I've wanted to build a bridged amplifier as a project for a while, so I decided to first simulate a schematic in LTSpice. However, I got really weird results. The first channel looks fine, but on the second, the waveform doesn’t go as low as the first channel (I’m not sure how to explain it better), and the top is heavily distorted. Also, when using the Fourier function, the first channel gets around 0.003148% THD, while the second one is around 0.4–0.5%.
Also in lt spice simulation, i used the universal op amp 5 if that matters.
Here is an image:
The green wave is the 2nd channel and the blue the first.
Heres the schematic:
Also in lt spice simulation, i used the universal op amp 5 if that matters.
Here is an image:
The green wave is the 2nd channel and the blue the first.
Heres the schematic:
Hi Angel23,
You haven't implemented the right side amp in the same way as the left.
I've taken the liberty of modifying your schematic and attaching below. I would have pasted a jpeg, but haven't discovered how. 😕
You haven't implemented the right side amp in the same way as the left.
I've taken the liberty of modifying your schematic and attaching below. I would have pasted a jpeg, but haven't discovered how. 😕
Last edited:
- Change the RIGHT amp to an inverting unity gain amplifier.
- The LEFT amp output then drives the summing junction of the RIGHT amp.
Distortion is much higher in the right side because it doesn't incorporate the output stages inside a feedback loop for distortion suppression. Said differently, all the circuitry between U2 output and op1 operates open loop. Compare with left side which takes feedback from left terminal of the load.
Just use LTspice to
and paste into your post (with the text). Note here I first used irfanview to crop it.
About the amp: The CCS and 1k35 are not a great idea. You are just adding load for the op-amp. You just need a small bias, maybe a couple 200K resistors ~0.1mA and a capacitor to reduce the AC impedance between bases. Then the op-amp is mostly driving the BD139+BD140.
But a 40V op-amp is not so common, and you would actually like more. Perhaps a better choice is CFP output with gain so that the op-amp can run at +/-15V and the OP at say +/-30V. There are other solutions including bootstrapping the op-amp, but they get messy.
Sorry, I'm replying so late, but thank you for pointing out the problem with my schematic. Looking at it now, it was a really dumb mistake, but I probably wouldn't have found it on my own. Thank you very much!
Also, to add images, you can use Ctrl + P or click on the top icon that says "Insert Image."
Hi, SteveU. Looking at it now, maybe that 1k35 resistor might not be suitable, but can't I just add an NPN transistor there so the op-amp only has to drive the NPN transistor? I tried it in LTSpice, and it seems to work fine.
Yes and no. A real op-amp does not work that close to the negative rail. You need some kind of voltage translation to move the Q15 base voltage closer to zero volts where the op-amp output is happy. It could be a resistor divider, or a PNP transistor. But be careful adding too much gain because you get instability. The main problem with using an op-amp front end is that the op-amp is stable with feedback from itself, but not if you add gain. And, of course, the op-amp supply needs to be reduced to the op-amp limits, typically +/-15V.
Regarding the supply voltage, I will lower it, but can't I just use a rail-to-rail op-amp to drive the base of Q15?
I wouldn't recommend using exotic parts to solve a problem that has better solutions. But it's done all the time. You may find that rail-rail op-amps are usually low voltage parts, which is contrary to the needs of a power amp. Using an op-amp for a power-amp input has drawbacks, which is why you won't see it in many professional designs. ~40 years ago, Crown, Phase Linear etc tried it and soon dropped the idea.
You can have a voltage divider to the transistor. With 390 ohms to negativ rail and 2,2k to the op you will have the OP working in class A and still output enough for any VAS transistor. I used NE5534 fed from - rail and +4,7v zener in 1984. 50w with 8ohm and distortion 1kHz 8ppm at full power.
The last broke down in 2023 not bad really.
Nowadays we have the OPA range with rail to rail output. They will work but only to 36v (-4,7) rail. A 2x22v transformer is max.
You can naturally have a zener paralleled with a capacitor in series with the 2,2k and reduce the negativ input voltage for the OP but then we have to check all voltages at start and stop so nothing gets out of limits.
The last broke down in 2023 not bad really.
Nowadays we have the OPA range with rail to rail output. They will work but only to 36v (-4,7) rail. A 2x22v transformer is max.
You can naturally have a zener paralleled with a capacitor in series with the 2,2k and reduce the negativ input voltage for the OP but then we have to check all voltages at start and stop so nothing gets out of limits.
Last edited:
Of course it is bridged / floating load so a 8 ohm load actually draws current of 4 ohms for each amplifier.
So real world you likely want more output devices for each amp.
OPA455 is 150 volt opamp 6,5 MHz 32V/us slew rate. OPA462 is 180 volt opamp
Likely easier to just use 65 to 70 volt rails Dual supply and get same amount if not more power.
Floating load will get more voltage swing but pulls way more current.
No need to add gain, numerous ways to make high current buffers.
Although bridge is possibly for using common single rail switching supply.
And not have to use output capacitors. Trade off for the excessive current they pull and usually dont give you actual 2x the rail swing
So real world you likely want more output devices for each amp.
OPA455 is 150 volt opamp 6,5 MHz 32V/us slew rate. OPA462 is 180 volt opamp
Likely easier to just use 65 to 70 volt rails Dual supply and get same amount if not more power.
Floating load will get more voltage swing but pulls way more current.
No need to add gain, numerous ways to make high current buffers.
Although bridge is possibly for using common single rail switching supply.
And not have to use output capacitors. Trade off for the excessive current they pull and usually dont give you actual 2x the rail swing
