F5 power amplifier

[Stefanoo]

yes, you are right.
That is why i am re-reading all my scrap books of Electronics from the University.
I guess i need to refresh many things as i have abandoned them in these years (not completely....but i am no longer able to do the things i used to do in such a simple manner).
Few days and my recap will be done....yuuupppieee!!
Obviously....even though a simple circuit can be figured out by a couple of calculations...i still think simulations are useful.
An example?
As i was building the output attenuator for the Aleph P, I had at that time the model simulated on my computer.
As i had to decide the value of the resistors for the 24 stepped attenuator i based the calculations on the simplest logarithmic law feeling sure since everybody did that using that excel table.

When i wired up the attenuator and plugged it in i realized that output voltage wasn't what i expected to be.
Why?
Because i didn't take in consideration that since low feedback value of the network is used, output impedance of th circuit isn't constant with the load and this messed up ALL the calculation...and thus the output volume control was a disaster from my view point.

So, i simulated the 24 steps trying the different values available till finding the right values that gave me what i needed.
This time i wired up the attenuator again with the new values ....plugged it in the circuit..and..no surprise on the test bench.

Result? the output volume control is now perfect.
No audible gab when switching position but only a linear audible variation.

Conclusion: if i would have simulated this last part from the beginning i would have saved lots of time and few money since i had to re-buy the resistors.

So never think to be too good for the simulator.
this is my personal thought.
Not even if you are a master degreed engineer who got the highest score...as i think many of you guys are...

i think....unless you have a big practice to be able to see things from far away....i think you might need a neutral confirmation that your ideas are correct....unless you are Nelson Pass...of courseeee....


so after this blameless post....is there anybody that might kindly help me to figure out why is my simulation going wrong?

[Zen Mod]

Originally posted by Stefanoo
.........
so after this blameless post....is there anybody that might kindly help me to figure out why is my simulation going wrong?

I'm not talking/writing anything against sims ;

answer is simple - if you put correct parts values in proofed schematic , then model of active device isn't good enough .

period

[Stefanoo]

that is what i think as well.
I have posted the schematic i am using in case you guys would see some mistake.
If no mistake on the actual draw then it must be something wrong on the models..i guess.

So, may anybody post here the spice model for the jfet and irf parts?


thanks.

[jacco vermeulen]

Originally posted by Stefanoo
what do you mean?

Your JFETs have an apparent resistance, if you put a source resistor behind it you can vary the current flowing into the drain.
But the actual current is determined by the Idss value of the device at hand.
Whether you play The-Sims or Curl-Ball with real devices, Idss is your input, different input is different output.

(NP suggested 8mA-Idss parts several pages back, which leaves over 6mA with the source resistor value in the schematic.)

[Stefanoo]

sorry jacco i don't get your suggestion.
i should add a resistor at the source of each the jfet where there is the two 10ohm resistors that regulated the gain of the second stage?

i am very ignorant as you can see if you could kindly explain me a bit more what you are trying to say i will appreciate it..
more if you or anybody else have a good spice model to share ...is obviously welcome.

[matejS]

Hi,

I am thinking of regulated PSU for F5. Regulated ZenV5 PSU is one way to go, but I want simpler. What about LT1083? It has 7.5A cont. current rating and within Vin-Vout <= 5V, it has cca 10A peak. I guess this would do...

... and I do not plan to put much capacitance after regulator, only one 4.7uF (polyester/polypropene).

What do you think? It's simple and should be good.

Matej

[jacco vermeulen]

Originally posted by Stefanoo
trying to say :

Your JFET models are set up for 12mA Idss, Spice models have NIENTE parameter spread.

[tms0425]

Originally posted by matejS
Hi,

I am thinking of regulated PSU for F5. Regulated ZenV5 PSU is one way to go, but I want simpler. What about LT1083? It has 7.5A cont. current rating and within Vin-Vout <= 5V, it has cca 10A peak. I guess this would do...

... and I do not plan to put much capacitance after regulator, only one 4.7uF (polyester/polypropene).

What do you think? It's simple and should be good.

Matej

In the article, Nelson said unregulated supplies work just fine though you could choose to use Z3 or Z5 as examples of regulated ones. The Z5 itself is apparently very sensitive to supply line noise, but I'm not sure whether the F5 is quite as sensitive (requiring regulation).

I planned to just use two Plitron 300va 2x18vac transformers (per amp) that I used in F4 with full wave bridges and CRC each for plus and minus rails. With the F4 load it yields a little more than +/-24vdc and seems pretty stiff, so I was hoping this would be adequate for F5 (6a continuous, 10a peak per channel) and keep the operating point where it is supposed to be for lowest distortion. Seems like a pretty simple solution unless noise sensitivity is just really high. I will probably do as the natives do and "just try it" except I really hate to take apart my F4 to find out.

Tom

[woody]

That LT1083 looks like a nice reg. but what will you use for
the negative regulator? By the way are there boards avaliable
yet?

[apassgear]

Just to add another iteration on the PCB, the extra extended one, V2.5 for anyone interested