Using 150mm wires from circuit board to FETS?

[Tony D.]

The enclosure I have designed is such that I have to run 100 to 150mm wires from the circuit board to the FETS. on the heatsinks.

I notice that not many constructors use this technique. Aside from the fact that it does not look as neat is there a problem with doing this?

I thought some of the FET. charectoristics might be changed but could find nothing in the Archive or on the NET.


Thanks for any informaiton.



Tony D.

 

[GRollins]

It's not so much that the MOSFET characteristics will change--it's more a question of whether you'll develop problems with oscillation. If you must run leads that long, you might find that it helps to put the gate resistors at the MOSFET end of the wire rather than the other end.
That said, I've got the prototype Aleph-X here next to my elbow with...you guessed it...6" (that's almost exactly 150mm) leads going to the output MOSFETs. But then, I seem to lead a charmed life as regards to oscillation.
Don't try this at home, folks, he's a trained professional. (Ah, but professional <i>what</i> is the question. Best not to go there.)
I'm also willing to take more chances than the average DIY critter. A fancy way of saying,"Do as I say, not as I do."

Grey

 

[Peter Daniel]

Aleph 0 had about that long wires from input board to FETs. The gate resistors were on ea. device.

 

[UrSv]

Check the original A75 article for this. Pass seems to use quite long leads in his case without problems.

Check Part 2 - Page 9 of the A75 article at http://www.passdiy.com/

/UrSv

 

[Samuel Jayaraj]

All of the Pass Labs designs seem intrinsically more stable. That is why people are removing the lead/lag compensation caps in the Aleph-X and are still a very happy lot.

If you were to try using long lengths of wires or wire-wound Source resistors in say some of the Elektor circuits, you sure will have an oscillator, not an audio amplifier.

IMO, avoiding wires to connect output devices is best and if due care is taken in the design of the PCB, construction is that much easier. Aren't most of us looking for short signal paths?

 

[jh6you]

If you must run leads that long, you might find that it helps to put the gate resistors at the MOSFET end of the wire rather than the other end.

The gate resistors at the MOSFET end might be the best. But, the resistors at the other end could be also sufficient because the resistors could be located at any place between the unwanted parastic positive feedback input and the gate.

If wrong, please let me know.

 

[Tony D.]

Thanks everyone for the replies,

I had a look over at PassDIY.com and I found this:

http://www.passdiy.com/images/gallery/a4-p1-1.jpg

and I see Mark used fairly long wires and did indeed connect the Gate resistors directly to the FETS so that is the way I will go.


Tony D.

 

[Nelson Pass]

I have not seen any difference as to which end of
the wire the Gate resistor is best, but if you have any
concern about capacitive coupling of the leads, then at
the transistor is best.

I don't have any information as to what lead lengths you
can actually get away with, but I always cut it short as
mechanically possible. 15 cm should not be any kind of
a problem. 15 m will probably take wing and fly away.

 

[jh6you]

15 m will probably take wing and fly away.

I will try the 15m. Thanks.

 

[mrgreg2u]

hi guys. i am building an aleph 1.2 and was wondering does all the leads have to be the same lenght? would it cause some sort of difference in cap. or bias?

 

[nar]

Yep. All wires same length

 

[Nelson Pass]

I am inclined to say all leads as short as possible.

 

[P.Lacombe]

Specially in class AB amplifiers, it's of extreme importance to keep source wires with same length. Otherwise, some inequality occurs in on/off transition time, causing a special kind of unsymmetrical crossover distortion at high frequencies (and also cross conduction, which can be destructive in very high power amplifiers)

Also, to avoid high frequency oscillations, it's necessary to connect the gates resistors directly on the transistors, as close as possible. High frequency oscillations can be very underhand (perfidious ?) if no oscilloscope is available for final tuning.

Regards, Pierre Lacombe.