F5 Turbo Builders Thread

Dear All,

Is there anyone here who used recently the JFETs Toshiba 2SJ74 + 2SK170 LAB MATCHED QUAD to 0.03mA AND 4mV (7-8ma range) from Punkydawgs instead of the LSJ74 / LSK170 from diyAudio store to build the F-5 Turbo v2 cascode version with 2SC4793+2SA1837?

Thank You

Pavel



Yes I am finishing my F5T V2 with these jfets and for having measured them they are really excellent well matched
 
Perhaps a stupid question:

Is it better to locate the ground loop breaking diode bridge near the Power Supply or near the signal grounds/speaker return?

In my amp, there are two ground loop isolators... one per channel, as this is dual mono construction. They can be located either near the back of the amp, along with the front end boards, or at the front of the amp, along with the power supply cap bank. If near the cap bank, the signal grounds are long. if near the back of the amp, the diode bridges are near the front end boards, and the speaker return signal will cause them to turn on/off.

THe question is in regard noise. Will the speaker return signal cuase any significant issues if the ground loop isolator is near the front end boards?
 
Moderator
Joined 2002
Paid Member
The ground loop 'breaker' creates a (relatively) high-impedance path between the amplifier circuitry and safety earth (and is illegal in many countries). It is also primarily a resistor, and not a bridge. The resistive element defines its impedance, the capacitor its impedance at high frequencies and the diode bridge limits the voltage across the resistor so it will continue to function at high leakage currents without blowing up the resistor.

Given the rest of your circuitry is properly wired up, it makes no difference because its impedance will be higher than any other path the return currents could take.

It is obviously possible to muck this up (I've seen people using both this and a resistor between the signal and power ground which is both redundant and dangerous) but if your have some basic understanding of how the currents flow, it shouldn't really matter.
 
Member
Joined 2013
Paid Member
I received the heat sinks for my v3 balanced amp, they are 14" tall and 10" wide with a 10mm base and 2 1/2" fins.
I plan on mounting the boards stacked, should the P channel or N channel go on top or bottom?
 

Attachments

  • IMG_0722.JPG
    IMG_0722.JPG
    802.1 KB · Views: 376
  • IMG_0723.JPG
    IMG_0723.JPG
    658 KB · Views: 374
Other than for some easy to remember location trick, I don't see that it matters. I put P on top, because positive is above negative.

However, I think you should put the mosfets to the middle of the sink and diodes above/below. There is a post in this thread in the last 3 months or so about distance between the mosfets and orientation -- has to do with heat dissipation, not polarity.
 
Moderator
Joined 2002
Paid Member
If there is a small capacitor on the input shield to case, is the capacitor on the ground loop breaker necessary?

Assuming correct wiring the capacitor will already be in parallel to the loop breaker.

Ripple current may be an issue depending on the capacitor size - be sure it can handle at least an ampere or two of current at frequencies >40KHz.
 
Moderator
Joined 2002
Paid Member
You're going to want a bigger capacitor in the breaker - something like X2 capacitor of 100nF.

Switching supplies (not necessarily in the amp) can put lots of high amplitude, high frequency noise on the line that can be unpleasant to touch, and these will find the resistor too inductive. In the case of a line fault these will safely pass through the cap.
 
I would kindly like your feedback on my problem:
I've purchased lsk170/lsj74 grade B matched jfets from dyaudiostore and mounted them on my F5 Turbo V1 pcb.
See image.

as you can see the jfets can be mounted and unmounted for test (red circle).


- I mount the first pair of lsk170/j74 matched.
- Power ON
- I expet 0 V bias (blu circled resistors 0.47 5W) and closed to 0 V at the output.
- Results are: 0 V bias (OK) and 0.012 V at the output (OK).
I can complete the biasing procedure SUCCESFULLY.

- I unmount the first pair of lsk170/j74 matched
- I mount the second pair of lsk170/j74 matched.
- Power ON
- I expet 0 V bias (blu circled resistors) and closed to 0 V at the output.
- Results are: 0 V bias (OK) and 0.150 V at the output (NOT OK Too high).
I can’t complete the biasing procedure SUCCESFULLY.


If I replace only the second lsk170 (suspect) with the first lsk170, I’ve:
- I expet 0 V bias (blu circled resistors) and closed to 0 V at the output.
- Results are: 0 V bias (OK) and 0.016 V at the output (OK)
I can complete the biasing procedure SUCCESFULLY but obviously they are not matched .


Besides I've an genuine 2sk170BL Toshiba. If I mount this in place of lsk170 suspect,



- Results are: 0 V bias (OK) and 0.010 V at the output (OK)
I can complete the biasing procedure SUCCESFULLY but obviously they are not matched .
What do you think?



So I think that the second lsk170 it’s the responsable.


Biasing procedure it’s obviously according to Pass (God bless him) procedure in the relative article.
I’m expert on this amplification, I’ve already built one F5 with Toshiba’s Jfets.
See my F5 (normal Not turbo).

Thank you for support.
 
Last edited by a moderator:
Hello

Do you connect the Gnd of power to the ground or do you lower the floating power supply?
Cordialy

Not sure I understand.

I use a center tapped transformer. The center tap is connected to the zero volt line ( power supply common ). The zero volt line is the amplifier side of the ground breaker circuit. The other side is connected to the case/safety earth.

Does that help?
 
Administrator
Joined 2007
Paid Member
:cop:

I've had to remove the links from Dr Freemans post as they ran to hundreds of thousands of characters long. This is why images wouldn't load on this page.

This is a very very tiny snippet... it took Notepad several seconds to load this even on a fast PC with SSD.
 

Attachments

  • Annotation 2019-06-25 203051.jpg
    Annotation 2019-06-25 203051.jpg
    724.2 KB · Views: 399