hi all,
i would like to build a "basic" V2 (32V rails, 4 outputs/channel) but i have a few questions before i commit. if i missed these being answered elsewhere, apologies. i have tried to read quite a few thread pages on the Turbo.
1) if i do the 0.3V "scaredy-cat" bias Nelson suggests in the write-up, what kind of Class A power am i looking at?
2) related to that question is, how much heatsinking will i need? i would appreciate any suggestions, particularly regarding domestically available stuff. in my basic F5, i'm running 6" lengths of the 10" stuff from Heatsink USA, and it seems to be enough, but i'm unsure how dissipation will change with the V2. at first glance, if the Class A power from 0.3V bias is similar to the basic F5, then i'd probably want at least a little extra to account for the extra MOSFETs and diodes.
3) i've done some poking around, and it appears the F5c board can be used without too much trouble for the simpler Turbos, as Nelson mentions. would that be the case for a V2 like i'm planning? except for P3, which i don't mind leaving out, i suppose the only trickiest part will be fitting in the diodes. can they be mounted with extension leads?
i should note i am flexible with bias -- 20-25Wpc in Class A is enough, frankly. i'm mostly looking for the extra headroom i get from having 50Wpc on tap because my speakers are probably not quite sensitive enough for my F5 as it is (sealed 3-way bookshelf, exact sensitivity unknown), and also a bit difficult to drive in the bass -- i usually only use a few watts, but i don't want to be clipping the big transients. plus, the V2 sounds more intriguing than just building the higher-wattage F5c.
thanks for the help!
-Chris
i would like to build a "basic" V2 (32V rails, 4 outputs/channel) but i have a few questions before i commit. if i missed these being answered elsewhere, apologies. i have tried to read quite a few thread pages on the Turbo.
1) if i do the 0.3V "scaredy-cat" bias Nelson suggests in the write-up, what kind of Class A power am i looking at?
2) related to that question is, how much heatsinking will i need? i would appreciate any suggestions, particularly regarding domestically available stuff. in my basic F5, i'm running 6" lengths of the 10" stuff from Heatsink USA, and it seems to be enough, but i'm unsure how dissipation will change with the V2. at first glance, if the Class A power from 0.3V bias is similar to the basic F5, then i'd probably want at least a little extra to account for the extra MOSFETs and diodes.
3) i've done some poking around, and it appears the F5c board can be used without too much trouble for the simpler Turbos, as Nelson mentions. would that be the case for a V2 like i'm planning? except for P3, which i don't mind leaving out, i suppose the only trickiest part will be fitting in the diodes. can they be mounted with extension leads?
i should note i am flexible with bias -- 20-25Wpc in Class A is enough, frankly. i'm mostly looking for the extra headroom i get from having 50Wpc on tap because my speakers are probably not quite sensitive enough for my F5 as it is (sealed 3-way bookshelf, exact sensitivity unknown), and also a bit difficult to drive in the bass -- i usually only use a few watts, but i don't want to be clipping the big transients. plus, the V2 sounds more intriguing than just building the higher-wattage F5c.
thanks for the help!
-Chris
Last edited:
between + and gnd
rule of thumb for series resistor , when common type LED is used - use 100R for each volt in PSU ; example - use 2K2 for 22V
500mW resistor is usually fine
if too bright , just double value
I made complete 'ritmetic for LED , but now I'm lazy to find it , or type again
As for heatsink requirements, look at the original F5 manual for a good hint --
(From the F5 manual) (Bold is mine)
"At 1.3 amps per channel, you will see idle heat dissipation of 62 watts. To keep the tempature rise of the heatsink to 20deg C. above the ambient temperature, you will want a heat sink rated at about .6 deg C./watt for each transistor. An example of this would be a chunk of finned aluminum, with a series of 2" fins attached to a 8"by 6" base. You will need 2 per channel."
So - There it is. A 8x6 inch finned heatsink for each transistor. That doesn't count the heatsink requirements of the bias diodes. Obviously a forced cooling system can be smaller.
Knowing this, the answer to the question that you are thinking is, Yes, the amp is going to be big.
(From the F5 manual) (Bold is mine)
"At 1.3 amps per channel, you will see idle heat dissipation of 62 watts. To keep the tempature rise of the heatsink to 20deg C. above the ambient temperature, you will want a heat sink rated at about .6 deg C./watt for each transistor. An example of this would be a chunk of finned aluminum, with a series of 2" fins attached to a 8"by 6" base. You will need 2 per channel."
So - There it is. A 8x6 inch finned heatsink for each transistor. That doesn't count the heatsink requirements of the bias diodes. Obviously a forced cooling system can be smaller.
Knowing this, the answer to the question that you are thinking is, Yes, the amp is going to be big.
As for heatsink requirements, look at the original F5 manual for a good hint --
(From the F5 manual) (Bold is mine)
"At 1.3 amps per channel, you will see idle heat dissipation of 62 watts. To keep the tempature rise of the heatsink to 20deg C. above the ambient temperature, you will want a heat sink rated at about .6 deg C./watt for each transistor. An example of this would be a chunk of finned aluminum, with a series of 2" fins attached to a 8"by 6" base. You will need 2 per channel."
So - There it is. A 8x6 inch finned heatsink for each transistor. That doesn't count the heatsink requirements of the bias diodes. Obviously a forced cooling system can be smaller.
Knowing this, the answer to the question that you are thinking is, Yes, the amp is going to be big.
yes, you read my mind
that 8x6 recommendation is assuming full Class A bias, right? i guess then my first question is the next one after that, which i might be able to answer myself. if i ignore the bias diodes (as i'm not sure what their effect on the biasing process is), 0.3V across 0.5R of source resistance should give 0.6A per transistor, or 2.4A per channel. that's nearly double the original F5 per above, so i would guess that translates to roughly 40Wpc Class A. another question, then. does the V2 need to be biased at least to 0.3V to take advantage of the diodes? as i said above, i don't need quite that much Class A, particularly because it makes the enclosure design that much more complicated (i.e. large).
thanks, that last part especially helps clarify my requirement. but don't you have to multiply the 0.6A times the four devices per channel in the V2 schematic? or can you safely run 32V rails with only a pair of devices if you limit bias to 1.2A per channel? my speaker load is 4 Ohms nominal with a dip to between 2.5 Ohms and 3 Ohms in the bass range, so i don't want to make life too hard for the outputs.
If you ignore the bias diodes you are kinda sorta not really building an F5turbo...
Nelson repeatedly says that for best sound you want to run as much bias as possible. Running more bias helps with the transients, and that seems to be your concern. Perhaps a speaker that's more efficient as well?
The additional output devices of the turbo instead of the 'classic' F5 will help with the current required of the lowish-impedance speakers. Don't worry about it.
Nelson repeatedly says that for best sound you want to run as much bias as possible. Running more bias helps with the transients, and that seems to be your concern. Perhaps a speaker that's more efficient as well?
The additional output devices of the turbo instead of the 'classic' F5 will help with the current required of the lowish-impedance speakers. Don't worry about it.
orignal F5 has 1.3A pr device
this is a push-pull
ah yes, i get what you're trying to say now, sorry.
If you ignore the bias diodes you are kinda sorta not really building an F5turbo...
i only meant ignoring their effect on the bias calculation. i guess if they don't conduct until above the bias point, then they don't come into the calculation.
Nelson repeatedly says that for best sound you want to run as much bias as possible. Running more bias helps with the transients, and that seems to be your concern. Perhaps a speaker that's more efficient as well?
a more efficient speaker may come along, but right now i like what i've got. and i'm sure Nelson is right about higher bias in general. at the same time, i'm still trying to balance trade-offs here. but i can see now that the V2 with a dissipation target of somewhat better than 0.31C/W per channel (as outlined by Audiosan) is quite doable.
If you ignore the bias diodes you are kinda sorta not really building an F5turbo...
Hi 6L6,
I'm not sure I'd agree. I think the diode thing has become larger than life because of this forum.
What I get from reading the article is Nelson offering the widest variety of choices for hopping up the F5 - different power levels, different topologies, and different costs to the widest group of people. Lots of options, lots of choices, right to the of the article.
http://www.diyaudio.com/forums/pass-labs/27831-high-speed-diode-220-package-aleph-x.html#post321830
Paps hasn't mentioned yet what they're using currently.
(MUR3020W is a likely candidate)
Paps hasn't mentioned yet what they're using currently.
(MUR3020W is a likely candidate)
I have come to believe that Nelson is a pretty calculated and thoughtful guy and that nothing he does is without purpose. I think the purpse of the F5T was spelled out in the article and in the same article Nelson states the improvement he believes comes from the addition of the diodes. No doubt it has been blown out of proportion, but then again, how many people even fully understand what is going on with the diodes. I think i do, but I am not 100% sure. EIther way, it is a clever trick for more transient power and that is a prcatically useful thing for most modern multiway speakers. You also have the ability to expand an idea beyond a presented concept. Consider all th past info he has released. If clever enough, you could have assmbled something very close to his commercial offerings. Of course they would not be the same and he has held back some things, but I believe if you ask folks who know him personally, they could tell you stuff that would blow your mind about his approach and intuitive planning in all that he does. i have heard rumours tha even his filing system is a bit revelatory. Folklore? Legend? Who knows. Here we are talking about
Mr. Pass wrote in his article on power supplies in 2001 that he likes slow rectifiers bypassed by small capacitors. Has he changed his mind since then? Anybody know what he uses in his commercial offerings?
Based on the photos on the 6moons review it looks to me like First Watt selected PC grade Electrolytic caps for the PSU in the F5 however in the Pass Labs amps that shade of blue looks like high quality and very large Vishay caps for the PSU. No doubt someone is going to correct my assumptions ;-)
Based on the photos on the 6moons review it looks to me like First Watt selected PC grade Electrolytic caps for the PSU in the F5 however in the Pass Labs amps that shade of blue looks like high quality and very large Vishay caps for the PSU. No doubt someone is going to correct my assumptions ;-)
Sir Buzzforb, I am sorry if I left doubt as to which rectifiers I was asking about. I was speaking strictly about the power supply. The little bypass caps are to attenuate the switching noise I think. More capacitance in the banks is always a good thing. I am building the F5T V3 exactly as drawn, account he's the engineer. I am only the lowly Conductor.
I find it very noteworthy that the 6moons review places the F5 right up in the top tier knowing that all the parts are very ordinary, albeit quality components, that came straight out of the bin at places like Digikey and Mouser etc.
There isn't a single example of anything like unobtainium based this or cryogenic that or oxygen-free single crystal whatever. And the rectifier bridges and main PS filter caps referred to above really aren't anything special.
But the F5 stands there; ranked with other best of breed tube and SS amps.
Very cool.
Kind of cuts through the you-know-what and gives an ordinary body hope.
There isn't a single example of anything like unobtainium based this or cryogenic that or oxygen-free single crystal whatever. And the rectifier bridges and main PS filter caps referred to above really aren't anything special.
But the F5 stands there; ranked with other best of breed tube and SS amps.
Very cool.
Kind of cuts through the you-know-what and gives an ordinary body hope.
Last edited: