rspangle@jarthur.Claremont.EDU (Randy Spangler) (04/11/90)
Well, Heath Roberts suggested I have the leads that were going to my stereo go to one side of an audio transformer, and then take the other side and attach that to my stereo. Result: massive improvement in sound quality Problem: I still get a high whine (mostly >2KHz) even when the speaker isn't doing anything. I figured it was probably RF interference from the monitor cable, since the whine changed timbre (sound, but not pitch) when my display changed (when a directory listing popped up or whatever). So I wrapped (separately) my entire circuit and the entire monitor cable in aluminum foil and grounded them to the chassis. But the whine remained. In fact, it remained even when I unplugged the monitor cable from the computer and turned off the monitor! The whine also still changed sound when I updated the display. So I figure it's the card which is putting out nasty RF somewhere inside the computer. How do I fix this? I'm a little nervous about waving aluminum foil near my computer's chips (crack! Oh, good, the noise went away. Oh. Oh. Oh.) Would wrapping the cord around an RF choke help? (BTW, the card is a Paradise Pro 16-bit SVGA) -- -------------------------------------------------------------------------- | Randy Spangler | The less things change, the | | rspangle@jarthur.claremont.edu | more they remain the same | --------------------------------------------------------------------------
gpz@bridge2.ESD.3Com.COM (G. Paul Ziemba) (04/11/90)
rspangle@jarthur.Claremont.EDU (Randy Spangler) writes: [...] >Problem: I still get a high whine (mostly >2KHz) even when the speaker isn't >doing anything. [...] >How do I fix this? I'll bet that the speaker is driven by the output of some TTL chip (maybe via a capacitor to block the DC component). If this is not the case (i.e., there is some special non-digital oscillator making the beep), the following explanation is not correct and I'd be very surprised :-) :-) In the real world, digital logic devices are actually analog circuits that generate variable output voltages even when they are generating a constant "logic level". That is, even if a gate is putting out logic "1", its output voltage may vary + or - depending on variations in the power supply voltage, the voltage at other gate inputs on the same chip, and the load on any of the chip outputs. In your PC, the load current on the power supply is a very complex function with high frequency components. The variations are caused by chip enables/disables, memory accesses, etc. These load changes translate into variations in the power supply voltage at your speaker driver chip (not to mention everywhere else in the computer) and show up as noise on the output. In general, the variations in output voltage are small, so that digital inputs in the computer ignore them. However, your setup involves a linear (we hope) amplifier and your (relatively sensitive) ears, so the noise becomes noticeable. To remove the noise, you need to remove the small amplitude variations but preserve the basic square wave. This process is called "limiting", and is used in FM receivers to remove amplitude variations in radio signals. A limiter is just a very high-gain amplifier that gets driven far into saturation by its input signal. It is usually followed by a filter to remove higher frequencies so that the output is close to a sine wave, but in your case that's probably not needed. You might try: 1. A TTL or CMOS buffer (inverting or non-inverting) with a power supply separate from the computer's. 2. An op-amp operating open-loop as a comparator with a reference voltage between Voh and Vol (again, with a separate power supply). Of course, these should be inserted _before_ any series capacitor at the output of the speaker driver in your PC. Hope this helps, ~!paul -- Paul Ziemba zapi!gpz gpz@3com.com (415)940-7671 Current nemesis: CA "winters", cold enough to drive the ants indoors but not cold enough to make them hibernate.
elliott@optilink.UUCP (Paul Elliott x225) (04/11/90)
In article <6062@jarthur.Claremont.EDU>, rspangle@jarthur.Claremont.EDU (Randy Spangler) writes: > Well, Heath Roberts suggested I have the leads that were going to my > stereo go to one side of an audio transformer, and then take the other > side and attach that to my stereo. > > Result: massive improvement in sound quality > > Problem: I still get a high whine (mostly >2KHz) even when the speaker isn't > doing anything. I figured it was probably RF interference from the monitor The whine is probably due to the switching power-supply in the PC. There is a significant amount of ripple in the +5V supply, and it will show up on the speaker output (translation of 'significant' = 'I don't know how much'). Here are some things to try: * Connect the transformer inputs closer to the timer chip to reduce ground-loop problems (the ripple current causes a voltage difference between different ground points). * Instead of connecting the 'common' input of the transformer to ground, try connecting it to +5V (at the timer chip). This might improve things, depending on the idle output state of the chip. Use a seperate coupling capacitor so you don't connect +5V to the speaker via the transformer winding. This one is a long-shot, but it might help. * If you don't need high-freq response, consider an R-C filter between the speaker connection and the transformer. * It is possible that the transformer itself is picking up the magnetic fields generated by the PC switching power-supply. Re-orienting the transformer might help. Electrostatic coupling is less likely, but grounding the transformer frame might help (where to ground it?). * For $799.95 I can sell you a whiz-bang DSP card that will filter out the noise. Such a deal! ;-) -- Paul M. Elliott Optilink Corporation (707) 795-9444 {uunet, pyramid, pixar, tekbspa}!optilink!elliott "Less than perfect, that's what I've been aiming for all along."