Circuit & Component Check

[Dominic-Luc Webb]

It is always a good idea to review the data sheet if you can find one:
http://www.onsemi.com/pub/Collateral/2N3055-D.PDF

I have several. I do not necessarily trust these. Note that some of them
currently available on the Net (I just found one) specified Vce in units
of amperes, for instance. In some cases, the units are mixed up. Thanks
for the link, John. I do not have this particular one, and scanning
through it, I do not see this error. Indeed, there appear to be some
numbers I have not seen. On the other hand, note that it is not clearly
stated that it is even an NPN at all. This appears further down, in the
sheet, but it is almost inadvertant. Further, this spec sheet does not
show any internal archecture and does not seem to state whether or not
this is a single NPN or Darlington config. In the case of the 2N3055,
there are so many newby's, this little bit of extra blurb would be
useful to quite a few customers.

So what do you make of the problem of getting the 2N3055 to shut off?
Could it be damaged?

Dominic

 

[Dominic-Luc Webb]

Study the data sheets and come back with anything you see that doesn't
make sense to you.

One thing that does not make sense is that Danny gets a varying low. I
would think low should be right at ground potential for this circuit.

I think BC108 is not a resistor, but a small signal NPN, about 0.1 Ic
and 300 Mhz with a large Hfe.

Dominic

 

[John Popelish]

I have several. I do not necessarily trust these. Note that some of them
currently available on the Net (I just found one) specified Vce in units
of amperes, for instance. In some cases, the units are mixed up. Thanks
for the link, John. I do not have this particular one, and scanning
through it, I do not see this error. Indeed, there appear to be some
numbers I have not seen. On the other hand, note that it is not clearly
stated that it is even an NPN at all. This appears further down, in the
sheet, but it is almost inadvertant. Further, this spec sheet does not
show any internal archecture and does not seem to state whether or not
this is a single NPN or Darlington config. In the case of the 2N3055,
there are so many newby's, this little bit of extra blurb would be
useful to quite a few customers.

The data sheet says on the first page that these are power
transistors, not power darlingtons. Every figure on page 3 labels the
2N3055 as NPN. Since this is a combined data sheet for a
complementary pair, one of then has to be a PNP and one has to be an
NPN.

Perhaps you like the initial description on this one, better.
http://rocky.digikey.com/WebLib/ST Micro/Web Data/2N3055, MJ2955.pdf

So what do you make of the problem of getting the 2N3055 to shut off?
Could it be damaged?

That is always a possibility ot be checked out. But I don't have
enough details on the application to say. Care to tell me the
details, again, of the circuit that gave you trouble?

 

[Danny T]

One thing that does not make sense is that Danny gets a varying low. I
would think low should be right at ground potential for this circuit.

It was bad code - it was only on part of the time, and my meter was
averaging the high/lows!

 

[Dominic-Luc Webb]

That is always a possibility ot be checked out. But I don't have
enough details on the application to say. Care to tell me the
details, again, of the circuit that gave you trouble?

John (and others),

The circuit is 555 in classic astable mode. In one example circuit:

Timing cap, Ra and Rb give 1 kHz 7 volt square wave with about 65% duty
cycle at pin 3. This is confirmed by oscillscope. Supply is 7 volts
(Vcc1).

Pin 3 goes (series) to 100 Ohm resistor followed by 20 kOhm pot followed
by NPN transistor base. The load is an 8 Ohm 0.25 Watt speaker in
series with a 500 Ohm resistor and receives power from supply that is 20
volts (Vcc2) relative the 555
ground:

Pin3 ----> 100 Ohm ----> 20k pot ----> NPN base

Vcc2 ----> 500 Ohm ----> speaker ----> NPN collector

Emitter goes to ground


BF658 as transistor (small signal type) can give loud sound by tuning the
pot. Oscilloscope shows current going on and off.

2N3055 as transistor shows continuous 16 volt plateau and barely 0.2 volt
superimposed square wave, or greatly increased frequency. No tuning
of the pot gives loud sound from speaker.


BF658 does this:


16.2 V ___ ___ ___
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
Gnd --- ---- ---- ---

Swing is from Gnd to 16.2 V.



2N3055 does this:

16.2 ___ ___ ___
| | | | | |
16V --- ---- ---- ---

Never drops to ground potential. It can in some cases go to roughly
12 times high frequency, as well.


Dominic

 

[John Popelish]

John (and others),

The circuit is 555 in classic astable mode. In one example circuit:

Timing cap, Ra and Rb give 1 kHz 7 volt square wave with about 65% duty
cycle at pin 3. This is confirmed by oscillscope. Supply is 7 volts
(Vcc1).

Pin 3 goes (series) to 100 Ohm resistor followed by 20 kOhm pot followed
by NPN transistor base. The load is an 8 Ohm 0.25 Watt speaker in
series with a 500 Ohm resistor and receives power from supply that is 20
volts (Vcc2) relative the 555
ground:

Pin3 ----> 100 Ohm ----> 20k pot ----> NPN base

Vcc2 ----> 500 Ohm ----> speaker ----> NPN collector

Emitter goes to ground

BF658 as transistor (small signal type) can give loud sound by tuning the
pot. Oscilloscope shows current going on and off.

2N3055 as transistor shows continuous 16 volt plateau and barely 0.2 volt
superimposed square wave, or greatly increased frequency. No tuning
of the pot gives loud sound from speaker.

BF658 does this:

16.2 V ___ ___ ___
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
| | | | | |
Gnd --- ---- ---- ---

Swing is from Gnd to 16.2 V.

2N3055 does this:

16.2 ___ ___ ___
| | | | | |
16V --- ---- ---- ---

Never drops to ground potential. It can in some cases go to roughly
12 times high frequency, as well.

Dominic

I can't think that your 2N3055 is good, unless you have reversed the
base and emitter pins or something similar. It has a lot more charge
storage than the smaller transistor, so it is harder to turn it on and
off quickly, but at this frequency, that should behave very similarly,
except for leakage and rise and fall time. What you describe is not
explained by the difference in the two types.

 

[Dominic-Luc Webb]

I can't think that your 2N3055 is good, unless you have reversed the
base and emitter pins or something similar. It has a lot more charge
storage than the smaller transistor, so it is harder to turn it on and
off quickly, but at this frequency, that should behave very similarly,
except for leakage and rise and fall time. What you describe is not
explained by the difference in the two types.

Thanks John!

I suspect it is bad, and I'll plan to try a new one. I do not think the
spec sheets, combined with the info I got with the transistor when I
bought it are very ambiguous, so wiring is certainly correct. The overall
circuit is correct since it works with all other transistors, taking into
account Ic and Ib, etc for each transistor.

Regarding the spec sheet you mentioned:
http://rocky.digikey.com/WebLib/ST Micro/Web Data/2N3055, MJ2955.pdf

I could not help noticing that here too, DC Current gain (hFE* in
Electrical Characteristics section) is specified as Vce = 4 Amps. Is
it is ever correct to state Vce in amps or typo? This is confusing.

For the next level of complexity, I believe that the low beta of the
2N3055 means it has a narrow dynamic range (varying Ic with fixed base
voltage and resistor) in which it will operate normally. If I understand
these transistor specs and math correctly, I believe this also means that if
I have a varying load, as would be the case with a photomultiplier tube
pulling current from a transformer driven via this transistor, then a much
larger beta could be desirable (i.e., more dynamic range). I think a PMT can
draw anything from 1 nA to 1 mA. The counts per second in pulse counting
mode can go from a few hundred to 10's of millions.

Dominic

 

[John Popelish]

Thanks John!

I suspect it is bad, and I'll plan to try a new one. I do not think the
spec sheets, combined with the info I got with the transistor when I
bought it are very ambiguous, so wiring is certainly correct. The overall
circuit is correct since it works with all other transistors, taking into
account Ic and Ib, etc for each transistor.

Regarding the spec sheet you mentioned:
http://rocky.digikey.com/WebLib/ST Micro/Web Data/2N3055, MJ2955.pdf

I could not help noticing that here too, DC Current gain (hFE* in
Electrical Characteristics section) is specified as Vce = 4 Amps. Is
it is ever correct to state Vce in amps or typo? This is confusing.

Definitely a typo. Since Hfe is not an inherent property, but a
result of a combination of factors, a meaningful Hfe spec always
includes both a collector current (Ic) and a collector to emitter
voltage (Vce).

For the next level of complexity, I believe that the low beta of the
2N3055 means it has a narrow dynamic range (varying Ic with fixed base
voltage and resistor) in which it will operate normally.

By "normally" I assume you mean as a switch.

If I understand
these transistor specs and math correctly, I believe this also means that if
I have a varying load, as would be the case with a photomultiplier tube
pulling current from a transformer driven via this transistor, then a much
larger beta could be desirable (i.e., more dynamic range).

Higher beta implies that you need less drive current into the base to
produce a given collector current. But you can usually provide excess
base current to handle the high current case, without messing the low
current case up, too much. The usual 'cost' is longer turn off time.
Driving any switching transistor with a large voltage and large series
resistor approximates a current source, which aggravates the turn on
and turn off time problems. Slightly more complicated drive schemes
(adding a base to emitter resistor to drain the stored charge out
faster at turn off and paralleling the series resistor with a small
capacitor to drive the transitions harder) can result in significant
efficiency improvements.

I think a PMT can
draw anything from 1 nA to 1 mA. The counts per second in pulse counting
mode can go from a few hundred to 10's of millions.

Sounds about right.

 

[Dominic-Luc Webb]

Higher beta implies that you need less drive current into the base to
produce a given collector current. But you can usually provide excess
base current to handle the high current case, without messing the low
current case up, too much. The usual 'cost' is longer turn off time.
Driving any switching transistor with a large voltage and large series
resistor approximates a current source, which aggravates the turn on
and turn off time problems. Slightly more complicated drive schemes
(adding a base to emitter resistor to drain the stored charge out
faster at turn off and paralleling the series resistor with a small
capacitor to drive the transitions harder) can result in significant
efficiency improvements.

OK. As you can already guess, I have a couple of batteries in series
that yield 20 volts and I now have a voltage divider. My 555 now gets, and
outputs about 5 volts, and the speaker, drawing from the top of the
divider through collector gets consirably more. Even the 5 volt output
exceeds the requirements for any of the transistors being considered here.


Dominic

 

[John Popelish]

OK. As you can already guess, I have a couple of batteries in series
that yield 20 volts and I now have a voltage divider. My 555 now gets, and
outputs about 5 volts, and the speaker, drawing from the top of the
divider through collector gets consirably more. Even the 5 volt output
exceeds the requirements for any of the transistors being considered here.

Dominic

If you run the 555 from a divider, it will have a hard time providing
any significant base drive current without collapsing its supply. I
would replace the divider with a 5 volt regulator.

That said, if you have a 100 ohm resister between a 5 volt powered 555
and the base of a grounded emitter NPN transistor, that leaves you
with 5 volts minus pull up saturation drop of the 555 minus the base
emitter drop of the transistor, across the 100 ohm resistor. Lets say
that the pull up saturation voltage is about (based on the middle left
graph on page 5 of http://cache.national.com/ds/LM/LM555.pdf )
1.5 volts and the base to emitter voltage of about .8 (saturated
switches have more base to emitter drop than ones operating in the
linear mode), so that leaves you with about 5 - 1.5 - 0.8= 2.7 volts
across the 100 ohm base resistor for a base current of about 27
milliamps.

If the transistor has a saturated gain (remember, you get a lot less
gain when the collector voltage gets near or below the base voltage)
of something like 20 to 50 that allows a collector current of
somewhere between .54 and 1.35 amperes. But the 555 supply has to be
able to deliver the 555 consumption and the 27 milliamperes of output
current while holding the supply steady at 5 volts.

Is all this making sense?

 

[Dominic-Luc Webb]

If you run the 555 from a divider, it will have a hard time providing
any significant base drive current without collapsing its supply. I
would replace the divider with a 5 volt regulator.

I actually did have a plan to use an LM317 to regulate the 555 Vcc at 5 V.
Another way I have run this circuit is with separate 9 volt (to 555 Vcc)
and 20 volt or more (to coil) supplies. Point of exercise was to give a
higher voltage to the primary winding of a step up transformer than the
555 is capable of. The idea was to drive a higher voltage from the 555 via an
NPN. In this case I am implementing what you call the "PUMP & DUMP"
method (volts * time product is increased).

At this point, I can get my transformers to work OK in the low kHz (1-2
kHz). I have also been thinking to try driving them via solid state or
induction type relays. For higher frequency transformers (more than 30 kHz),
a transistor would be needed because it would be difficult or impossible to
use a relay. But at 1-2 kHz, most relay types look like another option.

That said, if you have a 100 ohm resister between a 5 volt powered 555
and the base of a grounded emitter NPN transistor, that leaves you
with 5 volts minus pull up saturation drop of the 555 minus the base
emitter drop of the transistor, across the 100 ohm resistor. Lets say
that the pull up saturation voltage is about (based on the middle left
graph on page 5 of http://cache.national.com/ds/LM/LM555.pdf )
1.5 volts and the base to emitter voltage of about .8 (saturated
switches have more base to emitter drop than ones operating in the
linear mode), so that leaves you with about 5 - 1.5 - 0.8= 2.7 volts
across the 100 ohm base resistor for a base current of about 27
milliamps.

If the transistor has a saturated gain (remember, you get a lot less
gain when the collector voltage gets near or below the base voltage)
of something like 20 to 50 that allows a collector current of
somewhere between .54 and 1.35 amperes. But the 555 supply has to be
able to deliver the 555 consumption and the 27 milliamperes of output
current while holding the supply steady at 5 volts.

Is all this making sense?

I need some time on the rest. I think I have not been taking into account
the pull up saturation voltage of the 555. Until now, I have not really
understood this detail and have relied on a potentiometer at the NPN base
and using this to tune the circuit for best output. I did note about 2.7
volts as you mention at the resistor and wondered how it came about. I
need to read up on this.

Dominic

 

[Dominic-Luc Webb]

I have actually seen a couple circuits in which this was done, long ago.
No hint was given about why these components were in this config or how
to estimate values. Any idea where this is documented, and in particular,
with some idea how to calculate the resistance and capacitance needed?

Dominic

 

[John Popelish]

I have actually seen a couple circuits in which this was done, long ago.
No hint was given about why these components were in this config or how
to estimate values. Any idea where this is documented, and in particular,
with some idea how to calculate the resistance and capacitance needed?

I usually divert at least 1/10th of the base drive from the series
resistor to the emitter, assuming about .6 volts Vbe. For example if
you have 2.7 volts across your base resistor for a drive of 27 ma
drive current, I would put a resistor no higher than .6V/.0027A=220
ohms base to emitter. THis will drain charge out of the base, even if
the 555 output does not go all the way to zero volts (and it won't
while current is passing back to it from the base).

The optimum capacitor value depends on several factors, including the
base stored charge (smaller transistors generally have smaller stored
charge, so using the biggest one you can find does not improve
everything), the driving signal swing and rise and fall time, and the
pulse width. You don't want this cap to drive the base so far into
reverse bias that it breaks down the emitter junction, or it will
slowly degrade the current gain if the transistor. If this happens,
you can clamp the reverse voltage with a signal diode base to emitter.

And the series resistor and the paralleled speedup capacitor have to
have a time constant less than the minimum on or off time so things
get back near center before the next transition has to be handled. I
usually look at the base voltage waveform with a scope and select a
capacitor that causes Vbe to go at least to zero on turn off (all the
base charge sucked out before the resistors take over to hold zero
volts), after I have gotten the two resistors selected to provide the
required saturation voltage, Vce, during the largest load current. If
the driver swing is very large, compared to Vbe on, I sometimes put a
low value resistor in series with the speed up cap to limit the peak
driver current to a safe value during the transitions.

The whole idea of this network is a recognition that there are two
current requirements for base drive. One is the DC requirement to
force the transistor to maintain a low Vce during the on time, at the
highest load current and the other current is required only at the
transitions to drive stored charge into and back out of the base
emitter junction to produce fast turn on and turn off for high
efficiency switching.