Power amp under development

jethari said:
Hi Quasi, Key and everybody else,
Here's another version of my pcb design. I've managed to get the board width down to 4.05 inches. Take a look at it and tell me what you feel about it.
Hari

Hi Hari,

I have the following comments:

1) The track side copyright is now out of the board area. If you rotate the text 90 degrees and place is in the space between V POS and PCB edge there is likely enough space to do so.

2) The space between C7B and T7 is very very tight. I think it presents two issues. The first is the ability to access T7 so you can use a screw to attached it to a heatsink T6 and T7 will share. Second is thermal considerations. T6 and T7 will run very warm, which is a given, but that heat is going to radiate very much so to C7B and possibly cause C7B to fail in short time.

3) L1 and R35 should share the same point of connection, not around and about the trace as it is now. I would move L1 and the associated output pad to the space above C13 as the space above C11 and C13 look to be same.

4) T8 is not centered again, close, but not centered as I believe you once had centered in last PCB.

5) R23 is a smaller footprint than R24/R26/R27 are.

Did you want to include the Leach Feedforward NFB option to your PCB design?

I think I am forgetting a point or two, but all I can recall right now.


Regards,

John L. Males
Willowdale, Ontario
Canada
29 August 2007 (18:45 -) 19:36
Official Quasi Thread Researcher
 
Hi,
it's a pity V+ & V- are so far apart and also so remote from power ground.
I think these should form a tight loop.

Do you realise you can eliminate that link above T5?
Take the middle pin trace up, then left, then between the cap pins.
Take the left pin trace down and then right, then to the pot. It's also a shorter route.
 
It's very hard to mount T8. I can't fit it exactly. I wander ... what if we put all MJE's and T8 on the main heatsink .... ? and instead T8 something linke MJE to make a better contact ? MJE340 has lower amplification... it's that a problem if I want to use it instead T8 ? or other common transistor ?

I don't know if I have to solder the fet's on the PCB with a space between or without. I'm afraid if they are in contact with PCB will heat it up too, and maybe damage it. (now they are on the board with no space)
I don't use that PCB with glass fibre.

Quasi, what do you mean with Nmos350 and Nmos500 ? 350 and 500 it's the power on 8 ohms ?
In the Power Selection Guide I see that the only amp with 6 fet's (8ohms) have 350W - this seems to be Nmos350 , but here is powered by 85V rails and the schematics says 75V rails. I am a little confused ...
And where is Nmos500 ? Again, looking in PSG I see that only amp with 10 fet's it's powered by the same 85V rails and has 600 W but on 4 ohms ...
 
keypunch said:


Hi Hari,

I have the following comments:

1) The track side copyright is now out of the board area. If you rotate the text 90 degrees and place is in the space between V POS and PCB edge there is likely enough space to do so.

2) The space between C7B and T7 is very very tight. I think it presents two issues. The first is the ability to access T7 so you can use a screw to attached it to a heatsink T6 and T7 will share. Second is thermal considerations. T6 and T7 will run very warm, which is a given, but that heat is going to radiate very much so to C7B and possibly cause C7B to fail in short time.

3) L1 and R35 should share the same point of connection, not around and about the trace as it is now. I would move L1 and the associated output pad to the space above C13 as the space above C11 and C13 look to be same.

4) T8 is not centered again, close, but not centered as I believe you once had centered in last PCB.

5) R23 is a smaller footprint than R24/R26/R27 are.

Did you want to include the Leach Feedforward NFB option to your PCB design?

I think I am forgetting a point or two, but all I can recall right now.


Regards,

John L. Males
Willowdale, Ontario
Canada
29 August 2007 (18:45 -) 19:36
Official Quasi Thread Researcher


John
Thanks for the quick and concise observations.
1. I missed the copyright part. That will be shifted into the board.
2. I'm shifting C7B away from the transistor, more to the left and also little down.
3. You're right, I can very well shift L1 to the right side, the space available is the same.
4. Centering T8 is not an issue - there's a lot of space there.
5. R23 has the correct footprint for the given value. I've used larger footprints for the other three just to get across the large tracks/lands.

As for the Leach feedforward, I think I'll skip it for the time being. Maybe in a subsequent build I might just include it.

Thanks again
Hari
 
AndrewT said:
Hi,


Do you realise you can eliminate that link above T5?
Take the middle pin trace up, then left, then between the cap pins.
Take the left pin trace down and then right, then to the pot. It's also a shorter route.

Andrew
Thanks. You're dead right there, never looked at it that way:smash:
I've already done the changes, will be posting the updated version soon.
Cheers
Hari
 
Hi Marus

The 4 transistors mounted on the little heatsink will run very warm, some say hot. Do not worry about this, the transistors selected for these stages can take it. Constructors who use 85 volt rails will find these transitors will get hot and they should consider a larger heatsink. I run +/- 73 volt rails with a heatsink similar to yours and mine run very warm. Just make sure your case is well ventilated.

Regarding the output transistors touching the PCB, this is what I intended when I designed the board layout. The output FETs double as a PCB standoff and this allows the PCB to be quite compact. Most A/B amps that I have seen with this sort of power are much bigger. The average temperature of the FETs is not that high and well under what your PCB can handle.

Cheers
Q
 
Warning - Nmos200

A fellow constructor has identified an error in the PCB layouts for both versions of the Nmos200. Capacitor C12 was shown reversed biased. Anyone who did not notice this and correct it during their build would have had a dramatic failure of this capacitor. All new constructors should look at the layouts again and note the correction.

The corrected layouts are on my web site accessible from the Nmos200 link.

Sorry about any inconvenience (or surprises) caused. :scratch1:

Another small correction is the change in the value of R11 in the schematics for both versions of the Nmos200. The value of this resistor is now 22K.

Cheers
Q
 
Re: Warning - Nmos200

quasi said:
Another small correction is the change in the value of R11 in the schematics for both versions of the Nmos200. The value of this resistor is now 22K.

Cheers
Q

Quasi,

I am curious why you feel R11 Value for the NMOS200 needs to be changed from 27K to 22K? Should R3 then need to be changed to 22K? I know changing R3 affects the input roll off frequency, but for the purposes of the R11 and R3 question I am ignoring C1.


Regards,

John L. Males
Willowdale, Ontario
Canada
30 August 2007 22:48
Official Quasi Thread Researcher
 
Hi,

I'm finished drawing Nmos350. And change place between T9 & T6. No have jumper, If it's Ok I will make PCB. Can you help me find mistake.

Thank you very much.
 

Attachments

  • nmos350v2.jpg
    nmos350v2.jpg
    36.2 KB · Views: 757
Re: Re: Warning - Nmos200

keypunch said:


Quasi,

I am curious why you feel R11 Value for the NMOS200 needs to be changed from 27K to 22K? Should R3 then need to be changed to 22K? I know changing R3 affects the input roll off frequency, but for the purposes of the R11 and R3 question I am ignoring C1.


Regards,

John L. Males
Willowdale, Ontario
Canada
30 August 2007 22:48
Official Quasi Thread Researcher

R11 provides some bias to turn diodes D1&D2 on. These diodes set a voltage of around 1.2 volts which is the reference for the two constant current sources T4 & T7.

It was always my intention to have this at 22K for the Nmos200 and this is the value I used in the PCB layout, but forgot to change it in the schematic. The reason I lowered it from 27K as used in the Nmos350 / 500 was to maintain sufficient current with the generally lower rail voltages suggested for the Nmos200.

This resistor is not related to the feedback or input part of the schematic at all. Indeed it has nothing to do with the signal path.

Cheers
Q