how to lay-out audio circuit ground

Thread Starter

mozikluv

Joined Jan 22, 2004
1,435
:) hi to all,

am starting this particular topic to share with you guys who would love to make their own power amplifier.

for new builders, many would be unsure of how to lay-out the circuit ground of an audio amp. it is good if the circuit he/she is assembling has a foil trace lay-out all prepared but most of the circuits found does not have this.

i do not say that this is already the rule of thumb on ground lay-out, however based on my many experiences like learning from mistakes, these are the following ways;

A. avoid making ground loops (it invites unwanted signal interference) <_<
B. make a single point of entry to the circuit board for your ground connection
C. use the "star topology" when distributing your ground line
D. install your circuit filter caps as near as possible to your point of entry.
E. use seperate ground lines for input, decoupling, output & power line filters all converging to your ground single point of entry. (star topology)
F. item E is suggested for amps using discrete components
G. item E is also suggested for audio power IC, :p HOWEVER NOT ALL AUDIO POWER IC CAN USE THE STAR TOPOLOGY SYSTEM. PLS. TRY TO KNOW IF THE IC USES A DIFFERENT GROUNDING SYSTEM. this kind of power ic uses a single ground line for all connections. :angry: if you employ the star topology technique you will be creating distortions even at low level volume. it can also increase the crosstalk of the ic.

i have seen lot of guys make this kind of mistake and they will probably conclude that the ic they have use is defective after making all checks on connections and voltage measurement. :unsure:

i hope this topic would enlighten any member who will read this. :) :)
 

Dave

Joined Nov 17, 2003
6,969
Not being from a Power engineering background, I have never come across some of the ideas you have mentioned.

Have you made such a power amplifier, mozikluv?
 

Thread Starter

mozikluv

Joined Jan 22, 2004
1,435
:) hi dave

are you referring to the audio power ic that employ a single ground line? yes there are, philips, stmicro, toshiba, national, nec. they are usually old types but there's still lots of them in the market :)

am not a believer on the sound quality of these audio ic. its purely for convenient assembly. sorry to say that, no harm meant. if you really go for quality sound go for discretes or tubes. you are totally limited to what the manufacturer puts in and the parameters you can operate them. :rolleyes:
 

Dave

Joined Nov 17, 2003
6,969
Originally posted by mozikluv@Mar 29 2004, 04:07 AM
am not a believer on the sound quality of these audio ic. its purely for convenient assembly. sorry to say that, no harm meant. if you really go for quality sound go for discretes or tubes. you are totally limited to what the manufacturer puts in and the parameters you can operate them. :rolleyes:
A common problem with all mass-produced electronic components. The real skill is knowing what tolerances your design can work within and selecting the appropriate components to maximise performance.

Its also not only a case of convenient assembly but also cost :mellow:
 

Thread Starter

mozikluv

Joined Jan 22, 2004
1,435
hi dave,

am back from a nice short vacation with my kids. yes i totally agree with you as to cost and convenience. sorry for the sarcasm, but the only thing nice about this amps is the beautiful casing in which they are housed. they really look great but never mind the sound quality. anyway at low volume level they sound okay.

ψ

by the way, any member who would like to make their own electronic source selector with phono pre-amp and dc controlled volume/tone control can email me. this devices are my working models of my low budget amps. for source selector i use the LM1037, for dc tone/vol ctrl., its TA8184P and for phono-preamp i used the LF353.
 

Thread Starter

mozikluv

Joined Jan 22, 2004
1,435
hi,

since this topic is mainly how to lay out ground lines, i would like to add this slightly off topic but still about foil traces.

when you make your own circuit board layout please "highly consider" the size of your foil trace and the distance between traces base on the supply voltage.

some of use believe that the smaller we can make the circuit board the better. that is a fallacy. board size entirely depend on particular circuit application and supply voltage.
 

Thread Starter

mozikluv

Joined Jan 22, 2004
1,435
hi to all members,

here's an info about ground lay-out which i have deliberately left out at the time i posted this topic.

the main reason i left it out is because it runs contrary to the stated #A which says "we must avoid making ground loops coz it invites unwanted signal interference"

in a star topology lay-out, running ground lines in parallel to a single ground point prevents longitudinal voltages in the ground system to couple among different stages in the circuit. using a single ground line, ripple current can induce hum in the signal part in the circuit.

what i am referring here is for usage in discrete component power amp, not in ic amps. in this kind of amp, 2 ground lines are employed, 1 is used as signal current line and the other for ripple currents from the electrolytic decoupling caps all converging into a single ground point entry.

now here's the contradiction, these two ground lines must be connected on the other end thus a ground loop has been created which again as a general rule we must avoid. this connection is made via a resistor which has a value that can prevent or stop the current from one ground line to flow into the other ground line. now this same resistor must also have a low value to keep both lines at about the same ac voltage level at higher frequencies. impedance and inductance is involve in this. at higher freq. inductance causes the ground system impedance to increase.

as electronic students i am sure you are familiar with "electromagnetic compatibility"
 

Thread Starter

mozikluv

Joined Jan 22, 2004
1,435
hi,

this additional info is slightly off topic but still ground lines are still involve.

to prevent noise propagation at any stage like the input stage, pre-driver stage or the output stage we should decouple that stage by using two (2) decoupling caps in parallel. one for low frequency by using electrolytics and one for high freq. use any monolithic caps.

now at what frequency shall we consider, for input stage, filters for 200khz would suffice and for other stages 150khz will do.

we must remember that in ground lines ac signal, noise, voltage/current transienst, external noise will always be present.
 

AENEUMANN

Joined Feb 15, 2007
2
There are several ways to arrive at a power audio amp. One way is to use a number of power audio amp IC's and wire them in parallel. The inputs are driven in parallel while the outputs will need individual limiting resistors to compensate for differences in the amplifiers' gains, offset, etc. The limiting resistors, on the order of 10 ohms, also provide protection by allowing the speaker output to be shorted without actually shorting the IC output.

Another way is to roll your own. The most stable amplifier type to use is the transconductance, meaning collector (bipolar) or drain (MOSFET) output. The output of this type of amplifier appears as a power current-source and is quite stable, producing phase-margins on the order of 100 degrees. The other thing is that the transconductance output is easily expanded later by simply plugging in more current-source stages in parallel, up to a practical limit of about 1KW. The transconductance power amp also naturally has great power bandwidth, typically 1MHz - 5MHz.
Distortion in the roll-your-own amp has to handled by controlling the upper and lower cascode currents on-the-fly in real time. The equation which produces the least third harmonic distortion is |I+| x |I-| = Ko, that is the product of the absolute-value of the upper and lower cascode currents is equal to a constant. Making this happen requires a nested control-loop which monitors the two currents, multiplies them, compares them to the constant Ko, and then feeds back to both output-current cascodes to make this happen. The outer or main voltage control loop, which delivers the proper voltage to the loudspeaker, will force one cascode to drive the loudspeaker with the higher current while the distortion-servo will adjust the idling cascode current to satisfy the equation.
It is possible to force all of the power MOSFETs of one cascode to conduct identical currents in real time by using a string of bipolar transistors like Christmas tree lights to monitor the MOSFET source currents and make small gate-voltage adjustments to reach equality. This feature will harden the amplifier to overstress failure by forcing the MOSFETs to equally share current and therefore power. Theoretically this type of output should never fail if enough transistor is provided.
The current-equalizing nested-servo is the fastest, the distortion-servo next fastest, and the main outer voltage control loo the slowest.
For examples of this type of design, refer to the horizontal yoke drivers of high-resolution CRT displays such as used in medical imaging or air-traffic control. These are basically high-reliability ultra-linear high-power wide-bandwidth power amplifers and also happen to make wonderful audio amplifiers.
 
Top