I have made a simple, yet, very necessary, PC Speakers' Preamplifier :
PC Speakers Preamplifier - Google Drive
PC Speakers Preamplifier - Google Drive
Simple or not, there is a nasty mistake with the schematics :
The input capacitor, C1, with a value of 1nF must NOT be there and must be removed.
This is because the previous stage, the PC, would have a high value electrolytic capacitor in series to the output in order to remove the DC component. Because the PC analogue audio output is to be used by headphones with a typical impedance of 32 Ohms, this capacitor is extremely large. This capacitor requires a DC path to ground to charge. C1 prevents this and the output capacitor never charges ( or in a very long while ), because C1 is a standard ceramic capacitor which has very low leakage.
Although C1 blocks the DC, the fact the PC output capacitor cannot charge may introduce noise.
Thus, much better is to remove C1 and, then, the big PC output capacitor would charge through R1, R2 and the OA's output and or the schematics.
To rely on the OA's output and the schematics may not be a good idea. Thus, a better approach may be to use a non inverting amplifier with a 10K resistor from input to ground.
The inverting amplifier can still be used after the 10K input resistor to ground, but, R1 must have a large value of 100K and, R2, even much larger.
The input capacitor, C1, with a value of 1nF must NOT be there and must be removed.
This is because the previous stage, the PC, would have a high value electrolytic capacitor in series to the output in order to remove the DC component. Because the PC analogue audio output is to be used by headphones with a typical impedance of 32 Ohms, this capacitor is extremely large. This capacitor requires a DC path to ground to charge. C1 prevents this and the output capacitor never charges ( or in a very long while ), because C1 is a standard ceramic capacitor which has very low leakage.
Although C1 blocks the DC, the fact the PC output capacitor cannot charge may introduce noise.
Thus, much better is to remove C1 and, then, the big PC output capacitor would charge through R1, R2 and the OA's output and or the schematics.
To rely on the OA's output and the schematics may not be a good idea. Thus, a better approach may be to use a non inverting amplifier with a 10K resistor from input to ground.
The inverting amplifier can still be used after the 10K input resistor to ground, but, R1 must have a large value of 100K and, R2, even much larger.
Why is it "very necessary"?
Most PC sound cards are fully able to drive 30 ohm headsets to about 3 volts (p/p) which would easily clip any amplifier designed for Consumer Line Level (CLL) operation.
It seems to me that, while it sounds louder, all you would be doing is increasing the distortion in your PC speakers.
And especially so with a voltage gain of 25.
The circuit works. C1 reads 1 uF (not 1 nF) and is sufficient.
Better keep it there but add a bleeder in front of it (between the inputconnector and R1), say 100k (input imp is 9.1 kohm) to gnd.
If the outputcap is 1 mF, it has to be charged to Vr by the output stage of the opa2134; check current limits! Also, same cap-value to expect on powerrail.
Why do you want to amplify the pc output 25 times? On a 9V rail (with floating ref on 4.3V), inputsignal is ~ (9/3)/25 = 120mVac MAX!
Consider the audiocircuit in pc's are not neccessary high end, so why this 'golden' lunch box???
Yes, I do use lots of decoupling capacitors, large Nichicon, audio, electrolytic in parallel with Tantalum and a lot of Ceramics, X7R, 100V, etcetera.
The goal is to reduce the power supply and wire noise. The power supply is external.
In regards to this project, regardless I have used tons of capacitors, there is still rather strong noise. One of the reason is I use a Tin enclosure and not Aluminum for simple projects.
The most important problem is not only the noise which comes from the PC, but, mainly, the noise which comes from the wireless modem. Although high frequency, this noise is huge.
I, usually, use Ceramics, X7R, 100V or higher : 10uF || 1uF || 0.1uF : 10nF || 1nF || 100pF || 10pF to reduce high and all frequency noise.
I do not remember what I have used for this project, I only remember the noise is huge.
I have not been able to find amplified PC speakers which are loud enough. Most of them are not. The PC, usually, outputs very low signals, mainly, with movies.
Thus, without a preamplifier, the sound is very quiet. The software approach to overcome this is to use VLC with high preamplifier software gain or other preamplifier software, such as Equilizer APO.
The project removes the need of software gain and VLC.
You are right, C1 = 1uF. The capacitor best be removed not to interfere with the output capacitor of the PC audio, which, should be of high value, but, is unknown. Some sources cite 470uF.
The PC does not have a standard of how high the input impedance of the next stage, say, a amplified speakers, should be. Some say 4.7K, others, 10K, some may say 1K, etcetera. Because the PC output capacitor is not known, neither is the input impedance of the amplified speakers, the PC cannot be trusted.
In case the PC is designed to drive 32 Ohm speakers, this does not guarantee high output capacitor. They may cute frequencies, etcetera.
I am not sure whether modern PC's still provide analogue passive headphones output or the output is for amplified speakers only.
The output resistance of the PC is not known either. Passive headphones require low power level, so, there may be a high output protection resistors ( although most amplifier IC's are protected ).
Even, I am unhappy with R1 = 10K. Better is to use non inverting input amplifier to get huge input impedance. In case low input resistance is necessary, a resistor can be put between the input and the ground.
Even better, a fully differential ( instrumentation ) 3 amplifiers with two buffers and another amplifier best be used. One of the inputs is ground. This should reduce the noise.
Anyway, I do get huge noise. I am not sure whether I have resolved this or not. The noise is, probably, caused by the wireless internet. I have noticed this with other projects.
Also, the PC is open, without any lids. Also, the " lunch " enclosure is made on thin Tin and not Aluminium.
A document, called Errata.doc has been published here : Errata - Google Drive
The document correct errors I may have made in various projects I have worked on.
The document is to be updated.
Please, report all errors and problems.
The document correct errors I may have made in various projects I have worked on.
The document is to be updated.
Please, report all errors and problems.
In regards to the noise problems, I have carried out some tests and published a document, called : " Noise Problems with the PC Speakers Preamplifier.doc " in the original folder.
Here is the folder location, again : PC Speakers Preamplifier - Google Drive
Here is the folder location, again : PC Speakers Preamplifier - Google Drive
Yes, in the sense of the tests, C1 is OK to stay.
However, I still cannot find any information on output capacitance of a sound card. Most likely, they do not put anything and rely on low level DC of the output amplifier.
In case they do and in case their cutoff frequency is 20Hz and they do not compensate for multi order filter, also, in case they calculate the minimum headphone impedance as 16 Ohms, then, they must put a capacitor >= 497.4uF. To save money, they would, probably, put 470uF.
Either way, the 1uF input capacitor is OK as you have noted.
Thank you very much for your input.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.