F5 Turbo Builders Thread

That is not my understanding. According to kirchoffs laws the current flowing through the mosfet is shared between the parallel resistors and the diodes. The mur3020 datasheet tells you how much current will flow through the diode pair for any given voltage drop across the diode at any junction temperature.

The issue is that as the diode conducts more current, less current flows across the resistor pair. This means that all current will flow across the diodes once their effective resistance becomes so low as to be negligible, causing one of the mosfets to begin current hogging and explode. You will never see multiple mosfets blow on a single rail because of this.
 
Last edited:
Yes, You can think of it in terms of currents. I tend to view it from a temperature compensation point of view. Once the diodes conduct you loose the compensation provided by Rs.
If one of the diodes begins to conduct during idle conditions that mosfet will heat up and the heat will spread across the heat sink. How fast this happens depends on the thermal conditions. Who knows what will happen when a mosfet shorts, putting dc on the output. With feedback the rest of the circuit will try to correct that dc output voltage. With the huge power supplies built into some of the F5T builds I've seen, there is a lot of energy storage in the supply and also with big transformers and main fuses a lot of power is available from the line until the fuse or something else in the power supply blows.
I have about 1/2 farad of C in my monoblock supplies and 800va transformers. I don't want to see that supply shorted out. I decided against the diodes in my build. Not worth the risk to my amps or speakers.
 
Bear in mind that NP says that the V2/3 really starts to sound better when the diodes are conducting. In my system, with nominal 4 ohm voltage, that is around 1 watt:

P = i^2 * R = 0.5^2*4 = 0.25*4 = 1
V = I*R = 0.5*4 = 2 Volts.

This will drive my speakers to 93 dB, which I find too loud for regular listening.

This implies that the F5T would be a better match with less efficient speakers and/or a much larger room, as the diodes would be conducting at normal listening levels.

I can't follow your calculus concering as to when the diodes start conducting to be honest, the O.5R in your formula is the paralled source resistors ?
 
R is 4 ..the speaker load. I is 0.5 the current needed to drive the speaker to 93 db. Because that is 1 watt per 93 dB at one meter.

NP said the F5T starts flexing its muscles when 2 amps are flowing. Which through my speakers would be at 8 watts output.

But bias current could decrease this ?
 
Clarify:

The F5T starts flexing it's muscles with 2 amps *going to the load*.

This is independent of the bias.

If you want those 2 amps to be Class A amps, then the total bias current must be at minimum 1 amp.

With 4 F5 pairs in parallel, that's 0.25 amps per device, which occurs at 0.3V for junction temperature of 150 deg C.
 
Member
Joined 2018
Paid Member
F5turbo v2

I want to build an f5 turbo v2. I've built an F5 I'd like to recycle some parts.
Can I still use the 35 v bridge rectifiers or do I need to get the Vishay 3020
Diodes. I plan on another transformer with 24v secondaries, getting the store transistors and following 6l6's guide and the diy BOM. I'd appreciate any advice
Thank you
 
F5T V2 - More Gain ?

Hi all,

I have fired up my F5 Turbo with 45V rails feedback resistors are 220R in parallel as per schematic.

I would like some more gain and was wondering if I could increase the feedback resistors some more ? I understand it will decrease feedback at the same time.

Anyone out here has an idea how far to go or not to go ?

Any help is appreciated.

Max
 
Audiosan, I would say so. the Turbo is on the Woofers in an active setup and I don't want to make a preamp for it... it's fed by a Curryman DAc/Output with 2 VRms output only... so I need at least the 4dB (those have been considered OK by Nashbap/Bob Ellis with 380R Feedback resistors in //) - but while at it I'd like to go for 2 x 470R in // for the extra 2db ....unfortunately I don't know in what ratio it acts on the feedback and if/when I'll reach the "not enough feedback"-point as that will have an impact on stability if I am not mistaken...

EDIT: Rails are 45V
 
Official Court Jester
Joined 2003
Paid Member
Using the search function showed some comments on using 380R resistors in parallel for R7 - R10 increasing the gain to 26db instead of the 22db.

I was aiming for +6dB over the V2 turbo (21.6dB) could I use 500R resistors in parallel or is that too much decreasing the feedback ?

Thanks,

Max

just go for it

Papa is always teaching us that fun is greater out of (small ) comfort zone

:)

take usual care of possible oscillations with F5

considering that you're using it for bass duties (and no possible harm in rounding those squares way way high) , maybe is best to right away put lag caps across primary NFB resistors
 
So after agonizing about wire gauge forever, I have decided to wire it totally with 20 awg.

There are some places were a large gauge wire has to go directly to the back side of the boards, in particular, the input stage board.

To my view, this precludes using anything except for 20 awg -- unless you carefully drill out the holes and install solder cups to fit the larger gauge wire. I went down that road before, and do not want to repeat it.
 
I have used 14, 16 and 18 gauge. What I've done is to strip the wire with good quality wire strippers using jaws one size smaller, this removes a few strands enough to fit the conductor in the hole, then when soldering I make sure that those cut strands get close to the soldering pad and get "bonded" to the rest of the strands with solder. It takes time but works well. I used 14 gauge for all runs from power supply to front board and output boards, 18 gauge for inputs, multiple 16 gauge from output boards to speaker binding posts. Amps sound excellent, I have compared them to other amps with good reputations and they come on top.
 
just go for it

Papa is always teaching us that fun is greater out of (small ) comfort zone

:)

take usual care of possible oscillations with F5

considering that you're using it for bass duties (and no possible harm in rounding those squares way way high) , maybe is best to right away put lag caps across primary NFB resistors

ZM, thanks a lot.

I have fitted 240R gate resistors to the outputs and 1nF compensation cap on the output boards (DIIYAudio store boards) - so it's a go!

I'll post a pic later in the Pass Gallery.

Cheers,

Max
 
Here is what I can share: I finished the V3's last week. In 2017 I had built a V2 with 4 cascoded Jfets per front board running at +/- 43 volts, bias at 0.7 A per output devices, total bias 2.8 per channel, using the store's power supply boards it sounded fine but I wanted more power my NHT 3.3's.
This time a used Toshiba Jfets bought on eBay from a seller Nelson recommend the seller based on Sacramento CA (punkydawgs is his seller name), I upgraded the power supplies using Tea-bags dual rails boards running the same voltage. I took a chance buying Panasonic 27,000 uF 50 V capacitors from a guy on eBay (I think he has more of them, $8.99 a pair) they are doing fine so far, no humming on the outputs.
What I have realized is: Toshiba's Jfets are quieter, sound better (I gonna get in trouble by saying this but my ears tell me it's true) they are very low noise= more subtle sounds come through. More class A = sweeter sounds at loud volumes, currently running at .350 V across Rs's so total bias is 5.6 A per monoblocks. I think I'm getting close to 100 watts class A. The upgrades have made a big difference in the sound.