F5 power amplifier

[Melon Head]

Hi alazira
what is your question specifically?

[juma]

Originally Posted by alazira

....
If you could be so kind as to identify the 6 elements you used in your calculation....

Resistors in upper half of feedback loop (R5, R7 and R1) have equivalent resistance of 60 Ohms (R5 parallel R7 plus R1) and they are connected from output to ground i.e. parallel to the speaker. They (together) get the same output voltage as the speaker does. R1 (10R) has one sixth of that "chain" resistance and that's why he'll see one sixth of the output signal - if we have 20V on the speaker it means that there will be 20/6=3,33V over R1. P=(V*V) / R = (3.33 *3.33) / 10 = 1.11 W = power dissipated on R1.
The same goes for lower feedback network (R8, R6 and R2).
Got it?

[alazira]

Originally Posted by juma

Resistors in upper half of feedback loop (R5, R7 and R1) have equivalent resistance of 60 Ohms (R5 parallel R7 plus R1) and they are connected from output to ground i.e. parallel to the speaker. They (together) get the same output voltage as the speaker does. R1 (10R) has one sixth of that "chain" resistance and that's why he'll see one sixth of the output signal - if we have 20V on the speaker it means that there will be 20/6=3,33V over R1. P=(V*V) / R = (3.33 *3.33) / 10 = 1.11 W = power dissipated on R1.
The same goes for lower feedback network (R8, R6 and R2).
Got it?

Juma,

Thank you for taking the time to clarify that for me.
So it doesn't matter if a single 50R or paralleled 100R is used. Can I assume this formula also applies to the reduced feedback model of 20R/100R?

Garrett

[tinitus]

Seems like there could still be good reason to use IR devices

Looking at Toshibas they have gate-source voltage(vgss) +/-20V, which is similar to the IR devices
Fairchild is +/-30V, which also has higher thermal resistance too, if I understand it

Cant find anything on thermal resistance on the Toshiba
But it seems the one available is with "Y" rating, which has the higher Vgs-off voltage
Does this again mean its much less interesting
Am I on right track, or ?

[juma]

Originally Posted by alazira

So it doesn't matter if a single 50R or paralleled 100R is used.

It's just a matter of power rating - Google for: "resistors serial parallel connection"

Originally Posted by alazira

Can I assume this formula also applies to the reduced feedback model of 20R/100R?

The formula is Ohms Law (Google that too). But you can't use 100R/20R just like that. Changing 10R to 20R will change Id of JFETs and other parameters of the circuit consequentially. Visit Erno Borbely's site, there are good articles on "how JFETs work".

[jacco vermeulen]

Originally Posted by tinitus

on thermal resistance

Just compare the different temperature curves in the Vgs-Id graph of the various devices.

[tinitus]

Ok
I tried that, but dont really understand what it is I see on those graphs
Other than FQA19N20C and Toshibas are look alikes, and only the IRFP are different

Anyway, the Toshibas available are as said the Y types with higher OFF voltage
To me, the Toshibas were said to be an option because they should be easier to drive, with the single Jfet
Which to me means they should have lower OFF voltage, but it seems they really doesnt

Besides, we have worked out how to do the cascode, which I suppose makes it easier to drive multiple outputs
Or was the cascode just to make the Jfets take the raised supply voltage better, getting less hot
Or does it work both ways
I think Im a bit confused now

[Melon Head]

The cascode is for the jfet to run at higher power supply voltages

[Melon Head]

Originally Posted by jacco vermeulen

Just compare the different temperature curves in the Vgs-Id graph of the various devices.

Toshiba 2sk1530 looks excellent in this regard

However the testing conditions are different Vds =10V for toshiba and Vds =40V for fairchild.
I have no idea how sensitive the results are to Vds.

[salesmonster]

I've searched for every term I can think of to see if anyone has used heatsinkusa.com/eBay's BarredBoss heatsinks with the F5.

Has anyone used his heatsinks? What size did you end up using for proper cooling? I think my enclosure solution is using a Par-Metal Series 10 enclosure with the side panels being replaced with heatsinks.

Also, since BarredBoss' heatsinks are extrusions, should I cut the heatsinks to mount with the fins vertically? Or should I used them in their natural horizontal state?