F5 Turbo Circuit Boards

[buzzforb]

I guess your right. My feelings are that few will operate at those power levels. Even if you need it, the increased temperature willcontribute to heatsink temperature, still causing the change you suggest. More than likely, the changes will be instantaneous vs constant. I will try one channel each way and track temperature. Consider two amps through every diode in V3. You have only 1W of dissipation per diode,but 16A of available current to the load. Perhaps some will need more power, but it seems hard to imagine. This is a benefit of the increased output pairs IMO. More intial bias heat, but not as much fluctuation during playback as with lesser quantities. As always i could be viewing this in incorrect terms

[AndrewT]

Originally Posted by Bksabath

........ try solder an extra resistor on top of existing one that will give you 0.33 ohms........

Excellent solution.
Even a fourth might work. But the disadvantage of source resistor set to 0r25 is that there may not be sufficient thermal stability for the output device.

[buzzforb]

Originally Posted by AndrewT

Excellent solution.
Even a fourth might work. But the disadvantage of source resistor set to 0r25 is that there may not be sufficient thermal stability for the output device.

Closely matched devices will help in this regard, correct?

[AndrewT]

No.
Pass chose 0r6 or thereby.
The Vds and Rs are closely linked for thermal stability.

Going lower on Vds will allow lower Rs.
Euvl's F5x used this, +-16Vdc supplies and Rs= 0r22

But the F5T is intended generally for higher voltage and this may require a higher Rs than 0r22.

[buzzforb]

Are you saying that Rs value is related to rail voltage. Could you explain further? I thought it was mainly for stability against current hogging and linearization.

[jacco vermeulen]

Originally Posted by Rush

Am I missing something?

What you are missing is that the Turd-O is a 1-side sliding bias amp.

The static bias value only delivers part of the Class A output current, by the time there's 0.6V across the source resistors (and next to 0V on the other side), current through the diode could be at 1A.
For a 2-pair output stage, that would amount to 4.4A total, >50% over your 50W mark.

To estimate what the diode current would be at 0.3Vf, for other than datasheet temperatures, extrapolate the If/Vf graph curves (in the order of 10mA at 50-55C).
An accurate measurement value will be equal or lower, not higher (S-curve).

[AndrewT]

I still do not believe that adding the bypass diodes while keeping the bias voltage at 590mV will not compromise the performance of the amplifier.

I still believe the diodes should only pass when extra current is demanded. This can only happen if the bias voltage is reduced to substantially below the usual 590mV.

If one requires 1.3A of bias (of a standard F5 output stage) and expects the bypass diode to pass near zero current then the bias voltage needs to be ~ 400mV. This would require the source resistor to be reduced to ~0r3

I do not know if the F5 would work well with Rs=0r3.

That's why I put out a call for advice on what to do with the F5T biasing.

I am still waiting. Where have all the experts gone?

[buzzforb]

When i pull mine together today, i will only use one pair so that i can push it enough to ensure that the diode is conducting. I willtry to put together some measurements as to when it starts conducting as well as at what associated temperature.

[eyoung]

I am watching as i have wondered about adding the diode bypass in single pair f5...

regards,E

[Tea-Bag]

Done it already. The trick is getting your signal high enough for them to conduct.