would occur if TR8 turns off before
C10 has discharged. Speaking of C10, not sure if the polarity is correct. It
would be nice to know how this is supposed to work.

Tam

 

[Winfield Hill]

+10v +10v
-+- -+-
| | TR8
|1 '>| BC556
D1 V |-------
o--- -+- C10 /|
o--^ 3| 470uF |
.----o----||----o
||( 3| + |
||( -+- .-. R3
||( ^ D1/2 | | 4.7k
'----o | |
RL1-3 | '-'
| |
|/ TR2, |
---| TR3, ===
|>. TR4
|
===

It was hard to read. If your drawing above is correct, I see
one thing that worries me. What's to keep the collector voltage
of TR8 going positive at the coil flyback? Imagine some current
- it'll feel free to go up until the collector-base diode pulls
up the base, and then the 4.7k it's connected to, and finally
stopping only when TR8's Vbe breaks down in the reverse direction.
Some of this treatment and beta drops. After beta drops enough
the transistor's not too happy with its switching job and dies.

I'd like to see a collector-emitter diode across TR8.

 

[James Arthur]

I printed out the schematic, but still have trouble telling what are
connections and what are crossover. Anyhow, I am talking about what happens
when TR8 and RL2 turn off. The problem would occur if TR8 turns off before
C10 has discharged. Speaking of C10, not sure if the polarity is correct. It
would be nice to know how this is supposed to work.

Tam

It *is* hard to read.

To me, it looks like C10(+) drives a common bus that drives the "high"
side of all the relay coils.

That bus is clamped to +10v by one of the diodes (from pin 1 to 3)
inside dual diode D1. The remaining diodes in dual diodes D1-2 are
connected as inverse clamp diodes across the relay coils.

The negative side of C10 is returned to GND via 4.7k resistor R3.

OPERATION: when the relays are inactive, C10 charges to roughly 10v
via D1(1,3) and R3. To fire a relay, TR8 is turned on, which raises
C10(-) to +10v, and C10(+) to about +20v. That's the relay supply
voltage, which drives the "high" side of all the relays. The "low"
side of whichever relay is being fired is then pulled down by its
associated driver transistor, TR2,3, or 4.

HTH,
James Arthur

 

[James Arthur]

Pardon me if I return to fundamentals...

Get a current probe and plot a V-I load line on your scope.

You'll probably find a safe-operating-area (SOA) issue.

...Jim Thompson

I'll have to get a prototype first--it ain't my question, my thread,
or my circuit!

Grins,
James Arthur

 

[Joerg]

Pardon me if I return to fundamentals...

Get a current probe and plot a V-I load line on your scope.

You'll probably find a safe-operating-area (SOA) issue.

Kasper wrote that the coils of his relays were >1K and there is only one
engaged at a time assuming, ahem, no software bug. So there shouldn't be
more than 20mA. I'd suspect a spike of some sorts. Or they got several
reels of bad apples.

 

[Ben Jackson]

Also the software locks for not activating all 3 releays at once.
Then we could discuss just when the uC starts up, all 3 + boost is
activated... hmm...

You mentioned that you thought they failed during build. Have you
considered what happens when the uC is unprogrammed (outputs probably
floating)? Or *during* programming, in case programming lines are
shared with IO?

 

[James Arthur]

It was hard to read. If your drawing above is correct, I see
one thing that worries me. What's to keep the collector voltage
of TR8 going positive at the coil flyback? Imagine some current
- it'll feel free to go up until the collector-base diode pulls
up the base, and then the 4.7k it's connected to, and finally
stopping only when TR8's Vbe breaks down in the reverse direction.
Some of this treatment and beta drops. After beta drops enough
the transistor's not too happy with its switching job and dies.

I was tempted by that theory too. It's elegant, but the flyback
voltage will be in the opposite direction, and thus safely clamped by
anti-parallel coil clampers (D1 or D2, depending), right?

That is, if the C10-coil node has been driven to +20 volts and the
other coil node is assumed grounded, the C10-coil node would fly
negative.

If the C10-coil node is left at +20v (TR8 'ON'), the remaining TRx-
coil node would fly positive when released.

In either case, the TRx-coil node becomes positive with respect to
the C10-coil node, and the appropriate clamp-diode clamps the
excursion.

Best,
James Arthur

 

[James Arthur]

Did they migrate to a RoHS process with higher reflow temperatures?

That's not a bad guess. An open solder joint on D1 pin 1 or 3 would
remove any control on the relays' flyback. The resulting negative
transient could easily zap TR8...

--James Arthur

 

[Joerg]

That's not a bad guess. An open solder joint on D1 pin 1 or 3 would
remove any control on the relays' flyback. The resulting negative
transient could easily zap TR8...

I was more thinking about trauma to TR8 itself.

But I guess it's time to crack out the DSO and do some careful
measurements. Mostly this stuff happens because some spikes are pushing
a semiconductor too close to the cliff.

 

[Chris Jones]

No 18V 1276 Ohm for 1 and the 2 other is 1620 Ohm.
Also the software locks for not activating all 3 releays at once.
Then we could discuss just when the uC starts up, all 3 + boost is
activated... hmm...



The allowed peak current is 200mA.

What is the current gain (beta) of that transistor ** AT 200mA **?

http://www.nxp.com/acrobat/datasheets/BC856_BC857_BC858_6.pdf

From the NXP datasheet, it looks like the beta is about 40 at 200mA. Your
4k7 resistor cannot supply 200mA/40 = 5mA, because this would require more
than 20Volts across the 4k7 resistor which you do not have.

If the base current is not sufficient, then VCE will be large (perhaps 10
Volts) and so the power dissipation may be several watts.

I suggest you measure VCE during the switching process, to see if the
transistor is coming out of saturation. Unfortunately, this is probably a
once-only phenomenon, and so you may only see the VCE becoming large on
those transistors that are in the process of being destroyed.

If this is the problem, then you may be able to find a Zetex transistor that
is specified to have a high beta AT A HIGH CURRENT.

Chris

 

[Tam/WB2TT]

It *is* hard to read.

To me, it looks like C10(+) drives a common bus that drives the "high"
side of all the relay coils.

That bus is clamped to +10v by one of the diodes (from pin 1 to 3)
inside dual diode D1. The remaining diodes in dual diodes D1-2 are
connected as inverse clamp diodes across the relay coils.

The negative side of C10 is returned to GND via 4.7k resistor R3.

OPERATION: when the relays are inactive, C10 charges to roughly 10v
via D1(1,3) and R3. To fire a relay, TR8 is turned on, which raises
C10(-) to +10v, and C10(+) to about +20v. That's the relay supply
voltage, which drives the "high" side of all the relays. The "low"
side of whichever relay is being fired is then pulled down by its
associated driver transistor, TR2,3, or 4.

HTH,
James Arthur

I think you are right about the connections. Still causes problems for me.
The top ends of all relays are at near +10 volts through D1 (1-2); Now, RL1
will operate whenTR2 is on, RL2 will operate whenTR4 is on, and RL3 will
operate when TR3 is on. What does TR8 do? Or is he claiming the relays will
operate on 20V, but not 10V? I guess that is the big secret.

What he needs to do is to put a storage scope on the collector of whatever
transistor blows, and observe the waveform. In fact, I would do it with a
blown board, and simulate the operation of the transistor with a clip lead
between collector and base.

Tam

 

[James Arthur]

I think you are right about the connections. Still causes problems for me.
The top ends of all relays are at near +10 volts through D1 (1-2); Now, RL1
will operate whenTR2 is on, RL2 will operate whenTR4 is on, and RL3 will
operate when TR3 is on. What does TR8 do?

TR8 yanks C10's cathode up to +10v. Since C10 is already charged to
10v, the anode of C10 is then at +20v. That's bussed to the relays;
that's their supply voltage.

Best,
James Arthur

 

[Jim Backus]

I have a circuit with a BC857B where the transister gets destroyed.

The transistor in first place seems shorted between collector and
emitter, the ressistance is between 5 to 250 Ohm.
The funny part is both "diodes" in the transistor can be measured from
base.

The symptoms sound like secondary breakdown. Very localised
overheating causing collector to emitter short circuit.