Yep. That'll usually turn into a hair ball.
Yeah... But then you lose the advantage of having amps in parallel. Granted, the LM4780 does have two amps in one package, but you will have a very hard time with the thermal design in a parallel/bridged application.
What matters is the impedance of the traces. The impedance must be minimized. For audio frequencies, this means the resistance and inductance must be minimized. Use W-I-D-E traces or, preferably, planes and pours as these offer the lowest impedance.
Although not officially on-line yet, my "Taming the LM3886 Chip-Amp" page is nearing completion. I think a lot of builders will benefit from the content there.
~Tom
Thanks Guys you give me a very informative comments;
I have some question about building IC base audio amplifiers;
As of now, as far as I know;
:To do on PCB:
- Spider grounding (also known as star grounding)
-Thick traces in ground
Q1: I have difficult time to find the 1% tolerance resistor here in Philippines. the very common here is 5% tolerance,is the 1% tolerance is really important?
Q2: I already have center-tapped Trafo 25-0-25 8-Amperes, is the 8 amperes is good enough for bpa200?
and I need more additional info.
thanks again
I have some question about building IC base audio amplifiers;
As of now, as far as I know;
:To do on PCB:
- Spider grounding (also known as star grounding)
-Thick traces in ground
Q1: I have difficult time to find the 1% tolerance resistor here in Philippines. the very common here is 5% tolerance,is the 1% tolerance is really important?
Q2: I already have center-tapped Trafo 25-0-25 8-Amperes, is the 8 amperes is good enough for bpa200?
and I need more additional info.
thanks again
Last edited:
5% may be indicative of lower quality rather than only wider tolerance.
If they can't make +-1% which are very readily available in the E24 series then that probably tells you that their prodcution process is very variable.
Are they metal oxide or metal film?
What is the temperature coefficient?
What tests does it pass? There should be a long string of specifications.
If they can't make +-1% which are very readily available in the E24 series then that probably tells you that their prodcution process is very variable.
Are they metal oxide or metal film?
What is the temperature coefficient?
What tests does it pass? There should be a long string of specifications.
It's metal film, I'm not prepare smd type of resistor.(also available here) the test is sometimes < or/and > by 2 to 8ohms. anyways the store that I visited is selling made in china products, so expected to be low quality. I try to go to a certain place where there is a store that sell locally made products.
how about the trafo sir 25v-0-25v 8-Amperes, is the 8 amperes is good enough for bpa200?
thanks
how about the trafo sir 25v-0-25v 8-Amperes, is the 8 amperes is good enough for bpa200?
thanks
Alexan in Megamall ang Raon sells 1% metal film resistor for 1 or 2 php each,where you from? im from the south of manila, i can give you a few metal film 1% resistor if you live near my place.
If you don't have or can't get lower tolerance resistors you can always measure each one and hand match the ones you need.
The object is to have each stage have exactly the same gain as the rest of them.
I have lots of 5% values as well in my parts bin and for some projects I have to do this a lot in order to get the correct value if it is of a precision design such as this.
It makes no difference if the gain is 22,23,25 or even 30 or more, all that matter is that they are all the same.
And using .1% resistors helps to assure that right off of bat without worrying about it and having to go through the painstaking method of hand matching.
jer
The object is to have each stage have exactly the same gain as the rest of them.
I have lots of 5% values as well in my parts bin and for some projects I have to do this a lot in order to get the correct value if it is of a precision design such as this.
It makes no difference if the gain is 22,23,25 or even 30 or more, all that matter is that they are all the same.
And using .1% resistors helps to assure that right off of bat without worrying about it and having to go through the painstaking method of hand matching.
jer
From own experience, 0,1% tolerance resistors is required. Using 1% can work 5% is just stupid. You can of course take a bunch of 5% resistors and handmatch them to 0,1%. I usually do that with 1% resistors because they are cheaper and more easy to get in the correct value. 215k and 21,5k for example is a pain to get in 0,1% tolerance. You can always change the gain (and thereby the values) to resistors that's more easy to get. But then you also have to change the resistors in the DC servo to match.
If you have a lot of LM3886 you can always pick the ones closest in offset. But then you have to test them individually.
And always test the board first without the output resistors connected. Then you can measure all 4 LM3886 before connecting them together to see that they are close in offset and that they will not fight each other into oscillation.
If you have a lot of LM3886 you can always pick the ones closest in offset. But then you have to test them individually.
And always test the board first without the output resistors connected. Then you can measure all 4 LM3886 before connecting them together to see that they are close in offset and that they will not fight each other into oscillation.
A Hamon divider could be built that with care gives better than 0.1% ratio
.
That divider is then set in one side of a Wheatstone bridge to check against the other side for 9:1.
But you need to add an extra 1 to the 9 of the 9:1 to give a 10:1, for comparison to the 215:21.5
That is easy. just measure an extra 1 when one is selecting the resistors for the Hamon.
.
That divider is then set in one side of a Wheatstone bridge to check against the other side for 9:1.
But you need to add an extra 1 to the 9 of the 9:1 to give a 10:1, for comparison to the 215:21.5
That is easy. just measure an extra 1 when one is selecting the resistors for the Hamon.
I have been using a precision voltage sources from LT and AD and the ma or ua setting on my DMM.
I just ordered some precision resistors from LT to use as standards for such measurements and calibrations, eventually I will set up a wheatstone bridge as well.
http://www.linear.com/parametric/Precision_Resistor_Network
I was able to get accuracy's of better than down to 1 ohm with high valued resistors using such methods as only my DMM's resolution of 4 digits was the limiting factor.
I found that Visual Analyzer provided me with a much higher resolution than what I could get with my DMM.
I had to do this when I made the voltage divider for my Variable HV Bias supply and also an HV divider to measure my step-up transformers and amplifier outputs.
jer
I just ordered some precision resistors from LT to use as standards for such measurements and calibrations, eventually I will set up a wheatstone bridge as well.
http://www.linear.com/parametric/Precision_Resistor_Network
I was able to get accuracy's of better than down to 1 ohm with high valued resistors using such methods as only my DMM's resolution of 4 digits was the limiting factor.
I found that Visual Analyzer provided me with a much higher resolution than what I could get with my DMM.
I had to do this when I made the voltage divider for my Variable HV Bias supply and also an HV divider to measure my step-up transformers and amplifier outputs.
jer
Last edited:
If one uses +-1% resistors, then a Hamon divider will give a divided output that is accurate to better than +-100ppm
Use +-0.1% resistors and the divided output is accurate to <1ppm.
You don't need to buy high accuracy resistors.
+-0.05% matching resistors get the Hamon divider ratio accuracy to better than 250ppb
This is getting to the point that one needs to take account of soldered joint resistances and the effect of soldering on the resistor values and the long term drift of resistors. Baking at 100°C to remove the short term drift may help.
If you can find ways of reducing these effects, then
+-0.02% get to <40ppb
+-0.02% is equivalent to 1part in 5000 and is easily within reach for a 4½digit resistance meter.
Use +-0.1% resistors and the divided output is accurate to <1ppm.
You don't need to buy high accuracy resistors.
+-0.05% matching resistors get the Hamon divider ratio accuracy to better than 250ppb
This is getting to the point that one needs to take account of soldered joint resistances and the effect of soldering on the resistor values and the long term drift of resistors. Baking at 100°C to remove the short term drift may help.
If you can find ways of reducing these effects, then
+-0.02% get to <40ppb
+-0.02% is equivalent to 1part in 5000 and is easily within reach for a 4½digit resistance meter.
Last edited:
Very Cool !!!
I just looked that up and I was using those very same techniques and I had no idea this was called a Hamon Divider!!
http://conradhoffman.com/HamonResistor.html
Yes, I was able to get very accurate measurements with the series method once I had found two resistors that were exactly the same.
jer
I just looked that up and I was using those very same techniques and I had no idea this was called a Hamon Divider!!
http://conradhoffman.com/HamonResistor.html
Yes, I was able to get very accurate measurements with the series method once I had found two resistors that were exactly the same.
jer
Last edited:
thanks for all Response;
Thanks bro;
too bad there's no 12 amperes trafo in decco only 8amp is available, and all resistor is 5% tolerance, also in alexan raon, may be I try in e-gizmo or megamall.
I gathering first all components before I scheming the PCB.
as of now I have
- 20 pcs lm3886TF
-a couples of capacitor 4x-10,000/50v, etc,
-a couples of double-sided PCB and single-sided PCB,
-8-amperes trafo 25v-0-25v,
-30amperes packages bridge-diode
-1 DIY audio Amplifier housing
-LED green,light-green,yellow,red, 20pcs each, for VU LED
-opa134pa 10pcs
-LM3915 2pcs
-1A -+15v Regulator 7815,7915 4pcs
-100k double potentiometer for volume, 50k trimmer for vu LED
-1N456A 24pcs
-220AC metal seesaw switch 2pcs.
- and some connectors, screw,socket, fuses 5A, 1.5A,1A
-ferric chloride
Thanks bro;
too bad there's no 12 amperes trafo in decco only 8amp is available, and all resistor is 5% tolerance, also in alexan raon, may be I try in e-gizmo or megamall.
I gathering first all components before I scheming the PCB.
as of now I have
- 20 pcs lm3886TF
-a couples of capacitor 4x-10,000/50v, etc,
-a couples of double-sided PCB and single-sided PCB,
-8-amperes trafo 25v-0-25v,
-30amperes packages bridge-diode
-1 DIY audio Amplifier housing
-LED green,light-green,yellow,red, 20pcs each, for VU LED
-opa134pa 10pcs
-LM3915 2pcs
-1A -+15v Regulator 7815,7915 4pcs
-100k double potentiometer for volume, 50k trimmer for vu LED
-1N456A 24pcs
-220AC metal seesaw switch 2pcs.
- and some connectors, screw,socket, fuses 5A, 1.5A,1A
-ferric chloride
Last edited:
Ohh the transformer from deeco is not so good for audio amp,I suggest you visit roks they are just located on the other side of gonzalo puyat across rizal ave,i usually ask them to custom made a transformer if they dont have a stock,their price is fairly reasonable.keep us posted on your build.
Sent from my Desire HD using Tapatalk 2
Sent from my Desire HD using Tapatalk 2
Great... Scope creep... Rookie mistake #17.
Dude. I hope you test it in blocks rather than a big bang test.
~Tom