allFET circlotron

Founder of XSA-Labs
Joined 2012
Paid Member
Nice work Kees! :cheers:

I just got my J310's, are you saying that I need to use either DN2540 or PN4391? Can J310 still work?

This looks promising though - fast slew rate is good - that's what is reality if you have a fast transient. Maybe you need to add some caps - I notice your earlier board was devoid of caps anywhere.

Good to hear IRF610's act as tempco's at 500mA. I am perfectly fine with operating there. I have 2SK1058's but I prefer sound of IRFP240's - they have higher transconductance so bass impact and slew is faster. I do no not have distortion problems of IRFP240 in my other amps. Sounds fine with 80mA to 150mA bias if driver properly with low impedance drive.
 
Nice work Kees! :cheers:

I just got my J310's, are you saying that I need to use either DN2540 or PN4391? Can J310 still work?

The current feedback eats current, A j310 has to eat that to, that is why I have a 100 mA Jfet used to be save.

But I go try the mirror option so that Jfet can be removed if I can do that without trouble with unbalanced feed.

Anat there is no bootstrap as I see good, the resistors do git the gates of the powermosfets some resistance, so the VAS who is balanced sees a load, otherwise things get nasty unstable, jumping between the two irf9610 .

X I have heart from a music lover that musical instruments have no slew rate, the amp has to be width bandwidth but neets to be deslewed so to say getting a more liquid sound.
That you like the irf more is afcourse a taste of your ears, but I am really look out of what your experiences are, so get to work.

regards
 
Last edited:
Kees

It is nice to see progress !

If you remove cascode part and add current mirrors be sure that K170 Vds is not so high ,
yes I see now why those two 680R(1K) resistors have to remain , in between they add some level of bootstrap to driver stage ,
DC offset is usually very small by circlotron OPS , but check for OPS-IQ stability with those power IRP`s ,
btw ,sorry for my off topic but when you want check this allfet concept schematic based on Infinital A-class amp .

best regards !


Heee he try to copy me;)

joke afcause, it reminds me that I am on the right track.

it looks very much like my other one, however I need no feedback with that one, it is the balanced version we
have here the single ended version to circlotron, making it different.

This one is next when ready with our one. no feedback just local only in cascoded input amp, I like no feedback.
 

Attachments

  • ScreenHunter_478 Oct. 03 16.27.jpg
    ScreenHunter_478 Oct. 03 16.27.jpg
    393.9 KB · Views: 147
  • ScreenHunter_ Oct. 03 16.38.jpg
    ScreenHunter_ Oct. 03 16.38.jpg
    124.8 KB · Views: 217
Last edited:
Founder of XSA-Labs
Joined 2012
Paid Member
572614d1475505514-allfet-circlotron-screenhunter_-oct.-03-16.38.jpg


This is quite different Kees. No more P channel IRF9610. You are back to the J113 which was giving us such problems with not enough current. Are you sure it's enough his time?

You have 4 stages of amplification almost what is used for current driven BJT outputs here. 2sk170/j74 to ZVN to ZVP to IRF then to IRFP outputs. Does t need that many stages?
 
Last edited:
Impedance transforming and source follower need here some amplification, but it has no feedback, source degeneration is done here.

Yes we need this stages to be use of a source follower to drive irfp, the j113 is only in simulation, in real world we need change.

But this amp is for later care in winter maybe, making some pcb and test it.
This amp is balanced, I have even a cascoded one somewhere.

Ahh here it is, it has feedback now because it has high open loop because of cascodes..

Also long term project.

Anad the amp with one supply is only possible with balanced driver, mine is a single ended driver and when need 4 volts from ground to get the circlotron on give just 3 volts swing before hits ground, so I need negative supply 35 volts also.


regards
 

Attachments

  • ScreenHunter_ Oct. 03 17.25.jpg
    ScreenHunter_ Oct. 03 17.25.jpg
    758.5 KB · Views: 142
  • ScreenHunter_ Oct. 03 17.24.jpg
    ScreenHunter_ Oct. 03 17.24.jpg
    113.6 KB · Views: 137
Last edited:
Founder of XSA-Labs
Joined 2012
Paid Member
musical instruments have no slew rate

But a percussive clack of a drumstick on a rim, guitar picking, standup bass, piano, claps, castanets, etc. are close to impulse responses. So closer to infinite slew (without overshoot), the better. Makes a difference when you have 50kHz capable tweeters and time-aligned transient-perfect speaker.
 
Hi X

For the irfp 240 mosfets the joules yoyu need for driver them capacitances, and then special the switch of of the mosfets require in 20Khz already 1,5 amps driver capability, so our amp who has a drain follower needs to be biased miminal on 60 mA each.

Using HEXFETs in High Fidelity Audio

For slew rate to much is not oke, a good square on 20 Khz is way enough, most amps go higher tough, but worth to see how sound is affected.

I go look or to change the circlotron with a extra driver stage wh can deliver some amps can help, like the irf610 in soruce follower mode on the output.

Had somewhere also a compound circlotron design, this do amplifie on itself already 3 times with local feedback.


regards
 
Oke Mister X

The extra driver section in the circlotron who idles in class A did really put down HD.

onder the -80dB and can drive the irfp240 properly with current and is low impdedance. :D

regards
 

Attachments

  • ScreenHunter_486 Oct. 03 22.11.jpg
    ScreenHunter_486 Oct. 03 22.11.jpg
    120.8 KB · Views: 137
  • ScreenHunter_ Oct. 03 22.11.jpg
    ScreenHunter_ Oct. 03 22.11.jpg
    216 KB · Views: 124
Nice work. I assume all sections are class A?

For the output stage, if you set high idle you get class A automatically, for some extent afcouse.

The circlotron section with the extra driver this is clas A setup, the driver in the voltage amp
do not, but maybe a good idea to do also, except the LTP who get trouble with temperature.

Now with the extra stage we have for all stages a nice sinusoidal respons.

But when do this we need a new pcb.

I do not understand how a j170 can drive a hexfet properly in the pass amps seen the high Cn of them..
 

Attachments

  • ScreenHunter_ Oct. 03 22.20.jpg
    ScreenHunter_ Oct. 03 22.20.jpg
    358.7 KB · Views: 113
X

seems to me that those power Hex`s are pretty difficult to be driven and is thermally also unstable ,well at least in this fully balanced amp configuration ,
----------------------------------------------------------
btw, I don`t think that was floating is any sort of peppers but is paprika , any way after so good meal and couple glass of pivo (beer) is good to be in company of nice and talented pevaljka :) (pevaljka- Serbian female folk songs singer )
https://www.youtube.com/watch?v=BpxmbwkQMm0
 
Founder of XSA-Labs
Joined 2012
Paid Member
572685d1475523647-allfet-circlotron-dsc_0328.jpg


Quite certain those red bumps in a ring are the ends of a red bell pepper. The dish name means stuffed peppers doesn't it? Anyhow - I want to try the recipe someday.

I have not had problems driving heFETs in my other amps. I can get nice square waves and low harmonic distortion. Thermal compensation with BD149 Vbe multiplier had worked well too. CFH7 amp, FH9 amp, and FH11 amp.
 
It looks to me is ` punjena paprika` , yum ,yum , :D

btw , why don`t you try same amp but with OPS consisting of multiple pairs of standard K1058 device ? , they are relative easy do be driven in comparison with those power HEX`s .

Yes you are right punjena paprika I did learn to make in Joegoslavian restaurant I did work in 1982.

Now I now where familie of Marina do live, it is Brajkovac, I do now there is if I go out of serbia I get through a big mountain who has a transmitter on top and has a known name I did forget, from there it is going into the mountains 80 km only up and up, the house we build there is close to a Café in that village and after the café trun left we get a river who do in winter is a violent river, and that is as I did remember brajkovac.

I have to go to baco some day, when I have time and such to vistif the grave of Marina and met the mother.

For the amp, X did you read the hexfet article, these mosfets are not recommended for hi end, and inpact for bass, just use a tapped horn. who I have design and go build some day. But now we stick on the hexfet to see what happens.

But something else, I did see with the simulation that the amp swings also negative I do ask is this a disaster for the idle current causing harsing sound.

With the extra driver in circlotron I did see it swings from +1 to +20 and vica versa.

regards
 
Kees

From Belgrade you have to pass mountain of Avala which have that big radio tower on the top , and than following the road of Ibarska Magistrala to found that village Brajkovac which is somewhere in between city of Lazarevac and Krusevac ,
that harsh sound is caused probably from clipping of driver stage combined with OPS crossover switching , scope which can measure output balanced signal will reveal all those signals malfunctions ,
btw , remember this SET+SEPP simple hybrid amp ? , you can made one and use them as rough test rig for direct testing of the sound of any specific power FET , laterall,vertical ,Hex ,N-type,P-type , x,y,z, ....., regardless of any manufacturer specs or specific recommendation as , audio device , audio+switching device , pure switching device ,....
 

Attachments

  • Test Rig.JPG
    Test Rig.JPG
    499.7 KB · Views: 76