Canuck Audio Mart Hifi and Audio ForumCanuck Audio Mart Hifi and Audio Forum
It is currently Sun Dec 17, 2017 5:19 pm

All times are UTC - 8 hours




Post new topic Reply to topic  [ 187 posts ]  Go to page Previous  1 ... 7, 8, 9, 10, 11, 12, 13  Next
Author Message
 Post subject: Re: F5 Turbo
PostPosted: Sun Sep 11, 2016 6:09 am 
Offline

Joined: Wed Sep 14, 2005 11:47 am
Posts: 3838
Location: Ottawa, ON, CA
Frank Wang wrote:
Uunderhill, how is your F5 Turbo going? V2 or V3?
Curious what so called "high band-width" amplifier could be.
I'm using 24V transformers, so V2 it is...

Frank - thanks for asking, but it will be a while before I look at any amps.
I've got a kitchen door that I made and need to install
plus an oven range hood that I haven't put up yet.
Plus million other jobs around the house.

Bandwidth is simply B = | f1 - f2 |

where f1 and f2 are the - 3dB frequencies

The higher the bandwidth of an amp, the more easily it can go into oscillation.


dazed2 wrote:
Odd, both the F5 and F5Turbo V3 I built never had any turn on thump.
And they are differential pairs with feedback loop. And I took out any of the speaker protection that Pass originally designed into it.


The F5 and F6 are symmetrical designs - They don't have a differential pair

With a differential pair, a signal is applied to Q1 and another signal is applied to Q2.
Since they are 180 out of phase, only the difference between the two signals gets amplified (Vout1).
If a differential pair is made from a pair of jFET's its important that
they are matched pairs

The current source (labelled Iss here) added to the differential pair greatly improves
the circuit's ability to reject noise from the power supply rails.
This is expressed as PSRR (power supply rejection ratio).

Image


Erno Borbely's article on the use of jFET's in audio amps is considered a classic.
Look at the circuits, and you can see where NP got many ideas from.
Have a look at page 18 and you'll see the B1 buffer and the front end of the F5.
Its a tricky read, but try to slug through it - I should too.
http://www.linearsystems.com/assets/med ... ordely.pdf


Top
 Profile  
Reply with quote  
 Post subject: Re: F5 Turbo
PostPosted: Sun Sep 11, 2016 5:44 pm 
Offline

Joined: Fri Jul 04, 2014 8:37 pm
Posts: 294
Location: Toronto, ON, CA
Uunderhill wrote:
Frank Wang wrote:
Uunderhill, how is your F5 Turbo going? V2 or V3?
Curious what so called "high band-width" amplifier could be.
I'm using 24V transformers, so V2 it is...

Frank - thanks for asking, but it will be a while before I look at any amps.
I've got a kitchen door that I made and need to install
plus an oven range hood that I haven't put up yet.
Plus million other jobs around the house.

Bandwidth is simply B = | f1 - f2 |

where f1 and f2 are the - 3dB frequencies

The higher the bandwidth of an amp, the more easily it can go into oscillation.
...


Yeah, the possible oscillation...C3,C4 will be needed ..1nf, plus gate resistors use around 680R instead of 47R5 in schematics...

My F5 will also need take a while...crazy busy at work even over time during weekends...

I might post F5 progress from time to time and if you happened saw my post and found something not right, let me know...please


Top
 Profile  
Reply with quote  
 Post subject: Re: F5 Turbo
PostPosted: Mon Sep 12, 2016 8:00 am 
Offline
User avatar

Joined: Wed Feb 06, 2008 2:04 pm
Posts: 663
Location: London, ON, CA
Uunderhill wrote:
..............
Image


Erno Borbely's article on the use of jFET's in audio amps is considered a classic.
Look at the circuits, and you can see where NP got many ideas from.
Have a look at page 18 and you'll see the B1 buffer and the front end of the F5.
Its a tricky read, but try to slug through it - I should too.
http://www.linearsystems.com/assets/med ... ordely.pdf


+1
An interesting reference in so many ways and one that I was never aware of before.
Written in a very unusual way that seems to have targeted an intended audience of DIY audio enthusiasts who were just discovering SS back in 1999.
I suspect that you might have to go back to the 60's RCA transistor manuals to find descriptions of FET operation that use terms such as triode or pentode region :?:

Some similar circuit designs are here in this vintage App Note titled FET Circuit Applications:
http://www.ti.com/lit/an/snoa620/snoa620.pdf

Texas Instruments has since absorbed National Semiconductor and Burr Brown among other early pioneers in using semiconductors in professional and consumer audio :idea:

What the TI AppNote does not mention, as the technology was not mature in 1970, and the Bordely article does not mention almost 30 years later, perhaps to restrict the information to the use of discrete transistors :roll:
is that integration of FET transistor structures into integrated circuits, has resulted in very high performance audio op amps.

National Semi were masters of transistor fabrication.

Fascinating bit of history that at one point National did publish databooks of bipolar and FET die . These did not have 2N or 2SK prefixes.
For small signal transistors these might be 2mm by 2mm with bonding pads for laser wire connections to the outside world.
A clever company could buy and package these into your own transistor case, just as National did, or add it to an opamp die and create your own amplifier module, as National, Burr-Brown, Analog Devices and others did too.

The cost and performance breakthrough was to actually create FETs on the same silicon base as op amps.
Whatever features that need to be matched precisely was done by laser trimming on that silicon base before the package is sealed.
Matched pairs and current sources at inputs then track precisely over temperature.
Much less concern about external noise sources as the "leads" or "wires" between transistors are now only 100uM long :shock:

Amazing progress by so many talented people :D


Top
 Profile  
Reply with quote  
 Post subject: AJ ...temperature
PostPosted: Mon Sep 12, 2016 3:48 pm 
Offline

Joined: Fri Jul 04, 2014 8:37 pm
Posts: 294
Location: Toronto, ON, CA
FYI

My AJ is using hifi2000 5U 500mm full aluminum chassis ...biased around 0.9A
Normally, after turned on 1 hour, AJ will reach it's "norm" temperature...room temperature around 25 Celsius
-Heat sinks temperature around 49 Celsius;
-Inside the chassis also around 44 Celsius (power caps, transformers...everything);
-Parts on amp board over 45 Celsius...

Basically, it's like a oven ...I have to play AJ without top cover;-)



Comparing with F6, 4U 400mm full aluminum chassis...
-Heat sinks around 45Celsius;
-Inside is around 33 Celsius (power caps...everything);
-Amp board of course a bit "hot" as grilled over heatsink (I use 10mm standoff)


Top
 Profile  
Reply with quote  
PostPosted: Wed Sep 14, 2016 9:02 am 
Offline

Joined: Fri Jul 04, 2014 8:37 pm
Posts: 294
Location: Toronto, ON, CA
blue resistors are Vishay MRS25, black ones are Vishay Dale RS02M 1R


Attachments:
IMG_3565.jpg
IMG_3565.jpg [ 202.27 KiB | Viewed 951 times ]
Top
 Profile  
Reply with quote  
PostPosted: Sat Sep 17, 2016 9:39 am 
Offline

Joined: Wed Mar 12, 2014 5:47 am
Posts: 56
Location: Toronto, ON, CA
Fantastic work, Frank.

Oh your Aleph J, I assume you are running 0.9A per fet (and thus 1.8A total bias)?


Top
 Profile  
Reply with quote  
PostPosted: Sat Sep 17, 2016 10:24 am 
Offline

Joined: Fri Jul 04, 2014 8:37 pm
Posts: 294
Location: Toronto, ON, CA
AstroNewbie wrote:
Fantastic work, Frank.

Oh your Aleph J, I assume you are running 0.9A per fet (and thus 1.8A total bias)?


That's bout right. 2 output FETs per channel.


Top
 Profile  
Reply with quote  
 Post subject: JFETs matching...
PostPosted: Sat Sep 17, 2016 9:48 pm 
Offline

Joined: Fri Jul 04, 2014 8:37 pm
Posts: 294
Location: Toronto, ON, CA
Time consuming process...


Attachments:
J74.jpg
J74.jpg [ 190.84 KiB | Viewed 940 times ]
K170.jpg
K170.jpg [ 154.1 KiB | Viewed 940 times ]
Matching.jpg
Matching.jpg [ 159.32 KiB | Viewed 940 times ]
Top
 Profile  
Reply with quote  
PostPosted: Sun Sep 18, 2016 4:01 am 
Offline

Joined: Wed Sep 14, 2005 11:47 am
Posts: 3838
Location: Ottawa, ON, CA
Frank Wang wrote:
AstroNewbie wrote:
Fantastic work, Frank.

Oh your Aleph J, I assume you are running 0.9A per fet (and thus 1.8A total bias)?


That's bout right. 2 output FETs per channel.

Frank,

On my Class A MOSFET amp, I used a total output bias of just 1 Amp.

It drove B&W 110's so sweat in a large room.
The bass control was incredible.


Top
 Profile  
Reply with quote  
PostPosted: Sun Sep 18, 2016 5:52 am 
Offline

Joined: Fri Jul 04, 2014 8:37 pm
Posts: 294
Location: Toronto, ON, CA
Uunderhill wrote:
Frank Wang wrote:
AstroNewbie wrote:
Fantastic work, Frank.

Oh your Aleph J, I assume you are running 0.9A per fet (and thus 1.8A total bias)?


That's bout right. 2 output FETs per channel.

Frank,

On my Class A MOSFET amp, I used a total output bias of just 1 Amp.

It drove B&W 110's so sweat in a large room.
The bass control was incredible.


F6 is also class A, so does F5...and little ACA ;-)
Back to Aleph J, I measured voltage across R18 in schematics when adjusting bias and set to around 400mv(850mA), and R18 is just for one of the output MOSFET Q7, R19 is connect to Q8, total 2 output MOSFETs (Q7,Q8)...Q5,Q6 are for constant current source so they don't count for output MOSFETs.
850mA for Q7, and should also be around 850mA for Q8...add up 1.7A. I don't know if it's appropriate say "total bias" by doing the addition...probably should just say 850mA (per FET).


Attachments:
alephJ.png
alephJ.png [ 45.95 KiB | Viewed 922 times ]
Top
 Profile  
Reply with quote  
PostPosted: Sun Sep 18, 2016 7:13 am 
Offline

Joined: Wed Sep 14, 2005 11:47 am
Posts: 3838
Location: Ottawa, ON, CA
Frank,

Q5 and Q6 help form a constant current source.
They should be matched pairs and current flowing through them should be the same - or very close.

The current flowing through Q5 and Q6 is set by Vbe of Q4 / R16.
Remember that the base - emitter junction of a bipolar transistor is a diode.

Vbe = 850 mA x 0R47
So Q4 is just starting to switch on when Vbe = 0.4 V which is bang on as expected.

As the base - emitter junction of Q4 starts to switch on, the voltage drop across Vgs of Q5 and Q6 is maintained.
This sets Ids of Q5 and Q6.

R16 and R17 could be increased to something like 0.75 ohms
and that would lower the total bias to just over 1 Amp.
However, I think this would change the DC offset,
so other resistors would need to changed too.

.


Top
 Profile  
Reply with quote  
PostPosted: Sun Sep 18, 2016 5:38 pm 
Offline

Joined: Fri Jul 04, 2014 8:37 pm
Posts: 294
Location: Toronto, ON, CA
Uunderhill wrote:
Frank,

Q5 and Q6 help form a constant current source.
They should be matched pairs and current flowing through them should be the same - or very close.

The current flowing through Q5 and Q6 is set by Vbe of Q4 / R16.
Remember that the base - emitter junction of a bipolar transistor is a diode.

Vbe = 850 mA x 0R47
So Q4 is just starting to switch on when Vbe = 0.4 V which is bang on as expected.

As the base - emitter junction of Q4 starts to switch on, the voltage drop across Vgs of Q5 and Q6 is maintained.
This sets Ids of Q5 and Q6.

R16 and R17 could be increased to something like 0.75 ohms
and that would lower the total bias to just over 1 Amp.
However, I think this would change the DC offset,
so other resistors would need to changed too.

.


I used diyaudio boards and followed their build guide...that's all, I'm lazy. You did your own prototype boards, I admire that...even prototype following the exact schematics need a lot of thinking and thorough planning....another thing is that I always find diyaudio's boards premature, no matter which version, some places need be improved...;-)
I will draw PS boards and have factory made for me, something like the diyaudio PS board, but a bit simpler to fit my needs. To me, for class A amps, transformers and power supply is almost the most important...AJ apparently apply more pressure on PS and transformers than F6. The home made PS board is ok, but I need more robust stuff, something “permeant" that I can "glue" caps on...especially for coming F5 Turbo V2 and probably another F5T V3...


Top
 Profile  
Reply with quote  
 Post subject: PCBs...
PostPosted: Sat Oct 01, 2016 7:04 pm 
Offline

Joined: Fri Jul 04, 2014 8:37 pm
Posts: 294
Location: Toronto, ON, CA
Drawing some PCBs...

If don't want use Fuse here, e.g. have fuse on back panel, then short "---J---"...

In diyaudio build guides, use screw terminals, which I always find not "reliable"...I prefer the soldering...no need to worry about screws go loose or parts' legs oxidized...


Attachments:
screw.jpg
screw.jpg [ 33.67 KiB | Viewed 847 times ]
FW_TH.jpg
FW_TH.jpg [ 161.85 KiB | Viewed 854 times ]
Top
 Profile  
Reply with quote  
 Post subject: PCB...finalized
PostPosted: Mon Oct 03, 2016 5:05 pm 
Offline

Joined: Fri Jul 04, 2014 8:37 pm
Posts: 294
Location: Toronto, ON, CA
Some minor changes...
-Remove "J", can short from Fuse(if not intend to use Fuse);
-Add drill holes to fit more types of Fuse holder...AC terminal...


Attachments:
Final.png [8.28 KiB]
Downloaded 5 times
Top
 Profile  
Reply with quote  
 Post subject: PS Board...
PostPosted: Thu Oct 06, 2016 3:53 am 
Offline

Joined: Fri Jul 04, 2014 8:37 pm
Posts: 294
Location: Toronto, ON, CA
PS board...short in the middle to form +/- power supply, otherwise will be 2 set of DCs...


Attachments:
PS.jpg
PS.jpg [ 198.14 KiB | Viewed 800 times ]
Top
 Profile  
Reply with quote  
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 187 posts ]  Go to page Previous  1 ... 7, 8, 9, 10, 11, 12, 13  Next

All times are UTC - 8 hours


Who is online

Users browsing this forum: No registered users and 4 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
Powered by phpBB® Forum Software © phpBB Group