I used the attached circuit to add a preamp output to my stock Toyota head unit. The results were disappointing. For the 8V supply I used the output of one of the regulators on the head unit PCB. I remember reading a post on here about using COM8V for a clean supply, but I don't think I used that one, so that may be the cause of the noise. I can hear the CD mechanism in the output and it has crackles, pops, and hiss. It sounds like an AM radio. I am also wondering if there is something wrong with my circuit. Please take a look and let me know if there is anything I can improve. Thank you.
Attachments
It is not obvious to me what is coming in (from the cd unit?) and what the output is connected to? Are you perhaps trying to put an MP3 into the CD player to use its audio amplifier. Head unit is car language, I'm sorry I don't speak it. My car is too loud to hear a radio or CD's.
Nevertheless instead of using this IC to make some sort of fake -8v, I think I would use filtered 12vDC from the car feed to make +5 and +10 with two 5 v zeners like 1N5231. Put a suitable resistor from the end of the two point to line zeners as 100 ohm or something (calculate from the wattage of the zener and the voltage drop) and use your existing filter capacitors. Line goes to resistor bottom, resistor top goes to +12. Usually, a curved and straight capacitor symbol is used for electrolytic capacitors. I hope your 0.1 capacitors are actually ceramic disc capacitors with two straight lines. The plus 10 is the op amp +, the ground is the op amp minus, and the +5 is the signal reference point. Since your input is capacitor coupled, you don't have to worry about the signal references being in a different place. Capacitor couple the output also, if the power amp signal is ground referenced. 10 uf should be adequate. Plus towards you +5 referenced op amp output.
On the plus 12 VDC feed from the car, I would put an inductive hash filter, as cars have a lot of spikes and RF noise. You can wind 20 turns around a ferrite toroid core, or salvage one from the input section of a dead PCAT power supply. Another hash filter on the ground side might also help. If you are happy with this circuit otherwise, you could try it with the two hash filters on + and - in before changing to the zener power supply.
I have run ST33078 op amps at +-3.3 VDC with okay results for mag phono input (20 mv) and Cd player input (1.5 VAC). For a transistor radio input, however, that produced less hiss at 4 VAC , I had to boost my op amps to +-7 VDC, off 8.2V point to line zeners.
Don't know why you need both a unity gain buffer and a 10x stage. Depends in what is going in and going out.I would use the two halves of one op amp for the two stereo channels, at the same gain. If you want variable gain you can put a stereo pot in your feedback loop with the other resistor. Like a 4.7k resistor and a 5 k pot instead of the 10k fixed resistor. If you are injecting the output into the CD player someplace, you need a two resistor mixer, one for this op amp out and one for the other device output.
Nevertheless instead of using this IC to make some sort of fake -8v, I think I would use filtered 12vDC from the car feed to make +5 and +10 with two 5 v zeners like 1N5231. Put a suitable resistor from the end of the two point to line zeners as 100 ohm or something (calculate from the wattage of the zener and the voltage drop) and use your existing filter capacitors. Line goes to resistor bottom, resistor top goes to +12. Usually, a curved and straight capacitor symbol is used for electrolytic capacitors. I hope your 0.1 capacitors are actually ceramic disc capacitors with two straight lines. The plus 10 is the op amp +, the ground is the op amp minus, and the +5 is the signal reference point. Since your input is capacitor coupled, you don't have to worry about the signal references being in a different place. Capacitor couple the output also, if the power amp signal is ground referenced. 10 uf should be adequate. Plus towards you +5 referenced op amp output.
On the plus 12 VDC feed from the car, I would put an inductive hash filter, as cars have a lot of spikes and RF noise. You can wind 20 turns around a ferrite toroid core, or salvage one from the input section of a dead PCAT power supply. Another hash filter on the ground side might also help. If you are happy with this circuit otherwise, you could try it with the two hash filters on + and - in before changing to the zener power supply.
I have run ST33078 op amps at +-3.3 VDC with okay results for mag phono input (20 mv) and Cd player input (1.5 VAC). For a transistor radio input, however, that produced less hiss at 4 VAC , I had to boost my op amps to +-7 VDC, off 8.2V point to line zeners.
Don't know why you need both a unity gain buffer and a 10x stage. Depends in what is going in and going out.I would use the two halves of one op amp for the two stereo channels, at the same gain. If you want variable gain you can put a stereo pot in your feedback loop with the other resistor. Like a 4.7k resistor and a 5 k pot instead of the 10k fixed resistor. If you are injecting the output into the CD player someplace, you need a two resistor mixer, one for this op amp out and one for the other device output.
Last edited:
Sorry, I should have included more details.
I am tapping into the signal inside the head unit before it goes into the internal power amp to get the cleanest signal possible. When playing a CD with a 200Hz test tone at Max volume, the signal is ~1.3V peak-to-peak and has a bias of 2.5V.
The reason for the unity gain buffer is to prevent my circuit from loading the internal circuitry of the car stereo. I added the 10x gain stage since I need approximately 13-14V peak-to-peak output so I can turn the gain down on the external car amplifier (supports up to a 6Vrms input and this circuits provides it with a signal of approx. 4Vrms).
The output of my circuit is directly connected to an external amplifier through RCA cables. The car amplifier has an electrolytic coupling cap at the input.
I am tapping into the signal inside the head unit before it goes into the internal power amp to get the cleanest signal possible. When playing a CD with a 200Hz test tone at Max volume, the signal is ~1.3V peak-to-peak and has a bias of 2.5V.
The reason for the unity gain buffer is to prevent my circuit from loading the internal circuitry of the car stereo. I added the 10x gain stage since I need approximately 13-14V peak-to-peak output so I can turn the gain down on the external car amplifier (supports up to a 6Vrms input and this circuits provides it with a signal of approx. 4Vrms).
The output of my circuit is directly connected to an external amplifier through RCA cables. The car amplifier has an electrolytic coupling cap at the input.
Last edited:
Well, the accounting department says drop the unity gain buffer, the 10x op amp won't load anything down very much. Use a 4.7K input resistor and a 47 K feedback resistor,after the .1 input capacitor, that should keep your input current low. If you need 6 V RMS out then you need at least a 10V power supply(+5, +10). I would try the toroid hash filters on + and - in first, to minimize rewiring.
Lots of people use TL071 all the time, but it does have a latch up feature if it gets too close to the power supply rail. I like ST33078 and JRC4560 (now superceded by 4562 which I haven't tried but the numbers look good.) I did a lot of stupid things to some 33078's and none blew up, and a 4560 in my CS800s power amp withstood something that vaporized the 1k 1/8 watt input resistors. Probably a guitar amp output into the power amp input jack by the previous owner's bandmember. 33078 and 4560 are really fast, so make sure your .1 uf P.S. caps are within an inch of the op amp, and put a 20 pf disc cap across the feedback (the 10k->47k) to cut gain at radio frequencies and inhibit oscillations. When I upgraded my disco mixer from 4558 to 33078, the resulting 1 mhz oscillation was inaudible but it sure made the fan on the power amp run really hard. I have a scope, so I could see it.
The zener numbers I quoted before are 1/2 watt, you'd probably be better off with 1W zeners. BZX85C5V1 is a Vishay part number for a 5.1V 1.3W zener.
Have fun & good luck.
Lots of people use TL071 all the time, but it does have a latch up feature if it gets too close to the power supply rail. I like ST33078 and JRC4560 (now superceded by 4562 which I haven't tried but the numbers look good.) I did a lot of stupid things to some 33078's and none blew up, and a 4560 in my CS800s power amp withstood something that vaporized the 1k 1/8 watt input resistors. Probably a guitar amp output into the power amp input jack by the previous owner's bandmember. 33078 and 4560 are really fast, so make sure your .1 uf P.S. caps are within an inch of the op amp, and put a 20 pf disc cap across the feedback (the 10k->47k) to cut gain at radio frequencies and inhibit oscillations. When I upgraded my disco mixer from 4558 to 33078, the resulting 1 mhz oscillation was inaudible but it sure made the fan on the power amp run really hard. I have a scope, so I could see it.
The zener numbers I quoted before are 1/2 watt, you'd probably be better off with 1W zeners. BZX85C5V1 is a Vishay part number for a 5.1V 1.3W zener.
Have fun & good luck.
Last edited:
You are using 100k resistors in your feedback network on IC 1A and you then follow this with a 11x amplifier (IC 1B).
Reduce your feedback resistor values on IC 1A (I don't know what your source impedance is, but assuming its low). You should be at circa 2k each or lower.
Secondly, its not good practice to have a buffer and then follow it with a high gain stage, which is what you have done here. First amplify and then if you have to buffer.
Finally, even your gain stage feedback resistor values are too high and will result in excessive noise. Use 2.2k and 220 Ohms.
Reduce your feedback resistor values on IC 1A (I don't know what your source impedance is, but assuming its low). You should be at circa 2k each or lower.
Secondly, its not good practice to have a buffer and then follow it with a high gain stage, which is what you have done here. First amplify and then if you have to buffer.
Finally, even your gain stage feedback resistor values are too high and will result in excessive noise. Use 2.2k and 220 Ohms.
- Status
- Not open for further replies.