F5 power amplifier

[CanAm Man]

Nice piece of explanation, Andrew T. Nicely done.....

 

[AndrewT]

Note that the arithmetic is very easy.
Only a couple of formulae to remember. I=V/R and for parallel resistors is R1 * R2 / {R1+R2}

 

[CanAm Man]

Nice part of it, Andrew, was how you "backed into" the logic of the schematic. I made a paper copy for my files.....!

 

[Russellc]

Let's do some example numbers.

For Vgs = 4V @ 1.3A output bias.
Vdrop across the 0r47 resistor is 0V611
Total Vdrop across the jFET collector load is 4V611
Assume that the jFET is drawing 8mA = 0.008A
The resistance required to generate 4V611 @ 0.008 A is 4.611/0.008 = 576r375
This is made up (cold) with the 2k2 // 4k7kthermistor+2k2 // Vr = 1668r13//Vr
Set Vr to 880r66 This requires the Vr to be moved from 0r0 to ~17.6% of travel.

Now remove the thermistor. The Vr needs to be set to 780r98 or ~ 15.6% of travel. Note just 2% of travel difference.

Now back to the removed thermistor holes. You can insert a fixed trimming resistor here.
You need 576r when set cold. set the amp up using no thermistor. Switch off. Discharge all the caps. Measure the resistance across the group of jFet collector load resistors, checking both the top and the bottom loads.
Let's guess you measure 630r for top and 510r for bottom. Yes they could be very different because the mosFET Vgs can be very different.
Use the existing 2k2 by jumpering across the thermistor location. You now have 2k2//2k2//Vr in top and bottom.
Set top Vr to 1k474, set bottom Vr to 1k211. Note these new Vr values are much higher than previously, needing ~30% and 24% travel to get the 1.3A bias.

If you use a lower fixed resistor in place of the 2k2 you will find that the Vr needs to be moved even further making bias setting a little less sensitive to small movements.

Although these numbers are pure guesswork they are an indication of the sort of Vr values you might need.
Let's look at Vr current. When set to 0r0 the Vr wiper current is 8mA. This gives a maximum dissipation across the whole Vr when set to maximum and passing 8mA of ~ 320mW. This is quite high for a 400mW Vr. A higher Vgs and/or a higher Idss could take the Vr beyond maximum current capability when set to 0r0.
The Vr current when set to 780r is ~5.9mA. The equivalent maximum dissipation is 174mW or ~43% of maximum dissipation. This should be OK long term.

Now look at Vr dissipations when the low value extra parallel resistor is put in place. Vr is set to 1k2 or 1k4.
I'll use the 1k2 since this is a more demanding situation.
the 5k Vr set to 1k2 will pass ~ 3.84mA and the equivalent dissipation is 73mW or about ~18% of maximum rating. Well OK for long term use.

Now that was well put together!!!

Russellc

 

[Zen Mod]

fact that Andrew must write War & Peace means just one thing ....... someone somewhere somehow didn't read Papa's F5 manual - that crucial one more time

 

[CanAm Man]

ZM...... yeah, but still nicely done. And I can't wait until the movie version comes out!

 

[jpedro]

Only if your mains supply is 230V, what's the transformer secondary voltage, how many windings, and what is the VA rating?

You can save yourself a lot of headaches if you download and study the F5 service manual from the First Watt web site, that should answer 99% of your questions.

Hi Itsmee,
Thank you.
The transformer is this:

Best regards,
jpedro

 

[AndrewT]

A fact for those that want it.
I have read the F5 PDF and I have not yet started building the F5.

 

[Zen Mod]

fact is that you read it enough ;

building is irrelevant

 

[Woody1911a1]

A fact for those that want it.

appreciated , thank you

the knowledge you all have didn't come overnite . i know because mine is in woodworking and mechanics over many years .

my kick back time is spent here reading and reading and reading and chit it's been four months and i don't know everything ....... YET

cheers Woody

 

[bluevas]

Thank you guys for the explanation about the thermistor TH1 and TH2. I just read post # 2450. Peter Daniel and EUVL did this. how about increased the feedback resistor with 100 ohm.

 

[juma]

.... jFET collector load ...

That would be a heavy burden, carried by someone who has a great collection of JFETs...

 

[jacco vermeulen]

Well, somebody has to do it.

 

[juma]

Well, somebody has to do it.

You should know, no doubt, Mr JFET collector...

 

[AndrewT]

The jFET collector load is set to 0r0 for first power up and initial testing. Once the circuit is proved to be operating correctly, then the collector load resistance is increased to bias the output stages.
No harm comes to the jFETs with a zero collector load. That is part of the design.

 

[EUVL]

He was referring to the word "Collector" which applies to bipolar transistors and not JFETs (drain), straightly speaking.

But everyone understands what you are saying I guess.


Patrick

 

[NYCOne]

Cascoding outputs of the F5

Background:

I built the XF5 amp and cascoded the inputs. I used the Toshiba 2SK1530/2SJ201 outputs. I bought a bunch of Toshiba 2SK3497/2SJ618 to try as outputs. They're very cheap, and have nice Forward Transfer Admittance, but they come with a high Ciss.

I'm going to try them as a drop in replacement for the 2SK1530/2SJ201, but it also strikes me that cascoding may help with the Miller effect. I looked for information about cascoded outputs on the F5, but didn't find much. I'm an absolute newbie in the amp world, I'd take advice on cascoding the 2SK3497/2SJ618 - I'd like to try it.

 

[merlin2069er]

fuse?

I'm using 5x20mm 2.5A fuses (slow blow) for my F5. I just popped one with my transformer and bridges hooked up into the power supply unit.

I tested each primary side of the transformer to the outputs of each bridge ok.

I might have a short somewhere in the power supply I guess.

 

[Nelson Pass]

That would be a heavy burden, carried by someone who has a great collection of JFETs...

Who summons the JFET genie?

 

[juma]

I guess everybody does (once in a while, at least) - and he answers, most of the time...