Most probably I'll go with NE5532 for dual and MC33079D for quad opamp. TL07x are there on PCB because those are available in default Kicad library and are standard SO8/SO14 devices 🙂
If you are going to use bipolar opamps such as the NE5532 (which is an excellent opamp for the money) you will need to add input offset current balancing. This basically means the positive and negative input of the opamp need to see about the same resistance to ground. In this situation you can consider any driven point as being ground as well as the ground rail itself.
So if for your first stage you have 10K in parallel with 22K on the input and due to the feedback resistor you have another 20K in parallel with this. This work out to 5.1K so you either use a 5.1K or a 4.7K if you can't get a 5.1K.
This will balance the voltage on the positive input and the negative input due to the input bias current. The currents are about the same so if the pass through the same resistance you get about the same voltage, as one is positive and one is negative they cancel at the output. There will still be a small offset voltage as the bias currents don't match exactly.
If you don't do this and you don't have a coupling capacitor before any volume control the DC passing through the pot will make it sound scratchy when you adjust it. And if you DC couple to the amplifier you could end up with DC on the speaker.
Regards,
Andrew
So if for your first stage you have 10K in parallel with 22K on the input and due to the feedback resistor you have another 20K in parallel with this. This work out to 5.1K so you either use a 5.1K or a 4.7K if you can't get a 5.1K.
This will balance the voltage on the positive input and the negative input due to the input bias current. The currents are about the same so if the pass through the same resistance you get about the same voltage, as one is positive and one is negative they cancel at the output. There will still be a small offset voltage as the bias currents don't match exactly.
If you don't do this and you don't have a coupling capacitor before any volume control the DC passing through the pot will make it sound scratchy when you adjust it. And if you DC couple to the amplifier you could end up with DC on the speaker.
Regards,
Andrew
Thanks!
I'm sure I read about input offset current balancing some time ago (was it Small Signal Audio Design?) and I thought this is not critical. In my case before the crossover is a "regular" (not DIY) preamp (UHER UPA-400) and after the crossover are power opamps (LM3876) that are AC coupled.
The AC coupling saved my speakers at least once when input opamp (U1) lost its In+ connection to GNDA and locked to 15V DC on output 🙂 so I don't think I'll remove AC coupling at all.
I have DC protection for speakers lying around (with UPC1237) but it's not yet soldered and most probably I'll save it for my next project.
I think I'll use this input balancing nonetheless. I also have some LME49860 "in stock".
I'm sure I read about input offset current balancing some time ago (was it Small Signal Audio Design?) and I thought this is not critical. In my case before the crossover is a "regular" (not DIY) preamp (UHER UPA-400) and after the crossover are power opamps (LM3876) that are AC coupled.
The AC coupling saved my speakers at least once when input opamp (U1) lost its In+ connection to GNDA and locked to 15V DC on output 🙂 so I don't think I'll remove AC coupling at all.
I have DC protection for speakers lying around (with UPC1237) but it's not yet soldered and most probably I'll save it for my next project.
I think I'll use this input balancing nonetheless. I also have some LME49860 "in stock".
Last edited:
Yeah, bias current compensation is described in Small Signal Audio Design page 99. But D. Self says that Rcompen should have equal value to Rf (feedback resistor) and you gfiandy and other sources on Internet make Rcompen equal in value to a parallel connecion of Rf and Rin.
Is this an obvious error in Self's book?
Also Self advises to shunt Rcompen with capacitor that will lower Johnson noise of Rcompen, what should be the value of that capacitor? 100nF? 100uF?
Is this an obvious error in Self's book?
Also Self advises to shunt Rcompen with capacitor that will lower Johnson noise of Rcompen, what should be the value of that capacitor? 100nF? 100uF?
A question about 5532 and its input bias current.
Considering the worst case of input bias current from the datasheet (800nA) and missing resistor (5.1 k) for offset current balancing the worst case of offset input voltage caused by not using that resistor would be 4 mV on input (not considering input offset voltage from datasheet, max of another 4 mV) amplified to approx 8 mV on U1 output? If yes, then I don't care at all.
Considering the worst case of input bias current from the datasheet (800nA) and missing resistor (5.1 k) for offset current balancing the worst case of offset input voltage caused by not using that resistor would be 4 mV on input (not considering input offset voltage from datasheet, max of another 4 mV) amplified to approx 8 mV on U1 output? If yes, then I don't care at all.
Error is multiplied by gain but it may still be small. The anecdotal evidence is that the opamp will sound better with balanced input currents.
Unless you need to use large value resistors, greater than 47k jhonson thermal noise should not be significant again you need to reconsider in any stage that has high gain where it may become significant. For the value think of the resistor and capacitor forming a first order filter and calculate it to remove noise down to around 500hz or better.
Regards Andy
Unless you need to use large value resistors, greater than 47k jhonson thermal noise should not be significant again you need to reconsider in any stage that has high gain where it may become significant. For the value think of the resistor and capacitor forming a first order filter and calculate it to remove noise down to around 500hz or better.
Regards Andy
I settled on this so called "final" version and sent the order to PCB shop:
There is still much unused space, but board measures 111 x 60 milimeters, so I'll live with that 😉
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
There is still much unused space, but board measures 111 x 60 milimeters, so I'll live with that 😉
I hope to finish soldering during the weekend 🙂
An externally hosted image should be here but it was not working when we last tested it.
Hi all!
I just want to make a small update: my first active loudspeaker project is nearing completion 🙂
One last thing left to do -- tweak the tweeter level a bit higher and I'm satisfied. After correcting ground/power routing inside the cabinet (cables) there is no noise, no hum, no demodulating radio, and GSM susceptibility is on a "normal" level (not advised to leave a cell phone just by the loudspeakers, but >1m distance is enough not to cause audible interference).
My Uher preamp is more noisy that my crossover 🙂
Here are latest measurements, left and right loudspeaker, I didn't bother with room conditioning, hence the somewhat hilly frequency response (I've seen worse loudspeakers to be honest):
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
- Status
- Not open for further replies.