Eh... Probably not quite 10 seconds, but close. I'd like to thank everybody for the advice and help. I'm feeling a bit better about this.
Hy together,
looking forward to build some F5. Read many of the posts, but still have two questions.
1# Why the F5 have no capacitor on the Input (to block DC)?
2# is it necessary to have a Relays on the Speakerout, is there some noise on starttup/shutdown the Amp?
thanks!
looking forward to build some F5. Read many of the posts, but still have two questions.
1# Why the F5 have no capacitor on the Input (to block DC)?
2# is it necessary to have a Relays on the Speakerout, is there some noise on starttup/shutdown the Amp?
thanks!
The input JFETs are self-biasing. There is no DC need for DC on the input.
There's no need at all for an output relay. There are no thumps or pops whatsoever with this amp. Just power it up last.
There's no need at all for an output relay. There are no thumps or pops whatsoever with this amp. Just power it up last.
#1 I suppose you need to be sure that your preamp does not put out any DC.
Pass have some times written that his followers think capacitors is worse than poison 🙂
#2 My F5 has no pops on startup and shutdown.
Br Arthur.
Pass have some times written that his followers think capacitors is worse than poison 🙂
#2 My F5 has no pops on startup and shutdown.
Br Arthur.
Hi!
Please can i replace 0r47 resistors with 0r33 resistors?
What's probabily sonic and technical differences?
Thanks in advance.
Please can i replace 0r47 resistors with 0r33 resistors?
What's probabily sonic and technical differences?
Thanks in advance.
You can do that. The sonic benefits would be pretty minor. You'd get a little bit more power out of it. You run a higher risk of unstable bias though. Just keep a close eye on it.
Thanks Kosst
I would use 0r33 to use mur3020w in parallel as in turbo version but it starts to conduct from 0.42v so i'm afraid that with 1A bias (0r47=0.47v) the diodes are starting to run.
I would use 0r33 to use mur3020w in parallel as in turbo version but it starts to conduct from 0.42v so i'm afraid that with 1A bias (0r47=0.47v) the diodes are starting to run.
I would strongly suggest higher bias. The amp will work well at 1 amp, but really, it isn't sounding it's best. If you look back over the last few pages you'll see the story of my struggles trying to get the bias higher. I'm at about 1.25 amp and that is acceptable, but it runs quite warm. If you don't have a pretty robust heat solution I wouldn't try to hotrod it too much.
Hi! For me not is a problem power dissipation because i have a 3 units (12cm x 40cm x 4cm) hi fi 2000 Italian Cabinet.
I would biasing 1.3A as a other exemplar already built with no problem.
But now i woul to try a turbo mod (in F5 version) if it is possible.
Furthermore i would to try to reduce also feedback net ranging from 50r to 100r (r9-r10-r11-r11 ) for abtain more 2° thd harmonic component.
What do you think about?
I would biasing 1.3A as a other exemplar already built with no problem.
But now i woul to try a turbo mod (in F5 version) if it is possible.
Furthermore i would to try to reduce also feedback net ranging from 50r to 100r (r9-r10-r11-r11 ) for abtain more 2° thd harmonic component.
What do you think about?
I think that the original f5 feedback arrangement sounds better than the lower feedback arrangement. I also think 24v rails sounds better than 32v rails if you don’t need the wattage due to being able to achieve higher bias.
I also would rather have higher bias than have the diodes.
I also would rather have higher bias than have the diodes.
Hi Hikari!
What are your impressions about feedback reduction?
About turbo have you try with and without it?
I have had f5 with 1 pair mosfets and 24v rails, two pairs of mosfets and 24v rails, and the turbo with 2 pairs and 32v rails and less feedback. I have used IRF, Fairchild and the toshiba mosfets.
The answer really is not what sounds best but what your speakers need. If 25w is enough then I think 24v is better, it allows for more bias.
If you don’t need current and have 8 ohm speakers than a single pair of mosfets is best.
For me, with 87db 3 ways I need the wattage so I use a 32v turbo. I do believe the amp sounded better with 24v rails and more bias but I didn’t have enough headroom. The reduced feedback of the turbo version sounds less forward and dynamic.
To me, bias is king. I have experimented with stock bias and up to 2.5a bias. More is better. Same with capacitance. Preamp is very important with the f5 too.
The answer really is not what sounds best but what your speakers need. If 25w is enough then I think 24v is better, it allows for more bias.
If you don’t need current and have 8 ohm speakers than a single pair of mosfets is best.
For me, with 87db 3 ways I need the wattage so I use a 32v turbo. I do believe the amp sounded better with 24v rails and more bias but I didn’t have enough headroom. The reduced feedback of the turbo version sounds less forward and dynamic.
To me, bias is king. I have experimented with stock bias and up to 2.5a bias. More is better. Same with capacitance. Preamp is very important with the f5 too.
Can you tell us about the character of the different fets and how bias/rails influenced them? Thanks.
Well, my testing was less than scientific. For single pairs I have tried the Fairchilds and the IRFs...I remember the fairchilds being more hi-fi sounding and maybe thinner and faster. But I also remember there not being as much difference between them as expected.
For single pairs I have run 1.3-1.8A bias. Most of the bias experiments occured with the IRF mosfets. Bias makes a large difference. Far greater than the mosfets. Everything is better with more bias...it is dramatic. Personally, I think with real world speakers you need at least 1.5A bias to get the F5 really working.
I am using a 3- way speaker with a 3.2R minimum so with easier to drive speakers things may be different.
In search of more bias I went to a dual pair of Toshiba mosfets, still with 24V rails ad 2.5A bias. This was the best sounding setup I think. Highs and lows all great. Great imaging, bass sounded like a big hammer on a soft piece of wood (good) But I couldn't quite get the volume I needed so in went a 24V transformer...
So now I have 32V rails, dual pair Toshiba mosfets, 1.8A bias and reduced feedback. No diodes, no cascode. It satisfies my volume needs but I feel like the lower rail version with more feedback was richer and more dynamic. Probably due to the feedback changes but a lot of it has to do with the reduced bias due to the higher rails. It's kind of a compromise. I am sure with monoblocks and higher bias it would be my end of the road amp. If you need 50 or 100W class A things get big and expensive fast. Also, I have 200,000uf and chokes in the PSU.
I am using the 5U chassis so some of these scenarios may not be possible for those with smaller heatsinks.
For single pairs I have run 1.3-1.8A bias. Most of the bias experiments occured with the IRF mosfets. Bias makes a large difference. Far greater than the mosfets. Everything is better with more bias...it is dramatic. Personally, I think with real world speakers you need at least 1.5A bias to get the F5 really working.
I am using a 3- way speaker with a 3.2R minimum so with easier to drive speakers things may be different.
In search of more bias I went to a dual pair of Toshiba mosfets, still with 24V rails ad 2.5A bias. This was the best sounding setup I think. Highs and lows all great. Great imaging, bass sounded like a big hammer on a soft piece of wood (good) But I couldn't quite get the volume I needed so in went a 24V transformer...
So now I have 32V rails, dual pair Toshiba mosfets, 1.8A bias and reduced feedback. No diodes, no cascode. It satisfies my volume needs but I feel like the lower rail version with more feedback was richer and more dynamic. Probably due to the feedback changes but a lot of it has to do with the reduced bias due to the higher rails. It's kind of a compromise. I am sure with monoblocks and higher bias it would be my end of the road amp. If you need 50 or 100W class A things get big and expensive fast. Also, I have 200,000uf and chokes in the PSU.
I am using the 5U chassis so some of these scenarios may not be possible for those with smaller heatsinks.
Thanks for the detail Hikari. Seems you have really customised this build to suit your tastes, that's DIY done right🙂. I have Aleph 2s, and I can hear the difference when they are warm. The bias increases, and they sound much better, so I agree with you. I would be a bit nervous of running a mosfet that hot though. I think anything over 25W is maybe cutting into their life expectancy. I have not experimented with different mosfets, and I wan to try some Toshibas soon and see how they are, so thanks for the feed back.
Just a thought, as this is DIY, why not build two of them with 24 volt rails. Bi-amp and address your volume issue this way, best of both worlds🙂
Just a thought, as this is DIY, why not build two of them with 24 volt rails. Bi-amp and address your volume issue this way, best of both worlds🙂
Last edited: