F5 power amplifier

[R-K Rønningstad]

yes, and the output stage shold follow suit.

 

[Melon Head]

I just want to make a comment about low impedance speakers.
According to simulation (FFT) the distortion proflile doesn't look to pretty into 4Ohm loads at higher power output (ie 20W RMS/Channel and above).
I have simmed an F5 version with 3 Paralleled output pairs which improves things significantly.


Here is an image of the results.

 

[bobodioulasso]

s

I have simmed an F5 version with 3 Paralleled output pairs

Don't you increase distortion at low level, as bias is divided by three ?

 

[Melon Head]

No, distortion does not increase at low levels.

In fact it is better at low levels.


At 10 watts 3 Paralleled version is on left. (4Ohm Load)

 

[bobodioulasso]

10 watts is not exactly a low level in a home environnement.
If a live experiment could confirm this , it would help for spreading dissipation too.

 

[Melon Head]

OK then here is 1Watt into 4Ohm load.

 

[Babowana]

I usually double this rating for a 2x100w class AB amp (bottom in multi-amplification). Though, who will use F5 on a 2 ohm load? Is'nt the limiting current circuit set too low for this ( i do not remember, as i do not use any)?

F5 provides peak ac current (half-wave) of max. 10A, i.e. pk-pk of max. 20A . . . So, the class B operation upto 100W@2-ohm load is possible. Who will use this . . . ? I have no idea . . .

Of course, I like to use "big" transformers . . .

 

[bobodioulasso]

OK then here is 1Watt into 4Ohm load.

And what about this config and a 8ohm load?
Is it better or worse than the original?

Two pairs should be enough in this case and providing good sharing/spreading dissipation.

 

[cviller]

I just want to make a comment about low impedance speakers.
According to simulation (FFT) the distortion proflile doesn't look to pretty into 4Ohm loads at higher power output (ie 20W RMS/Channel and above).
I have simmed an F5 version with 3 Paralleled output pairs which improves things significantly.


Here is an image of the results.

Have you checked that it is not just the extra gate capacitance which lowers the cut frequency in the low pass filter?

 

[Melon Head]

It is also better into 8 ohm loads.

However I should say there is a cost to this: 3 x the total power dissipation = bigger heat sinks, and bigger transformers (ie VA rating)

Just so there is no confusion bias current through each output mosfet is the same as original version.

Measurements were performed at 1000Hz.

 

[AndrewT]

I usually double this rating for a 2x100w class AB amp

1kVA for 100W+100W ClassAB is unusually high. I don't think there are many here who would recommend that VA to W ratio and few builders who would follow that advice.

I do not give any advice, just telling what i do.

For a 2x 100w/8 ohms capable of 2x400w/2 ohms

so which is it that you use?
a 100W+100W amplifier powered by a 1kVA transformer
or
a 200W+200W amplifier powered by a 1kVA transformer
or
a 400W+400W amplifier powered by a 1kVA transformer.

I will answer "advice" by private Email, it's too far off topic to expand here.

 

[tinitus]

Multiply by factor 10, and you are in the clear

But that gets exceedingly difficult with high power amps

Maybe one reason to choose smaller amps

 

[bobodioulasso]

Just so there is no confusion bias current through each output mosfet is the same as original version.

I was understanding 1,3/3 amps each, sorry...
So your proposal is power hungry.

 

[bobodioulasso]

Andrew,
Large transformers having a better regulation, the rails are stiffer, giving a tighter response which is a plus , something like an iron fist, never mind calculations.
Just make it big and enjoy!

 

[Melon Head]

Yes it is power hungry, but it is not extreme.
It certainly is still doable. You would need 500VA transformer per channel.
The the heatsinks would need to be something like this:
MF35-151.5 http://www.conradheatsinks.com/products/flat100_350.html#MF35

 

[bobodioulasso]

The the heatsinks would need to be something like this: MF35-151.5 http://www.conradheatsinks.com/prod...0_350.html#MF35

These are just enough for a normal F5. So you'll need three pairs of them here for stereo.

 

[Melon Head]

There is 30 Watts (heat dissipation) per device, multiply by 6, and you get 180Watts total per channel.
Hence the heat sinks suggested earlier should be adequate.

I know this can be done because I have a JLH class A which dissipates 160W (heat) per channel, and the heat sinks I have are smaller. and I can touch the heatsinks no problem (temp is below 60 degrees Celsius)

Well to be on the safe and conservative side you could buy 4 and make a cube just like the Aleph 3 http://calab.kaist.ac.kr/~cgkim/audio/amp/zen/www.passlabs.com/images/a3_lg.jpg

 

[AndrewT]

Large transformers having a better regulation, the rails are stiffer,
Just make it big

enough.
Yes, I agree.

But you have chosen to not answer the question I asked.
What size of two channel amplifier is your 1kVA transformer driving?

 

[56oval]

Hi All

A mate of mine gave me a pair of F5 boards (peter danials design)
I check out diy-tech site and they don't have matched pairs of the jfets ,how important is it to have these matched .

Cheers

Mal

 

[bobodioulasso]

According to N.P. recommendations, it is not necessary to match them, just use the same group. ( GR, BL...).