An illustrated guide to building an F5

hehe welcome, you have some reading to do if you ask this question. not the case, although i'm almost certain the answer is no, but i dont think even the obscenely large cases Papa has been playing with lately, will house and cool 5 x f5 with integrated power supply for all.

are you aware of exactly how much heat 5 class A power amps will put out at idle?. you will need minimum 1.5-2kva of transformers, 300000uf capacitance and also, just to make sure you understand the nature of the f5, in 5.1 it will sound amazing, but a standard f5 is generally 25-30wpc, while dissipating considerably more than that again into the heatsink. not really what people usually think of for a home cinema.

now if you are cool with that and you are ready for the custom case needed, have the skills and are prepared for many many hours of planning before you even order your first part, then great power to you and dont forget pictures.
 

6L6

Moderator
Joined 2010
Paid Member
But, judging by your comments, clearly this would not be a place to start my DIY hobby.

Actually, it's a great place to start the hobby... the F5 is a great amp, easy to build, simple to diagnose, and sounds fantastic. Many people have cut their teeth on this amp, with great results!

It is a truly great project, which is why it's so popular.

But if your requirements are for multi-channel only, there might be a different project out there for you. For example, a 5-channel chipamp might be more appropriate.
 
I have to agree with 6L6 on this one.
Fan idea preaty good mosfets sound better at hi bias.
(At least the Tosh I have tried)

As I am planning on dubling the boards on a single heath sink I got a simple controll made up with parts you may have already if building the F5 that may came handy.

I have tried CPU fans and even the noisy ones are silent at 1/2 off V nominal

Just set pot for fan to start tiking over at wanted C the hotter the faster.
and it only start up when you want to so no long time waiting for temperature/ offset to stabilise.


I am Not shure about the next bit could some one cek the math please?
On the other hand you may decide that your normal listening level is quite low
say for example 6 V
6V / 8 (?) homs speaker impedence = 0.75 A The amplifier will stay in class A up to 4.5 W.

I have 4 homs speakers and 32V rails at the moment I have dubled the number of mosfets and current bias is set at 3A

I have decided on a listening level of 12 V and therefore 12/4 = 3
3X12 = 36 W class A

It will pump out much more than this as it leave class A.

One thing I have noticed sound is better at hier voltage.
Don't get me vrong F5 is already realy good but at 32 V rails is a bit more open a bit faster attack.

So fans are worth looking in to and you could use 2 in series and skip on the LM317 bit of the drawing
 

Attachments

  • Testing%20The%20dudas%20003.jpg
    Testing%20The%20dudas%20003.jpg
    593.5 KB · Views: 2,329
Quite right
And not at the moment

But would you realy think that she let me keep the mule in the livingroom for long especialy in our litle UK cubiholes.

Not all can afford mannors so

Option 1 separate suply and 2 small towers

2 as option 1 but balanced (pulling 3 A at 32 V X 4 quadrants = 384 W dissipation )

Option 3 as option 1 but By amp 2 towers 1 for left and right speaker but 2 F5 in each
384 dissipation

Option 4 as 0ption 3 but 2 balanced F5 = 700 W dissipation

now we are talking cpu fan X3 in series tiking away at 70 %

This based on 12 V class A requirment on 4 Homs Puke speakers so 3 A bias.

Option 5 6 and 7 no worth mentioning and off topic

At times I do sound to put it mildly whakooo but is just my brain running away and asking the next question even before the first one was answered.

By the way is it wort the effort?

Yes absolutley the mule sound fantastic just cant get enough of it
Even started buiing the same copy of records I already have becose the old ones are a bit worn and then having to look for the same all over again because the original Shanakie sound a milion time better than the CBS pressing.

And this with a questionable MC pre and a SS1 Akay duda.
 
I've been able to get a hold of 3 DVMs, and I've got the one side @ .59v, the other side @ .58v and dc offset @ .4v.

I'll leave it connected for about an hour or so, but I can't enclose the unit yet until I get the other heatsink mounted.

The heatsink has been heating up a bit, not overly hot.

Should I check these results again with one meter? I'm not so sure how well each are calibrated?

Hopefully, I'll be able to drill & tap, and mount the mosfets, and heatsink today.
 

6L6

Moderator
Joined 2010
Paid Member
Just move the meters to the different test locations, and if Meter 1 reads very close to meter 2 in the same position, and meter3, than you are good. If one reads very different, that one may be suspect.

Keep the amp on for a half-hour or so minimum (to get the sink hot) and re-adjust. Is anything significantly different?

.4v offset is not good. .04v (point zero-four) is on the high side of acceptable. What is the highest bias you can get with zero offset? (It's going to be lower than your .59v)
 
ok, thanks. I'll move the meters around. I could have push the thermistors closer to the mosfets. (I'll do that. )

I'll try lowering the bias to get near 0v dc @ the dc offset.

I did leave the amp on for a good 1.5hrs , and the bias (left & right) didn't vary much.

I'll attempt it again soon.
 
Member
Joined 2009
Paid Member
I think you would be better just using two meters one on the output and the other across one of the source resistors and aim for a reading near to 0.6v then tweak the trimmers to achieve the lowest offset whilst staying close to the bias setting. This is what Zenmod recommended and it worked for me; like you I was trying to get 0.59v across both source resistors and found that the offset was all over the place. If you get it set across one resistor and a low offset reading the other resistor setting will be very close but maybe not exactly the same.
 

6L6

Moderator
Joined 2010
Paid Member
ok, thanks. I'll move the meters around. I could have push the thermistors closer to the mosfets. (I'll do that. )
...I did leave the amp on for a good 1.5hrs , and the bias (left & right) didn't vary much.

If the thermistors are touching or almost touching, that's good enough -- they are working properly because you said that there wasn't much difference in the bias readings between cool and warmed-up

I'll try lowering the bias to get near 0v dc @ the dc offset.

Yes. This is very important right now, to see if you have enough R in series with the pot to get them to bias. I have a suspicion that you will need to change out the resistors (R3, R4) like I did...
 

6L6

Moderator
Joined 2010
Paid Member
I think you would be better just using two meters one on the output and the other across one of the source resistors and aim for a reading near to 0.6v then tweak the trimmers to achieve the lowest offset whilst staying close to the bias setting.

I can personally attest to the benefit of the third meter.

As you are turning the pots, it starts to look like one of the pots adjusts the bias, and the other trims the offset. But both need to be turned up in similar amounts as you get closer to full bias, and then you trim (slowly) for the exact setting.


This is what Zenmod recommended and it worked for me; like you I was trying to get 0.59v across both source resistors and found that the offset was all over the place. If you get it set across one resistor and a low offset reading the other resistor setting will be very close but maybe not exactly the same.

And you are describing the same thing I am getting at, which becomes clear with 3 meters -- one pot pulls the bias up, the other adjusts the offset (and usually pushes the bias down a bit as you zero the offset, requiring a couple of turns on the other one to get the bias back up, which then requires a turn on the opposite pot to remove the offset... ad infinitum... until you get both very close, and you are down to 1/4 or 1/2 turns to get everything perfect.)
 
Member
Joined 2009
Paid Member
I guess your right. At the time I only had two meters so just followed Zen and cvillers advice and adjusted each trimmer alternately by a half turn until things started registering and then tweaked until I arrived at the correct readings. As you say its best to set to a slightly lower bias and leave things to warm up before making final settings.
 
I was able to get everything mounted this weekend. yippee!

I tested my right channel (the channel I just installed)

after 5min -> .532v & .535v - dc offset -> .000
after 45min -> .568v & .565v - dc offset -> .016
after 60min -> .569v & .567v - dc offset -> .016

my left channel (previously connected)

I moved the meters after 60mins

after 60min -> -.575v & .565v - dc offset -> -.502

(this was with top cover and front plate on)

So, it looks like the right channel needs a little pot tweaking to get to ~ .0vdc.

But my left is way out of wack. I'll attempt to look at the board closely to determine if I've mixed up any resistors, hopefully I can check these values with a meter.

Any suggestions?

Thanks