F4 power amplifier

Jump to Latest
#3,422
flocchini said:
Hooked up the 4 Jensen JT-123-ALCF transformes wired as Nelson and Choky pointed out in posts 3399 and 3400 to my F4 headphone board.

I lifted R3 and R4 on the version 2 boards currently in the DIYAudio store on the cap side. I connected both leads to the transformer input (grey/brown)
The output was connected to the F4 board where R3 was lifted and the ground was also connected to the F4 board.

The sound is absolutely superb !!!! 😉😉 Have listened to Etta James, Eva Cassidy, Nina Simone, Phoebe Snow, Judy Collins, Doug McCloud, Laura Nyro and Van Morrison. Voices are amazing as is sound stage. Plenty of detail and outrageous bass. No problem driving the my HiFiMan 500s

Thanks Nelson and Choky for the help. Choky I see why you like the M2 !!!
Should I call mine a M2F4headphone amp?





Might have to build a SE version to drive the Ravens in my system.😀

Best

Bob[/QUOTE


I remember reading that output Fet s like a hot heatsink - My heatsinks are obviously overkill - I could try small heatsinks - any feedback appreciated.
Thanks
Click to expand...
 
Last edited:
#3,425
O.K. 😀😀

for all having an F4 only with one pair of IRF....

for instance ZMs crippled headphone version
or for the Magma version with a transformer (Jensen, etc.)

http://www.diyaudio.com/forums/pass-labs/188109-introducing-magma-front-end-ba-2-f4.html

I can highly recommend after some hearing sessions the matched quadrupels of IRF 240/9240 of alweit (Alexandar Weitzmann, Israel), ebay

4 X IRFP240/IRFP9240 (Matched quad) | eBay

I always thought matching of N and P channel is not necessary and I am not sure if this is really the reason for the very good sound....

maybe also possible that the two channels for stereo are close together I do not understand it theoretically but .... it works.

Maybe he even matches you two or three quadrupels for your F4....I did not ask him...😀 because I use a step up transformer 1:5 and only one pair...

This offer is seldom and not very expensive, big recommendation....😀😀😀

maybe the FW without feedback, F4, BA-2 are profiting more from the very close matched quadrupels than an F5 with feedback....
 
#3,427
normally the distributors do not measure themselves they trust in the limits given by the producer.....(I was told this by somebody much more involved with these things)

I do not know how he is doing this miracle...🙂

and my friend with the curve tracer is always enchanted with the close matching alweit is doing and the correctness of the values....!🙂🙂
 
#3,429
generg said:
normally the distributors do not measure themselves they trust in the limits given by the producer.....(I was told this by somebody much more involved with these things)

I do not know how he is doing this miracle...🙂

and my friend with the curve tracer is always enchanted with the close matching alweit is doing and the correctness of the values....!🙂🙂
Click to expand...

Not having matched any of those types, I am unsure. I know with the FQA devices, If matching a large enough sample, it would be possible to get close pulling from the extremes of the two devices. Just basing comment on datasheet and small knowledge i have of limited tru complimentary mosfets. If you have curve tracer and it says it is good, I believe it. It should also make a big difference in response of the amp as well.
 
#3,433
luvdunhill said:
I'd bet if doing a single point match at unspecified parameters that you could find a point on the curve that would have high enough match rates to sell on EBay. I bet at an actual operating point it will differ.

Can you post in circuit measurements?
Click to expand...

Luvdunhill,

Interesting argumentation! The one point argument!🙂

Do you think I should look for the Vgs of N and P channel at the 1,4A bias I have?

Or I could look for the THD value with these types and unmatched p and n type!

Nice idea thank you!😀😀

I did no write this to make advertisement in the favor of alweit, there are some other guys delivering also very good stuff....
I wrote it because the sound got more details and color.....🙂

ZM please do not write that I have golden ears...😀

Hasta la vista!
 
#3,434
:clown:
 

Attachments

  • GF.jpg
    GF.jpg
    121.3 KB · Views: 418
#3,435
I still have other golden parts......😀😀

Alex Weitzman wrote:

"
Dear Gerd,
Thank you for the warm words!
I match at 100 mA (39 Ohm), heating time 1.5 sec (very precise, controlled by microcontroller) , MOSFET pushed to
heatsink, using Fluke 45, synchronized through serial port - I selected this preheating time in order to get repeatable results. I am a software engineer, so it was not a big deal for me to write software to control my station.
Measurements are dependent on ambient temperature, but still MOSFETs stay matched on different temperature - although absolute VGS voltage different.
Best regards,
Alex"

So at least some more information.....At the moment...

The FQA and IRF I measured up to now had very different Vgs values at 200mA.
I never met similar values, mostly the P values were higher if I rember right.
Of course a three point measurement like Euvl suggested for the input J-FETs of xF5 would be better...
 
#3,436
Its basically impossible if they do not have natural similarities. The only ones i know of,...I dont know much, are the Toshibas. THe FQA16n I got was closer to P part in terms of Vgs than 19n part, but this is irrelevant if you look at Yfs. All the sudden, the disparity is obvious.
 
#3,440
luvdunhill said:
Yup just measure Vgs after the amp is warmed up and sounds good. Should not take more than 5 minutes.
Click to expand...

You do not know me.... Warming up of the amp one hour....way from basement with DMM to the salon 5 minutes.....adjusting DMM for the right value 2 minutes...opening the lid of the case 1 minute.... Finding the right place two times in the amp 2 hours.....time for the light from the LCD display to my eyes ....difficult to add, will take some time....hope to give the answer next week...😀😀
 
You must log in or register to reply here.
Top Bottom