Hi Lars,
I am posting photo of the layout , which I made on graph paper , of the output section of the ZETA , with 6 pairs of output devices , most of the space is takenup by the source 5w resistors , I have also added resistors from the Source which will go to the protection circuit , the input section can be added to this .
This the component side of the board , the lines in brown are for the copper track , blue are the components , and green are jumpers if required . G is for gate , P is for protection , FB for feedback .
The mosfets can be mounted on bothsides of the pcb horizontalily or vertically depending on the heatsink .
I am posting photo of the layout , which I made on graph paper , of the output section of the ZETA , with 6 pairs of output devices , most of the space is takenup by the source 5w resistors , I have also added resistors from the Source which will go to the protection circuit , the input section can be added to this .
This the component side of the board , the lines in brown are for the copper track , blue are the components , and green are jumpers if required . G is for gate , P is for protection , FB for feedback .
The mosfets can be mounted on bothsides of the pcb horizontalily or vertically depending on the heatsink .
Sorry I was unable to post the layout , I tried many times and failed , I think he size of attachment should have been smaller , however I have sent it by e-mail to Lars and have requested him to post it with his comments
As Lars has offered to make the boards with components I think that will be THE BEST
Rajeev
As Lars has offered to make the boards with components I think that will be THE BEST
Rajeev
Thanks Lars!
Nice way indeed one can use as many FETs as he may want. I would surely like to give it a try.
Thanks to all.
Regards
Rahul
Nice way indeed one can use as many FETs as he may want. I would surely like to give it a try.
Thanks to all.
Regards
Rahul
What's the advantage?
Whats the advantage of having only 1 connection between VAS and the next stage? In LC's Zeta, The VAS's output (T9-T10) is only connected by 1 line to the next stage (R15-R16). Ordinary amp put Vbe multiplier between T9-T10, so there is 2 connection in upper and lower section of Vbe multiplier to the next stage.
I encountered this "1 point connection" also in an commercial amp. But I dont know what is the advantage of this, because it will add more component for the next stage to translate full swing.
Whats the advantage of having only 1 connection between VAS and the next stage? In LC's Zeta, The VAS's output (T9-T10) is only connected by 1 line to the next stage (R15-R16). Ordinary amp put Vbe multiplier between T9-T10, so there is 2 connection in upper and lower section of Vbe multiplier to the next stage.
I encountered this "1 point connection" also in an commercial amp. But I dont know what is the advantage of this, because it will add more component for the next stage to translate full swing.
Lumanauw: The voltage across the Vbe multiplier is going to be missing from the voltage swing of the entire amplifier. In a typical bipolar output stage the loss is in order of 2-4 Volts, so often ignored. Here you would see something like 8 Volts, so it becomes more of a problem.
In other words by making the 1-line connection you get an extra 8 Vpp of output swing.
🙂
In other words by making the 1-line connection you get an extra 8 Vpp of output swing.
🙂
LC,
You're right. Is there any side good/bad effect on sound quality?
ps: Would you care to explain why you designs non-feedback amp so much?
You're right. Is there any side good/bad effect on sound quality?
ps: Would you care to explain why you designs non-feedback amp so much?
Hello Lars,
You wrote ;--
"I suggest you post your layout here in the thread, so we can all try and help you optimize it before you go ahead and build it."
Now you have an idea of the output section layout I had in mind , I was only worried of the length of the track going to the gates of the output devices , please let me know if this is OK
I have a few more questions
1, Holton in his 1000w 4 ohms has used 10 pairs of devices in the output inspite of fan cooling , will 6 pairs in the Zeta be enough , I hope the the temprature of the output devices at 1000w 4 ohms , will not be too high .
2, can we use wirewound resistors in the source of the output mosfets and R54 of the zobel network.
3, Is there is no need of a RFC choke in the output , we see it in most poweramps ??
With Regards
Rajeev
You wrote ;--
"I suggest you post your layout here in the thread, so we can all try and help you optimize it before you go ahead and build it."
Now you have an idea of the output section layout I had in mind , I was only worried of the length of the track going to the gates of the output devices , please let me know if this is OK
I have a few more questions
1, Holton in his 1000w 4 ohms has used 10 pairs of devices in the output inspite of fan cooling , will 6 pairs in the Zeta be enough , I hope the the temprature of the output devices at 1000w 4 ohms , will not be too high .
2, can we use wirewound resistors in the source of the output mosfets and R54 of the zobel network.
3, Is there is no need of a RFC choke in the output , we see it in most poweramps ??
With Regards
Rajeev
Rajeev: First question no. 0: 😀
0..Yes it looks fine, if the tracks get too long, you can stabilize the amp by using slightly higher gate resistors instead of 100 Ohms.
1..6 pairs will be enough if you follow my suggestion to mount directly on the anodized surface. If you are going to use silicone or MICA insulators, better use 10 pairs instead.
2..Source: yes, Zobel: No use MOX resistor.
3..I didn't encounter any problems with different capacitive loads.
If you want to be sure in all cases, just drop in the RFC anyway.
Another substitute for the RFC is a 10pF cap from the gain stage back to the feedback point. This way the output stage can't see capacitive loads, in the feedback loop, and thus becomes less sensitive to the problem that a RFC solves. (But NOT immune).
My claim is though that you CAN avoid the RFC by designing the circuit layout with your mind on stability at high capacities. (I have never built an amplifier with a RFC network in my life, and i'm sure not gonna start now 😉 ).
However i can not give you exact instructions that will guarantee that result.
0..Yes it looks fine, if the tracks get too long, you can stabilize the amp by using slightly higher gate resistors instead of 100 Ohms.
1..6 pairs will be enough if you follow my suggestion to mount directly on the anodized surface. If you are going to use silicone or MICA insulators, better use 10 pairs instead.
2..Source: yes, Zobel: No use MOX resistor.
3..I didn't encounter any problems with different capacitive loads.
If you want to be sure in all cases, just drop in the RFC anyway.
Another substitute for the RFC is a 10pF cap from the gain stage back to the feedback point. This way the output stage can't see capacitive loads, in the feedback loop, and thus becomes less sensitive to the problem that a RFC solves. (But NOT immune).
My claim is though that you CAN avoid the RFC by designing the circuit layout with your mind on stability at high capacities. (I have never built an amplifier with a RFC network in my life, and i'm sure not gonna start now 😉 ).
However i can not give you exact instructions that will guarantee that result.
Thanks Lars ,
1, What is the max value of gate resistor we can go upto without it causing any cross conduction , can we go upto 330 ohms,
2, I do not know if I will get metaloxide resistor of that value for the zobel , but I can get a carbon resistor , I will use two 1 ohm 2w in series , even for the source will it be better to use three carbon 1 ohm 2w in parallel for each mosfet ??.
3, you meen a 10pF across R19 ,
Now can I go ahead and paint the pcb .
Rajeev
1, What is the max value of gate resistor we can go upto without it causing any cross conduction , can we go upto 330 ohms,
2, I do not know if I will get metaloxide resistor of that value for the zobel , but I can get a carbon resistor , I will use two 1 ohm 2w in series , even for the source will it be better to use three carbon 1 ohm 2w in parallel for each mosfet ??.
3, you meen a 10pF across R19 ,
Now can I go ahead and paint the pcb .
Rajeev
Thanks Djk ,
Please give an example with values
Hi Lars ,
Could you please test your prototype of the Zeta with a 330 ohm gate resistor in the output mosfets and with Djk,s recomendation
Rajeev
Please give an example with values
Hi Lars ,
Could you please test your prototype of the Zeta with a 330 ohm gate resistor in the output mosfets and with Djk,s recomendation
Rajeev
At work I use an IRF driven with 390R and a 10R that has a glass signal diode (sorry, can't remember the #, but it is similar to a 1N4148) in series with it. The diode only conducts on turn-off.
The values were optimized for the layout by using a 7µS pulse and going for the least amount of overshoot and ringing on both the leading and trailing edge.
The values were optimized for the layout by using a 7µS pulse and going for the least amount of overshoot and ringing on both the leading and trailing edge.
Rajeev: Sorry i have disassembled my prototype for now. Next step will be a full scale PCB. For freq's below 50 kHz i think 330
Ohms will work fine.
The resistor / diode solution might be good here, even though it seems like solving a problem that isn't really there. This design is not very critical, and will not easily make cross conduction.
🙂
Ohms will work fine.
The resistor / diode solution might be good here, even though it seems like solving a problem that isn't really there. This design is not very critical, and will not easily make cross conduction.
🙂
Thanks djk ,
Thanks Lars,
I was thinking what to do regarding the protection ckt , is it ok to sink the emitters of T12 & T13 to the centre , in the +ve side the current can be senced by the voltage across the Drain resistors and fed to the base of an npn transistor , lets call it T27 which will sink the emitter of T12 to the centre , however in the -ve side the voltage across the Drain resistors can be fed to the base of another npn transistor T28 , emitter of which is going to -ve supply and collector to base of a pnp transistor T29 which can sink the emitter of T13 to the centre . The values or resistors will be quite critical hence please help out here .
I would like this type of protection to a timer based protection as in PA if the amp turns off and you have to wait for the timer to reset its quite a neiucance .
Thanks Lars,
I was thinking what to do regarding the protection ckt , is it ok to sink the emitters of T12 & T13 to the centre , in the +ve side the current can be senced by the voltage across the Drain resistors and fed to the base of an npn transistor , lets call it T27 which will sink the emitter of T12 to the centre , however in the -ve side the voltage across the Drain resistors can be fed to the base of another npn transistor T28 , emitter of which is going to -ve supply and collector to base of a pnp transistor T29 which can sink the emitter of T13 to the centre . The values or resistors will be quite critical hence please help out here .
I would like this type of protection to a timer based protection as in PA if the amp turns off and you have to wait for the timer to reset its quite a neiucance .
"The resistor / diode solution might be good here, even though it seems like solving a problem that isn't really there. "
It may not be a problem here. I mention it as a fix if someone does have a problem. The amplifier it fixed for me was a modulator amp for a radio transmitter. It really cleaned up the spectrum for me.
It may not be a problem here. I mention it as a fix if someone does have a problem. The amplifier it fixed for me was a modulator amp for a radio transmitter. It really cleaned up the spectrum for me.
djk: aaa ok 😉 Now i understand.
Rajeev: Yes you can do that with no problem, but this 'old fashion' kind of safety system only has a couple of downsides.
In case of a short circuit, the amplifier will be running very hot. It is running at it's highest possible output power (and not higher), into a short circuit.
The other thin is a more hifi oriented thing, that the peak current is limited rather close to normal operation current, and so a little bit of bass control may be sacrificed. The protection circuit is audible, some would say.
The pulse mode protection can be safely set at a much higher level of peak current.
The timer does not have to delay more than say 0.2 sec, so you will hardly notice any occational downtime. At severe overload you will/should of course notice the problem.
Rajeev: Yes you can do that with no problem, but this 'old fashion' kind of safety system only has a couple of downsides.
In case of a short circuit, the amplifier will be running very hot. It is running at it's highest possible output power (and not higher), into a short circuit.
The other thin is a more hifi oriented thing, that the peak current is limited rather close to normal operation current, and so a little bit of bass control may be sacrificed. The protection circuit is audible, some would say.
The pulse mode protection can be safely set at a much higher level of peak current.
The timer does not have to delay more than say 0.2 sec, so you will hardly notice any occational downtime. At severe overload you will/should of course notice the problem.