Op-amp/transistor/transformer connection problem in inverter circuit

[xeno]

Hello.
I want to build a little inverter circuit and i am a bit clueless about how to combine
the circuit part with the Op-amp/transistors with the actual transformer.
When i connect the emitters to ground( as seems to be common for wiring a transformer
to a class B push pull transistor set-up) then i can't get the signal back to feed the op-amp, because it would be effectively put to ground level.
Could anyone advise me on alternative ways to connect to the transformer in my current set-up?
Thank you.

**broken link removed**

**broken link removed**

 

[MrAl]

Hi,

Your amp is single ended. The transformer shown uses a two output amplifier.
You might build one more amp the same way but with inverted signal and connect that to the other side of the transformer and leave the center open.
Not sure how much power you are looking for here.
If you want to try to do it the same way as the lower drawing you'd have to change your feedback.

 

[xeno]

Thank you MrAl.
I would prefer a solution with one output amplifier, if it is at all possible.
I feed from a XR2206 and the juice coming out of it is already so low )
As i have it now with a cap before the transformer the sine wave is a bit deformed.

EDIT: Here the connection of the primary to the single output amp.

**broken link removed**

**broken link removed**

It does work better with a DC offset correcting cap, but then the sine gets distorted a bit:

**broken link removed**

 

[alec_t]

Try increasing the value of C3.

 

[MrAl]

Hi again,

I would think it would work with a coupling cap like that if you arent too worried about the power level.

Why did you change the feedback from the original drawing? It didnt work like you had it?
Keep in mind that if your amplification factor is big enough you'll saturate the system and see flat tops like that. Lower the gain and see what happens.

 

[xeno]

Thanks Alec_t and MrAl
Using higher values for the cap also does not make it better.
Even at 1:1 voltage gain (by choosing same resistor size for R1 and R2)
the sine looks the same. It might have to do with the charging curve of the cap.

With the actual built circuit, there is a lot of issues with distortions in the waveform
which i can not explain myself at this point, because they are not clippings
but spikes.
I appreciate your patience in this matter and will make some scopeshots

XR2206 Output with no transformer primary coil connected to the transistors:

**broken link removed**





XR2206 Output with transformer primary coil connected to the transistors:

**broken link removed**





Emitter with no coil connected:

**broken link removed**





Emitter with coil connected Before cap:
**broken link removed**



Output Voltage measured after cap across the coil:

**broken link removed**

Far from ideal :/

Looks like maybe the cap forms oscillations here.

My goal was to get a pretty clean sine wave across the coil.

 

[MrAl]

Hi xeno,


In my previous post i mentioned that you changed the feedback. What i meant by that was the connection itself.
Try connecting the right side of R2 to the two emitters of the two transistors.

You can get away with connecting the feedback as you have it if the signal to the transistors is large, but with smaller signals like 1 or 2 volts you'll see a ton of distortion. Connecting the feedback resistor to the emitters instead of the bases allows the op amp to compensate for the voltage drops of the two base emitter diodes, and thus reduce distortion. You still might get some crossover distortion, but much less than what you are seeing now.

Start with a low gain like both resistors equal to 25k and work up from there. Gains too large will also cause clipping depending on the input amplitude.

To model the inductor correctly you should also put a small series resistance in series with it. Say 1 ohm.

 

[ericgibbs]

hi xeno,

Look at this ETO thread , another member was having problems with a XR2206.

https://www.electro-tech-online.com...-unstable-when-using-above-10v-supply.120476/

 

[xeno]

@MrAl. Thank you, i have gone ahead and connected the feedback to the emitter again and have both R1 and R2 30k now. The waveform across the coil looks like this now:

**broken link removed**



@ericgibbs: Thank you for the link, i will read the thread.

EDIT: I have read it, but my problem starts with connecting the coil after the transistors.
The pure XR2206 output waveform looks very satisfying (First image in post #6)
Also after the buffer it still looks nice.
For some reason my coil connection might be less than ideal.

 

[ericgibbs]

hi,
Try a 10R and a 0.047uF in series across the output load.

The output of the push pull stage MUST be coupled to the output load inductor via capacitor, say a 220uF or 470uF for starters

 

[MrAl]

@MrAl. Thank you, i have gone ahead and connected the feedback to the emitter again and have both R1 and R2 30k now. The waveform across the coil looks like this now:

**broken link removed**



@ericgibbs: Thank you for the link, i will read the thread.

EDIT: I have read it, but my problem starts with connecting the coil after the transistors.
The pure XR2206 output waveform looks very satisfying (First image in post #6)
Also after the buffer it still looks nice.
For some reason my coil connection might be less than ideal.

Hi again,

Yes there must be something very basic wrong, like a connection or something. Connecting the feedback properly should have worked when both resistors are 25k or nearly so.
What kind of NPN and PNP transistors are you using? If they are too large there may not be enough drive. Try small ones like 2N4400 and 2N4403. If that doesnt work there's really something very basic wrong.

 

[xeno]

@ericgibbs

Just to make sure i understood you correctly, you mean like this:
**broken link removed**

@MrAl: I am using TIP120 and TIP125. Without connecting the coil, there is a good signal at the emitter, so i don't think it might be due to the transistors alone.

 

[ericgibbs]

hi,
On the other end of the 220uF

Post your asc file.

 

[xeno]

Asc file: **broken link removed**

 

[ericgibbs]

Asc file: **broken link removed**

ReportsL Object not found....

Post it to this Forum using Advanced and Manage Attachments

 

[xeno]

Sorry, okay attached

 

[ericgibbs]

Sorry, okay attached

Got it thanks, the problem is the OPA output drive to out transistors

 

[MrAl]

Hi again,


As i suspected, the transistors look too big to drive directly from the op amp. Try smaller transistors like i mentioned earlier. Once you get that working you can always work your way up to larger ones being driven from the smaller ones.

 

[xeno]

Thank you ericgibbs and MrAl.

I will experiment with smaller transistors.
Your help is much appreciated.

 

[crutschow]

The problem is that the impedance of the 10mH inductor is only 3 ohms at 50Hz, thus the transistor has to drive a lot of current. An actual transformer will need to have a much higher inductance than that to minimize the magnetizing current. You probably need at least a few hundred mH.

Attached is my simulation with a 500mH inductor and a larger output cap (it resonates somewhat with a 220uF cap). That gives a little over 10mA peak inductor (transformer) magnetizing current.

The small glitches are crossover distortion.