Hi guys, im trying to design a new gainclone to build, now i want to start using more components, and filters (like zobel).
I draw this.
I want to know if can give me some tips.
I just finished draw it, some values are wrong.
Thanks for your support.
I draw this.
An externally hosted image should be here but it was not working when we last tested it.
I want to know if can give me some tips.
I just finished draw it, some values are wrong.
Thanks for your support.
Why do you put 4700uF at the amps output??! Thats my biggest concern. Everything else looks pretty standard. You say that "some values are wrong" and yes they are but I will not go any further about that since you say so your self.
Back to the output-cap, again, why? One big rule is to avoid output-caps as much as you can...
Back to the output-cap, again, why? One big rule is to avoid output-caps as much as you can...
CJ900RR said:
I put 4700uf Cap because its the value that are used in the Datasheet of the LM3886. Btw how can i calculate wich value will be the best?
The V+ and V+ are decoupling diferent because i forget to Fix the V-, both will be like V+AndrewT said:
btw, what you means with the Zin Resistor? in the input im using Ri Resistor, i will use an Pre to control Volumen.
pacificblue said:
I seccond that!
The best most naturally sounding configuration regarding to my opinion is the first schematic in the datasheet for the LM3886, page 1.
For Rm use 10K, for Cs use 470uF - 2200uF depending on listening-impression plus 0,1uF at supply-pins on the LM3886 to ground.
A zobel network is a resistor plus a cap in series from output to ground, 2R7 and 0,1uF is often used there. More easy to use then the inductor that can be hard to find.
In my opinion you dont need a zobel if cables is short and only one LM3886/channel is used, but I know that other say that you must use it. Make space for it and try both with and without it.
To keep it simple is the best way to go to not affect the sound. Later on you can design a more advanced layout with DC-servo, loudspeaker-protection and so on.
Many people use the LM3886 in a wery simple way and are happy with it. Look at the kits from Audiosector, Chipamp.com etc, popular and simple.
Keep it up
Re: the best tip!
a few months ago i build an GC using the first schematic of the LM3886 Datasheet, now im trying to get it a little more complex. Im just trying to add other components to compare the sound. im really happy with the sound of my first GC. Btw, i read that adding the Zobel is an nice addon.
Ok, will remove that Cap.
I build my first gainclone without Zobel Network, this time will try to add it.
lanchile said:Well vecino! my best tip for you is "Keep it as simple as you can".if the design is not right!adding more parts does not mean you will get better sound. keep it simple ...keep it right!
a few months ago i build an GC using the first schematic of the LM3886 Datasheet, now im trying to get it a little more complex. Im just trying to add other components to compare the sound. im really happy with the sound of my first GC. Btw, i read that adding the Zobel is an nice addon.
pacificblue said:
Ok, will remove that Cap.
CJ900RR said:
I seccond that!
The best most naturally sounding configuration regarding to my opinion is the first schematic in the datasheet for the LM3886, page 1.
For Rm use 10K, for Cs use 470uF - 2200uF depending on listening-impression plus 0,1uF at supply-pins on the LM3886 to ground.
A zobel network is a resistor plus a cap in series from output to ground, 2R7 and 0,1uF is often used there. More easy to use then the inductor that can be hard to find.
In my opinion you dont need a zobel if cables is short and only one LM3886/channel is used, but I know that other say that you must use it. Make space for it and try both with and without it.
To keep it simple is the best way to go to not affect the sound. Later on you can design a more advanced layout with DC-servo, loudspeaker-protection and so on.
Many people use the LM3886 in a wery simple way and are happy with it. Look at the kits from Audiosector, Chipamp.com etc, popular and simple.
Keep it up
I build my first gainclone without Zobel Network, this time will try to add it.
Now Redraw it with some tips that you give me.
can someone tellme chich is the purpose of Ri2?
THanks in advance.
An externally hosted image should be here but it was not working when we last tested it.
can someone tellme chich is the purpose of Ri2?
THanks in advance.
Cz and Rz as you draw them are not a Zobel network.
In the datasheet on page 6 there is the Zobel (or Boucherot) network composed of Rsn and Csn.
The Zobel network has little to no influence on the sound. It is a snubber that helps to prevent the amplifier from oscillating, especially in situations, when no load is connected to the output. It is highly recommendable to have this.
Lo and Ro from your previous post may be necessary, if your speakers are highly capacitive, e. g. magnetostats or with unusual crossover design. If you are sure that you won't connect speakers of that kind you may skip this.
In commercial amps you usually find both circuits, because a manufacturer has to take in to account that any kind of speaker can be connected to the amp. And to avoid guarantee claims manufacturers prefer to walk on the safe side.
In the datasheet on page 6 there is the Zobel (or Boucherot) network composed of Rsn and Csn.
The Zobel network has little to no influence on the sound. It is a snubber that helps to prevent the amplifier from oscillating, especially in situations, when no load is connected to the output. It is highly recommendable to have this.
Lo and Ro from your previous post may be necessary, if your speakers are highly capacitive, e. g. magnetostats or with unusual crossover design. If you are sure that you won't connect speakers of that kind you may skip this.
In commercial amps you usually find both circuits, because a manufacturer has to take in to account that any kind of speaker can be connected to the amp. And to avoid guarantee claims manufacturers prefer to walk on the safe side.
An externally hosted image should be here but it was not working when we last tested it.
What about that?
Ri2 forms the input resistance of the amplifier, i. e. the resistance your source sees and must be able to drive. If you don't have that or a potentiometer then the input impedance will be extremely high (several MOhms). That has influence on impedance matching, on the sources output filter, where applicable, on cable termination and so on. Please google that or use the search function in the forum, if you are interested in that topic.
Ri=R2 and Ri2=R3 strips the signal of DC and improves common mode rejection. That is a recommendable configuration, if you want to skip Ci and Cin.
Ri=R2 and Ri2=R3 strips the signal of DC and improves common mode rejection. That is a recommendable configuration, if you want to skip Ci and Cin.
samsagaz said:
Why did you put Ro1 and Ro2 in? Are you planning on a parallel configuration? Then they should be in, and then Ri, Ri2, R2 and R3 should be matched to 0,1%, Ro1 and Ro2 to 1%.
If you don't use a parallel configuration, leave Ro1 and Ro2 out, because in that case they only waste precious output power and have no function at all.
pacificblue said:
If i use Ri too high (20K like Ri2) i dont will loose input power? Im using the PC as PRE.
Btw, will google abt that too see if can get more tips.
pacificblue said:
I used both res in paralell because i see some schematics that use 0R22 Ohms at output but rated at 3watts, so recommend using 0R47 carbon film resistors in parallel to avoid to use 3W wire-wound resistor .
Btw, if this waste power i will leave it out.
On the contrary. The higher the resistance, the lower the current your source has to deliver. Just like with speakers. 4 Ohm speakers are more critical to drive than 8 Ohm speakers.samsagaz said:
However the input resistance must also not be too big, because it also has to match the output impedance of your source. 20k should be okay
Yes, for parallel operation to assure equal load for all output stages. Not for single operation.samsagaz said:
No. At audio frequencies Cz poses a high resistance that does not consume significant power.samsagaz said:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.