Need help on phase splitter

[guwakzhai]

hi there, im constructing a simple phase splitter using BC550..

as attached is the schematic..
[IMGDEAD]http://img688.imageshack.us/i/schematic.png/[/IMGDEAD][IMGDEAD]http://img688.imageshack.us/img688/617/schematic.png[/IMGDEAD]

problem is when i measure the THD for both the output, the in-phase signal turned out to be OK.. but not the inverted output.. as attached are the ARTA results for in-phase and inverted respectively..
[IMGDEAD]http://img403.imageshack.us/i/phasesplitterinphaseout.png/[/IMGDEAD][IMGDEAD]http://img403.imageshack.us/img403/5309/phasesplitterinphaseout.png[/IMGDEAD]
[IMGDEAD]http://img839.imageshack.us/i/phasesplitterinvertedou.png/[/IMGDEAD]
[IMGDEAD]http://img839.imageshack.us/img839/4305/phasesplitterinvertedou.png[/IMGDEAD]

so i thought it might be the inbalance output during measurement.. i then construct a differential circuit using INA137.. the schematic and measured THD is as attached..
[IMGDEAD]http://img23.imageshack.us/i/ina137.png/[/IMGDEAD][IMGDEAD]http://img23.imageshack.us/img23/2726/ina137.png[/IMGDEAD]
[IMGDEAD]http://img718.imageshack.us/img718/8932/phasesplitterina137diff.png[/IMGDEAD]


the low freq hum is still there.. how to get rid of it? sorry for the noobness of this question.. noob here
[IMGDEAD]http://img718.imageshack.us/i/phasesplitterina137diff.png/[/IMGDEAD]

 

[SY]

There's a big difference in power supply rejection between the phases. You're picking up hum, for which the collector side has lower rejection. If your power supply is clean (check it!), you'll need to play around with the grounding a bit to get the hum out. If you can diagram how you have all grounds arranged, I'm sure that someone will have some suggestions.

 

[guwakzhai]

thx for the fast response...

since this is the testing stage, i've yet to draw the PCB layout.. i was just using breadboard to conduct the test, but i've did make sure the ground is connected at the same point (star grounding)

as for the power supply, i'm quite certain that it is clean.. it's the same power supply i used for my previous power amplifier.. it does hav some hum on lower frequency at the output of the amp (same as in-phase output, hum on 50 Hz) but it's not as serious as this 1..

may i know why the power supply rejection between the phases can be this much difference? what can i do to improve it?

 

[DF96]

The collector has a high impedance, so the potential divider formed by the 1k resistor and the collector impedance is like a volume control near maximum. Any noise on the supply goes straight through to the inverting output.

The emitter has a low impedance, so the opposite effect happens there. Any noise on the ground supply is attenuated (like volume turned down) in the non-inverting output.

The easiest solution is to add more smoothing/decoupling to the supply.

 

[bbent91745]

The two 100K base resistors are equivalent to one 50K to ground (1 part vs 2 = better). The 1K emitter resistor (and collector resistor) are pretty small. I assume the transistor beta is ~100, so looking into the base, it looks like about 100K. Thus, the 50K base resistor forms a divider with the reflected emitter resistor ... base is probably about -8V. The circuit input resistance is about 100K || 50K or about 33K, which is a little low too. I think you could safely increase the emitter/collector resistances a little, depending on the input resistance of the stage you will be driving. BTW, when you hooked your little differential amp to the circuit, if you didn't AC couple it, you also changed the biasing of the circuit.

The other suggestions about supply noise on the outputs are very true. The collector output is 'referenced' to the positive supply voltage, so any noise there comes straight out. If you replace the 2 base resistors with one resistor to ground, the emitter output will be referenced to ground and will have fairly good supply noise rejection.

Have a great holiday season.

Regards,
bbent

 

[DF96]

33K should be OK, unless the prevous stage has high output impedance. Increasing the emitter and collector resistors will reduce the load on the bias chain, which will raise the base voltage, which will reduce the headroom. I would leave it as it is. Alternatively, increase all resistors by the same ratio.

 

[megajocke]

The biasing is off. The transistor sits almost saturated at idle.

You need to move the bias voltage for the base closer to the negative rail.

 

[guwakzhai]

thx guys for the replies! after a few days of troubleshooting, i've find out that the power supply is the main culprit.. i modified my power supply, a normal transformer-rectifier with RC filter to a buffered RC filiter.. the results are much better now..

In-phase output
[IMGHTTPDEAD]http://img841.imageshack.us/img841/5309/phasesplitterinphaseout.png[/IMGHTTPDEAD]

Inverted output
[IMGHTTPDEAD]http://img130.imageshack.us/img130/4305/phasesplitterinvertedou.png[/IMGHTTPDEAD]