Hi:

I have bread boarded several emitter-followers. I have tried the 2N3904
and the TIP31. The input is a 1 volt 1 Khz rms sine wave. The output is
always riding on a dc level even though I am taking it off of a coupling
capacitor. There are no loose connections or wiring errors.

I have also tried the TIP120 (Darlington) and it doesn't work at all. I even
tried two different devices.

This is really confusing.

David

 

May be leakage thru C2, try .1 uF ceramic or something close to 1 uF film cap just to confirm.

Just out of curiosity what is bias level at emitter ? ~ 6V would be good.

Regards, Dana.

 

what device exactly is used to "take output" from this circuit... ?

 

May be leakage thru C2, try .1 uF ceramic or something close to 1 uF film cap just to confirm.

Just out of curiosity what is bias level at emitter ? ~ 6V would be good.

Regards, Dana.

Hi Dana:

Thanks for the suggestion.

Tried capacitor change with multiple caps. Output waveform is noisy and
gittery and still riding on a dc level.

Bias level at the emitter is 5.98 V.

 

Hi panic_mode .
I am just scoping the output with no load.

 

The ONLY way there can be a long-term DC level at the output is if there is DC leakage current through the capacitor. There is NO OTHER WAY there can be DC at the output (assuming, of course, that the measuring instrument isn't injecting DC). Any DC level at the output, other than due to leakage, means that the capacitor is accumulating charge.

BUT...
If you use a 100 µF capacitor and an oscilloscope with 10 megohm input resistance (at probe tip), the time constant is 1000 seconds. It will take 4000 seconds, over an hour, for the capacitor to reach 98% of charge. Until it reaches 100%, there will be DC appearing on the scope. If a 1:1 probe is used, the time will be reduced by a factor of 10.

Unless they are low-leakage types, about the lowest specified DC leakage current for an aluminum electrolytic capacitor is around 3 µA, though in practice the leakage can be lower if the temperature is moderate and the cap has been subjected to its DC bias for several hours. 1 µA into 10M ohms is 10 volts!

The jitter is probably due to poor regulation by the power supply, resulting in shifting of the DC level. Make an RC filter for the power using one of those 100 µF capacitors (or 1000 µF or more, if you have such on hand) and 50 to 100 ohms. This will of course introduce both some DC and AC error, but it will at least help you determine if the supply is at fault.

 

Add a 10k ohm load at the capacitor output to ground.

 

The ONLY way there can be a long-term DC level at the output is if there is DC leakage current through the capacitor. There is NO OTHER WAY there can be DC at the output (assuming, of course, that the measuring instrument isn't injecting DC). Any DC level at the output, other than due to leakage, means that the capacitor is accumulating charge.

BUT...
If you use a 100 µF capacitor and an oscilloscope with 10 megohm input resistance (at probe tip), the time constant is 1000 seconds. It will take 4000 seconds, over an hour, for the capacitor to reach 98% of charge. Until it reaches 100%, there will be DC appearing on the scope. If a 1:1 probe is used, the time will be reduced by a factor of 10.

Unless they are low-leakage types, about the lowest specified DC leakage current for an aluminum electrolytic capacitor is around 3 µA, though in practice the leakage can be lower if the temperature is moderate and the cap has been subjected to its DC bias for several hours. 1 µA into 10M ohms is 10 volts!

The jitter is probably due to poor regulation by the power supply, resulting in shifting of the DC level. Make an RC filter for the power using one of those 100 µF capacitors (or 1000 µF or more, if you have such on hand) and 50 to 100 ohms. This will of course introduce both some DC and AC error, but it will at least help you determine if the supply is at fault.

Hi ebp:

Thanks for the reply.

I get the capacitor thing. It makes perfect sense.

I have also tried many different caps on the output with no change. I
guess it isn't a leaky cap then.

My test equipment and power supply are all brand new Techtronix gear. I haven't
had issues when building and testing other circuits so I am not sure what is going on.

I guess I am snookered.

 

Add a 10k ohm load at the capacitor output to ground.

Thanks for the reply crutschow.

I put the 10K load on but still getting the jittery output riding a dc level.
I have changed all components multiple times to no avail.

David

 

Put that 10k load on the output.

Make sure you have the transistors E, B, and C connected to the right points. You should see 12.0 volts on the collector, 5.92 volts on the base, and 5.22 volts on the emitter.

 

Rather than using a capacitor to couple to the scope input, try putting your scope probe directly on the emitter and setting the scope for AC coupling.

AC coupling will make the average at the output side of the cap exactly zero, again assuming no leakage, which is a safe assumption with the scope's internal DC blocking cap. If there is any asymmetry in the waveform at the emitter for any reason, there may appear to be a DC offset - the positive and negative peaks will not be exactly equidistant from the baseline (think about what would happen with a rectangular pulse of 10% duty cycle, swinging between -1 V and +1 V).

---
Verify with your scope:
- AC input signal is good with no signs of jitter
- DC voltage at the emitter is stable with the input disconnected

Is there any chance that you might be getting some line-frequency noise coupling into your circuit? What are you using to connect your signal generator to the breadboard (coax, or ??)?

 

Put that 10k load on the output.

Make sure you have the transistors E, B, and C connected to the right points. You should see 12.0 volts on the collector, 5.92 volts on the base, and 5.22 volts on the emitter.

Collector 12.0V
Base 5.8V
Emitter 6.6V

10K load is on the output.

 

Base voltage should be higher (more positive) than emitter by about 0.6 to 0.7 volts!

Pinout for 2N3904 with leads pointing down, flat facing you is EBC, left to right.

 

Base voltage should be higher (more positive) than emitter by about 0.6 to 0.7 volts!

Pinout for 2N3904 with leads pointing down, flat facing you is EBC, left to right.

Checked out okay. Actually the pins are marked which is rare.

 

Checked out okay. Actually the pins are marked which is rare.

I actually have it working now.

The transistor readings are:
Collector 12.0V
Base 5.7V
Emitter 5.0V

I am using an old proto board and some of the connections are not tight.
If I hold some things in place, I get the thing working.

Thanks for all the help and sorry if I frustrated you guys.