I'm refreshing my electronics knowledge a bit (after 30+ years), and I'm trying to build an audio amplifier from scratch without integrated circuits (transistors, and mosfets only).

I'm trying to build the input stage, which is a differential amplifier using 2 transistors. I'm currently simulating it with LTSpice, but I seem to amplify only the positive part of my input signal. I guess it's because of the values of the emitter and collector resistors.

I understand the principle of a diff amplifier, but I'm struggling with the practical setup.

Can somebody explain to me (or point me to a site) how to calculate those resistors?

I guess there should be current flowing through R2 while out1 should be at ground level, if no signal is applied?

I included my schematic and the waveforms.

Thanks a lot!

schematic and simulation

  • \$\begingroup\$ There are plenty of books and courses that explain how to do circuit design. An BJT diffpair with no degeneration can only handle an input voltage of around 50 mV so you're over steering this pair by a factor 10. \$\endgroup\$ – Bimpelrekkie Jan 21 at 16:47
  • \$\begingroup\$ The working principle of a "good" diff. amplifier requires a large resistor in the common emitter leg - much larger than the collector resistors. Remember: Very often we treat this resistor - together with the supply voltage - as a current source. \$\endgroup\$ – LvW Jan 21 at 17:19
  • \$\begingroup\$ @Bimpelrekkie I was expecting oversteering, and I hoped to see a square wave between +20 and -20V. Now I only see the upper half and a almost nothing in the neg part. Next would be adding feedback. \$\endgroup\$ – WimDH Jan 21 at 19:14
  • \$\begingroup\$ When do you expect the voltage to be -20 V? The lowest value at out1 occurs when in1 is at a the top of the sinewave (+0.5 V). Now investigate that! Copy this schematic to a new sheet, then remove V3 and replace it with a DC voltage source of 0.5 V. Then do a DC operating point simulation. With the result you can observe the state of the circuit as if it is "frozen in time" at the top of the sinewave. ... \$\endgroup\$ – Bimpelrekkie Jan 21 at 19:41
  • \$\begingroup\$ ...What are the values of the voltages of Q1 like \$V_{BE}\$ and \$V_{CE}\$. What does that mean for the operation of the transistor? Suppose there was -20 V at out1, would that be OK for the transistor or would something blow up? Think of the diode model of the bipolar transistor. \$\endgroup\$ – Bimpelrekkie Jan 21 at 19:41

Firstly, you probably need to set V4 to be 0 volts (midway between the power rails) if you are testing the front end differential amplifier. Currently it appears to be set to a value called "V" and I can't see what this is anywhere.

But, your circuit has a massive voltage gain so reduce the input voltage AC amplitude to produce a decent sinewave on the output. Remember that when you "close the loop" the circuit gain will become much lower due to negative feedback and the actual levels at the bases of the transistors will be around a milli volt or so. That is about the right sort of level you'll need to use when testing your differential front-end.

  • \$\begingroup\$ Tnx Andy. I was expecting to see a square wave between +20 and -20V. V4 is set to 0V, I even tried to ground it. \$\endgroup\$ – WimDH Jan 21 at 19:16
  • \$\begingroup\$ @WimDH it’s not a rail to rail circuit. The 200 ohm emitter resistor is going to eat up 25% of the available supply range. Did you try it with a lower voltage input? \$\endgroup\$ – Andy aka Jan 21 at 19:38
  • \$\begingroup\$ Indeed Andy, I replaced the resistor by a current source and changed the collector resistors. I added a feedback loop to limit the gain. Now I got it solved. Thenks for the hints ;-) \$\endgroup\$ – WimDH Jan 21 at 20:38
  • \$\begingroup\$ @WimDH Do you know what to do next? \$\endgroup\$ – Andy aka Jan 21 at 21:22
  • \$\begingroup\$ I think I do... I'll have a look at the driver stage now and select the mosfets. I'll try to read some stuff about it and if I'm stuck, I know where I can find help :-) \$\endgroup\$ – WimDH Jan 22 at 19:04

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.