Ah 'xrk',
Surprisingly, the TIP41, etc have shown some problems with stability in the past with the working conditions of this amp - the Rod Elliot site (sound.whsites.net/projects.htm) offers different transistor choices - those Darlington type combinations of the active devices have a fairly specific 'behaviour' - there was quite a bit of work done on these circuits here in different threads some years ago, possibly archived now but still available
Also, I'd suggest you read thru the TeddyRegs literature (see PinkFishMedia) for more detailed info and options as these types of circuits are well proven in use - once you get 'the hang' of these circuits, you'll use them all over the place
Also, have a look at the K-Multiplier thread here - pretty simple cct that works really well indeed but it too has quite specific components.
Surprisingly, the TIP41, etc have shown some problems with stability in the past with the working conditions of this amp - the Rod Elliot site (sound.whsites.net/projects.htm) offers different transistor choices - those Darlington type combinations of the active devices have a fairly specific 'behaviour' - there was quite a bit of work done on these circuits here in different threads some years ago, possibly archived now but still available
Also, I'd suggest you read thru the TeddyRegs literature (see PinkFishMedia) for more detailed info and options as these types of circuits are well proven in use - once you get 'the hang' of these circuits, you'll use them all over the place
Also, have a look at the K-Multiplier thread here - pretty simple cct that works really well indeed but it too has quite specific components.
> And what's the upcoming amp?
It is public knowledge that Stixx is beta testing the DAO SE.
The F5X-HA is also working for a while now as posted here before, and there is one last one in the pipe.
We'll publish when it is fully working.
And that woud be about all the headamps I would want to do, except for one last very secret project for electrostatics.
Patrick
It is public knowledge that Stixx is beta testing the DAO SE.
The F5X-HA is also working for a while now as posted here before, and there is one last one in the pipe.
We'll publish when it is fully working.
And that woud be about all the headamps I would want to do, except for one last very secret project for electrostatics.
Patrick
My f5 headamp (point to point) is done. Thank you to EVUL for posting this project (and Nelson of course).
Transformer is 50VA Antek, 2107 case, 317/337 regs (no cap mult), 10K stepped SMD pot. 180ma bias/16V rails. Caps are KG and PW in 317/337 and silmic on the amp boards. Gate resistors are 2K, the amp is similar to the veroboard layout in post 1. MSR860 diodes.
Here's what I think:
Headphones: Grado 125e, AKG 272
Amps compared with: Hifiman EF-2, Macbook out, CD player out
Positives:
This is a fast amp, soundstage is wide and upfront and detailed. More detailed and faster than anything I compared it to. If you like "visualization" of the music...placement, etc. this amp (like the speaker F5) is very, very good.
Vocals very "there" and detailed (but thin, we will get to that later).
Amp stays pretty cool w/180ma bias.
Noise is nonexistent. Background very black.
Loud listening is about 60-70% of the dial with AKGs. Grados maybe 40% of the dial.
Sounds good w/vinyl front end. Some amps (like the hifiman) have too much gain and bring out all the noise.
Negatives:
This is amp has a "bleached" sound to it. Lacks some depth and decay (and bass). It's almost like it needs more contrast if it was a photo or video. It's doing a lot of things right, especially in the space/location/visualization area but it is failing on tone. Going back and forth between the Hifiman amp (or Macbook) and the F5 it's very obvious. There is just a lack of body or sense of drama. Best I can describe it is "bleached". Sounds thin. No midrange body.
I have not put the amp on a scope or anything. I may try my computer-based oscilliscope. But it seems stable and correct.
This is not meant to be a complaint or anything. EVUL has been generous with his time, I respect that. He knows what he is doing. The problem is either MY amp or me.
I have learned stuff, it's fun to make things. Some people just don't like things...even good things.
What I really want to know is "is this the sound signature of this amp"? Do you guys feel the same way?
I have had a F5 which turned into a F5 turbo in my main system for about 4 years. It does not have this sound signature at all.
Is it the PSU? 317/337 letting me down? I know the amount of capacitance in my speaker F5 makes the amp "weightier". Perhaps the cap mult. is needed. Also, more bias in my speaker F5 brings more weight. But I am already running 180ma in the headphone amp and as far as headamps go, that's a lot.
Headphone amps for me come down to tone...sense of space is secondary. Speaker amps are the opposite for me. I don't like the tone of this amp. I want to love this amp, any suggestions?
Transformer is 50VA Antek, 2107 case, 317/337 regs (no cap mult), 10K stepped SMD pot. 180ma bias/16V rails. Caps are KG and PW in 317/337 and silmic on the amp boards. Gate resistors are 2K, the amp is similar to the veroboard layout in post 1. MSR860 diodes.
Here's what I think:
Headphones: Grado 125e, AKG 272
Amps compared with: Hifiman EF-2, Macbook out, CD player out
Positives:
This is a fast amp, soundstage is wide and upfront and detailed. More detailed and faster than anything I compared it to. If you like "visualization" of the music...placement, etc. this amp (like the speaker F5) is very, very good.
Vocals very "there" and detailed (but thin, we will get to that later).
Amp stays pretty cool w/180ma bias.
Noise is nonexistent. Background very black.
Loud listening is about 60-70% of the dial with AKGs. Grados maybe 40% of the dial.
Sounds good w/vinyl front end. Some amps (like the hifiman) have too much gain and bring out all the noise.
Negatives:
This is amp has a "bleached" sound to it. Lacks some depth and decay (and bass). It's almost like it needs more contrast if it was a photo or video. It's doing a lot of things right, especially in the space/location/visualization area but it is failing on tone. Going back and forth between the Hifiman amp (or Macbook) and the F5 it's very obvious. There is just a lack of body or sense of drama. Best I can describe it is "bleached". Sounds thin. No midrange body.
I have not put the amp on a scope or anything. I may try my computer-based oscilliscope. But it seems stable and correct.
This is not meant to be a complaint or anything. EVUL has been generous with his time, I respect that. He knows what he is doing. The problem is either MY amp or me.
I have learned stuff, it's fun to make things. Some people just don't like things...even good things.
What I really want to know is "is this the sound signature of this amp"? Do you guys feel the same way?
I have had a F5 which turned into a F5 turbo in my main system for about 4 years. It does not have this sound signature at all.
Is it the PSU? 317/337 letting me down? I know the amount of capacitance in my speaker F5 makes the amp "weightier". Perhaps the cap mult. is needed. Also, more bias in my speaker F5 brings more weight. But I am already running 180ma in the headphone amp and as far as headamps go, that's a lot.
Headphone amps for me come down to tone...sense of space is secondary. Speaker amps are the opposite for me. I don't like the tone of this amp. I want to love this amp, any suggestions?
Last edited:
> "is this the sound signature of this amp"?
No, but you should wait for a neutral answer from other builders, not from me.
> This is amp has a "bleached" sound to it. Lacks some depth and decay (and bass).
> Sounds thin. No midrange body.
That is not my own experience, though in general a F5 is not known for warm tube sound.
I also have not heard such complaints from the other builders, be it with IRF or Toshiba or Fairchild FETs.
Perhap you should first do a frequency response measurement up to 1MHz, to see if it is working as it should.
I do not think adding a cap multiplier would change things dramatcally.
It will just reduce noise further.
Patrick
No, but you should wait for a neutral answer from other builders, not from me.
> This is amp has a "bleached" sound to it. Lacks some depth and decay (and bass).
> Sounds thin. No midrange body.
That is not my own experience, though in general a F5 is not known for warm tube sound.
I also have not heard such complaints from the other builders, be it with IRF or Toshiba or Fairchild FETs.
Perhap you should first do a frequency response measurement up to 1MHz, to see if it is working as it should.
I do not think adding a cap multiplier would change things dramatcally.
It will just reduce noise further.
Patrick
Last edited:
You are right, I should check the response of the amp and I will before I make a final conclusion.
I feel it might be a reaction to the bandwidth of this amp...it feels so wide, it may be that it is hard to focus on an area (like the midrange) because it seems so flat and wide. Everything is there, nothing sticks out. Some amps with midrange magic are just magical because there is nothing else there. It may be my brain...that's why I am asking if others feel the same way.
But regardless, I will make sure it is functioning correctly. I also don't think it's a PSU issue but I was hoping the cap mult would bring the magic out. Funny thing is the amp sounds MUCH better with vinyl...maybe limiting the bandwidth or the vinyl "EQ" is more to my liking.
I just talked to one of the builders over email. He will post his experience.
So I leave him to do that, freely, without any influence from me.
But he thinks it is the power supply.
(I have not tried all variations as some other guys did.)
If you are used to a certain signature, you will hear the difference immediately, and think the new one has a deficit.
I used to listen to a Krell year ago, and when I listened to the JLH for the first time, I thought it was flat and lacking sparkle.
I then listened to the JLH for a week, and then turn the Krell back on to compare.
The Krell was sold within a month afterwards.
Patrick
So I leave him to do that, freely, without any influence from me.
But he thinks it is the power supply.
(I have not tried all variations as some other guys did.)
If you are used to a certain signature, you will hear the difference immediately, and think the new one has a deficit.
I used to listen to a Krell year ago, and when I listened to the JLH for the first time, I thought it was flat and lacking sparkle.
I then listened to the JLH for a week, and then turn the Krell back on to compare.
The Krell was sold within a month afterwards.
Patrick
Last edited:
I built 2 versions of the F5-HA, first is the standard version using the official PCBs and power supplies. Sounded excellent to my ears.
Second is with my own PCB (picture attached). While building that second version and tested it with a SMPS, it didn't sound very good, then I filtered the power supply with a big RC filter, sound improved a lot. Finally I used the PS from post #137 - all the magic from the official version came back.
I think the power supply used have a strong influence to the sound of this amp.
Second is with my own PCB (picture attached). While building that second version and tested it with a SMPS, it didn't sound very good, then I filtered the power supply with a big RC filter, sound improved a lot. Finally I used the PS from post #137 - all the magic from the official version came back.
I think the power supply used have a strong influence to the sound of this amp.
Attachments
Ahhh...another satisfied customer...I mean builder.
1) How can the mids be "very" there, yet thin??
2) Sounds about right.
3) YOUR amp and mine makes a grand total of two.
I know I didn't do anything wrong, plus I used only quality components to construct mine.
Hikari, it's just not good to doubt yourself.
4) Yeap.
5) Yes, build something else and put it in that nice case.
Really?
I have several shunt supplies and a couple of LM-based supplies that I use.
With all the discrete and op-amp circuits I've built and tested, I could detect virtually NO difference in sound quality.
Maybe this F5 needs a "better" supply to sound its best.
The F5 needs a supply with sufficiently wide bandwidth to ensure stability, not necessarily one with better spec on paper.
That is why we chose an open loop cap multiplier.
The Salas shunt was e.g. a disaster. As Stixx found out.
And I would bet the Didden regulator would have the same problem in this application.
In my very first prototype, the amp on Vero board even caused my linear lab supply to trip at the mains fuse.
Patrick
That is why we chose an open loop cap multiplier.
The Salas shunt was e.g. a disaster. As Stixx found out.
And I would bet the Didden regulator would have the same problem in this application.
In my very first prototype, the amp on Vero board even caused my linear lab supply to trip at the mains fuse.
Patrick
Last edited:
If You want to try something different regarding PS,
You can add "Push-Pull/Master-Slave" stage after your 313-337 stage,
based on RNMarsh very old idea.
Look at http://www.diyaudio.com/forums/power-supplies/277235-317-337-based-virtual-ground-ps.html for schematics and origin.
After this stage, PS ground is hard fixed exactly between rails.
Cap multiplier is solution for residual HF PS noise as a last PS stage, if needed.
Wish You happy experimenting
You can add "Push-Pull/Master-Slave" stage after your 313-337 stage,
based on RNMarsh very old idea.
Look at http://www.diyaudio.com/forums/power-supplies/277235-317-337-based-virtual-ground-ps.html for schematics and origin.
After this stage, PS ground is hard fixed exactly between rails.
Cap multiplier is solution for residual HF PS noise as a last PS stage, if needed.
Wish You happy experimenting
The circuit from Ric is a virtual Gnd splitter.
It does not solve the problem of limited bandwidth on any regulators interacting with the F5.
The cap multiplier not only reduces HF noise.
More importantly, it provides a flat Zout from 10kHz onwards to >> 5MHz.
That is also the case for the Nazar regulator.
Patrick
It does not solve the problem of limited bandwidth on any regulators interacting with the F5.
The cap multiplier not only reduces HF noise.
More importantly, it provides a flat Zout from 10kHz onwards to >> 5MHz.
That is also the case for the Nazar regulator.
Patrick
Not my experience at all. My build is on vero board as well (Patrick's proto), and I am using a simple cap multiplier power supply (or as alternative the Nazar reg posted somewhere in the thread).
It could well be that the quality of the power supply has an influence on the sound since my F5 HA devellopped the "magic" only with the cap multiplier, but not with the 7815/7915 supply used for initial testing.
The high bandwidth of the amplifier contributes to speed and detail but doesn't make the midrange sound thin. One of the key qualities of the F5 HA (and the most important to me) is that it sounds so natural... warm but not wooley or bloated. Fast but not aggressive... and so on. Definitely not thin in the midrange.
Last edited:
I haven't tried a cap multiplier before. I'm working on a layout for the multiplier supply already mentioned here.