F5 Turbo Builders Thread

regarding this

Primary target of the PS capacitors is to increase S/N of the amp.
Secondary one is enabling the amp to reach the most continuous power at a target impedance, ripple lowers the usable range of the rail voltage.
Third is to stock ample energy for short bursts, when the transformer is not delivering or can't deliver enough juice.

No 1 is a reason why Class A amps need a larger cap reservoir : higher bias level means higher ripple level at zero output current.
No 2 is important for power amps which are required to deliver a lot of power in 4 Ohm and lower impedances.

For one and two, total uF size of the capacitors is the main theme.
For the third category, total energy of the capacitor bank is the important parameter.

Those two are not the same.
Think a balanced/bridged amp, compared to a regular one that delivers the same power.
For the same bias level, each half of the bridged amp requires the same uF number as the straight amp does, for identical ripple level.
Ideally, rail voltages of the bridged amp are however half the number of the ones in the normal power amplifier.
Total energy of the caps is proportional to squared voltage level, so a bridged power amp would require 4 times the cap size of the basic amp, at half the voltage rating (ideally).

CRC is a good one to drop ripple, more efficient than C+C.
For peak output level, CRC is less favorable.
Prettiest for a high bias turbo is a choke PS. For a reference, see e.g. Mr Gerhard Cerubin monster amp at the MPP thread.
For peak current delivery, also think a large bank of motor run caps (reference : PBN Olympia from Denmark)

For a quicky estimate of PS cap size, pick a ripple number and desired impedance : C number rolls out of mains frequency, impedance, output voltage and associated current.
 
Tanks Jacco

I could not have put this on paper any better (but you all know that)

Now for the Mechanical side of things ...

First one AC V
2 and 3 DC offset I have seen beter than that once is warm
No cheating (I newer do) meter give min and max V and if you look range is same for pictures taken same time apart.

This is what you get with 150mU on CRCLC
 

Attachments

  • F5 turbo 013.jpg
    F5 turbo 013.jpg
    854.3 KB · Views: 6,093
  • F5 turbo 005.jpg
    F5 turbo 005.jpg
    986.8 KB · Views: 5,640
  • F5 turbo 010.jpg
    F5 turbo 010.jpg
    800 KB · Views: 5,322
Jacco realy great to see you here

Could you please give me (and those interested) a run down regarding the use of mosfets with Ciss of 2550pF (P chanell)Gfs 19 S and 1250pF (N channell) Gfs 20 S
compared to 920pF (P channell) Gfs 6.6 S and 1200pF (N canell) Gfs14.5 S?

Plan is to stick 4 of those in parralell
Transconductance looks close for top pair beter than second?
Many Tanks
 
Last edited:
About dinamic range
1 whole minute of the New Cold Play CD at enoug volume to fill bass on tummy.

Range = pup (crapp with explosive duble p at end would not do)

It's always interesting, and enlightening, to see what the actual power requirements are for music. The problem is that when you hit a big peak, and it comes within a frequency range where your speakers are at their impedance minimum or high phase angle, you want as much power available as possible. I think that this F5 Turbo v3 at 100W Class A/8ohms, 200W Class AB/4ohms is going to be perfect for the majority of people out there!
 
Hi all! :)

I have some 2SK1530/2SJ201 pairs here and while they aren't exactly matched, their Vgs are within 100mV range. Most of them has higher value though. Is it ok to use them to build F5 Turbo?

And also, I couldn't understand how thermal runaway is prevented on this design. Can someone please explain me?

Many thanks
Daniel
 
Last edited:
Hi all! :)

I have some 2SK1530/2SJ201 pairs here and while they aren't exactly matched, their Vgs are within 100mV range. Most of them has higher value though. Is it ok to use them to build F5 Turbo?

And also, I couldn't understand how thermal runaway is prevented on this design. Can someone please explain me?

Many thanks
Daniel

Nevermind, I've just noticed the thermistors preventing thermal runaway. :p
 
mosfets with Ciss

For a power amp to reach a desired bandwidth, voltage is required to rise fast enough, which leads to rise-time, the time it takes to do 90% Vpp.
Combined with the peak output voltage of the amp leads to the slew rate figure, volts per second.
To get there, the input capacitance of the output device gate has to be (dis)charged fast enough.

Capacitance is amps time seconds, per volt : C = A *s / V
Tossed around : C * V/s = A
Slew rate is volts per (micro)second, V/s.
So : required drive current of the gates = input capacitance times slew rate.

Again : C = A *s/V => C = A/V * s
Transconductance : A/V
Means that for a given output stage, charging half the input capacitance leads to half the output current.

Regular output stage is a source follower (aka common drain), signal in at gate, signal out at source.
Ciss = cgs + Cgd

If you double the number of output devices in a source follower output stage, the drive current may charge the capacitance half only.
But as the output devices number is doubled, output current will be the same: 1/2 times 2 is 1.

So for a single device, the input capacitance to drive is Cgs + Cgd.
For two devices : Cgs + Cgd + Cgd

If you look at the F5, it's signal in at the gate, signal out at the drain.
Implies that it's a different drive than a source follower.
Or as papa says : 800KHz with 2 pairs, 80KHz with 20 pairs (85Khz, due to 1 time Vgd)

Different also is that the N-channel JFET handles both the charge and discharge of the P-channel output device(s), same thing on the other side.
(upside is that the bias current of the input stage is used twice, the driver in a regular output stage has to charge/discharge the gate of both P- and N-channel)
Which is why each half has it's own steering system.

Recall C = A/V * s and gm = A/V
Means that the weakest half sets the pace.
In this case, it's the FQA12P20, 544 pF/A/V
(4 times Vgs plus 1 time Vgd, divided by a transconductance of 6.6 A/V)

Me bit distracted by UKtoecutter's board :clown:
 
Last edited:
Tanks Jacco
I will have to digest that.

In the mean time have a pint on me.

for now I am going to crak open a botle of apple wine is only 19 proff but not sometink one can get legaly

So duble the drive currants I going to think about 2 Jfets in parralel bit.

By the way If you ever this side of the water Just give us a call I am going to keep a botle of Nardini for that
 
regarding this

Part Deuce (bigelow) :

For cap size, also read the Pass A40 paper, the WCR part.
W is 2 *Pi * mains frequency, so 2*Pi*f is ~314 for 50Hz and 8 Ohm.
(for full wave rectification : ripple current = ripple voltage * 2WC)

Other post :
The FQA36P15/FQA28N15 look the better combo for Ciss & transconductance.
They're also good for 1/4 more bias current (the lower power one sets the pace for dissipation)
Not a garantee that they'll sound better than the other ones.
(thanks to Patrick's mini group buy, i still have plenty of matched FQA19N/12P left at bottom rate, so easy for me to say)
 
Last edited:
More power than what F5 can provide would be welcome for my speakers. Having F5 in hand, I’m looking for an easy way out, which -per my very limited understanding- seems available. I’m considering the following.

In the F5 Turbo text:
Increasing the power supply voltage is the obvious way to get more power out of an F5. You can simply raise the supply rails to +/-32 Volts and get 50 watts into 8 ohms right away without other modification. 24 V AC secondaries on the power transformer will do it. Don't forget to use higher voltage power supply capacitors. Probably you should also upgrade R9 through R12 to 5 watt resistors. Depending on your heat sinking, you will probably want to adjust the bias so that the power transistors don't run too hot.
I have Peter Daniel boards and 36V power supply caps. Am I ok, if just I swap the transformer to 2x24V 500VA (to replace 2x18V 400VA) and above named resistors to 5W? Or is there some downside I should be aware of, except that I need to adjust the bias current with the trimpots per heat?

Text also states:
You can remove the current limiting (Q5, Q6, R15 to R18 of the original) as long as you are cautious about shorting the output.
Am I to get what benefit out of this? I do understand that it brings along a risk, but I’d assume there’s something positive about it as well. Or is there?

Further:
You may also notice that the feedback resistors R7 through R10 have been increased, increasing the amplifier's gain to about 22 dB and decreasing the amount of feedback by about 7 dB. If we are going to put out more power it is appropriate to have some more gain, and it gives us more margin for feedback stability – the amplifier still is flat to about 800 Khz.
In F5 these resistors are 100ohm and in Turbo V1 220ohm. Should I change these as well, if I’m to follow the simplified path described above? To 220ohm? Or is this relevant only if I’m to use double output devices (i.e. per V1 turbo)?

Comments welcome. ;)
 
More power than what F5 can provide would be welcome for my speakers. Having F5 in hand, I’m looking for an easy way out, which -per my very limited understanding- seems available. I’m considering the following.

In the F5 Turbo text:
Increasing the power supply voltage is the obvious way to get more power out of an F5. You can simply raise the supply rails to +/-32 Volts and get 50 watts into 8 ohms right away without other modification. 24 V AC secondaries on the power transformer will do it. Don't forget to use higher voltage power supply capacitors. Probably you should also upgrade R9 through R12 to 5 watt resistors. Depending on your heat sinking, you will probably want to adjust the bias so that the power transistors don't run too hot.
I have Peter Daniel boards and 36V power supply caps. Am I ok, if just I swap the transformer to 2x24V 500VA (to replace 2x18V 400VA) and above named resistors to 5W? Or is there some downside I should be aware of, except that I need to adjust the bias current with the trimpots per heat?

Text also states:
You can remove the current limiting (Q5, Q6, R15 to R18 of the original) as long as you are cautious about shorting the output.
Am I to get what benefit out of this? I do understand that it brings along a risk, but I’d assume there’s something positive about it as well. Or is there?

Further:
You may also notice that the feedback resistors R7 through R10 have been increased, increasing the amplifier's gain to about 22 dB and decreasing the amount of feedback by about 7 dB. If we are going to put out more power it is appropriate to have some more gain, and it gives us more margin for feedback stability – the amplifier still is flat to about 800 Khz.
In F5 these resistors are 100ohm and in Turbo V1 220ohm. Should I change these as well, if I’m to follow the simplified path described above? To 220ohm? Or is this relevant only if I’m to use double output devices (i.e. per V1 turbo)?

Comments welcome. ;)

There is nothing wrong with replacing the transformer, and the required parts, but if you do that and don't want new boards, you could wire another pair of output mosfets to share the heat dissipation.

Regarding the speaker protection circuitry, some people report that amplifiers sound better without this parts. You can try it yourself, if you hear a difference leave them out, if not, put them back in.
 
power supply

Hi,

Concerning the F5Turbo power supply:

(1) I was wondering if there is some advantage in the center-tap style wiring of the secondaries vs the original F5 power supply that uses bridge diodes (other than needing four fewer diodes and one less thermistor on the primaries)?

(2) The paper says "The power transformer needs to be VA rated at least twice the actual dissipation of the amplifier." Can someone please give me an example of this calculation using V1 or V2?

(3) Would using a bleed resistor like in the original have a deleterious effect?
 

Attachments

  • F5 Turbo Power Supply.jpg
    F5 Turbo Power Supply.jpg
    206.2 KB · Views: 4,596
  • F5 Power Supply.jpg
    F5 Power Supply.jpg
    131.5 KB · Views: 3,737