Bobo,
They're not heatsunk.
It just happens that the devices Papa specified are TO220 package and the only room I had for them was at the top of the board.
Parralel imputs
There you have it.
AS far as I know paralel imputs give a bit more current and maybe same gain.
Imput impedance is given by R1/R2 (as far as I know as well )
They need to be close togheter for termal coupling / Traking
easy done just move close togheter or just put another set of pads close to existing ones.
I like the board wery much by fitting the Mosfets / diodes paralel to the board plane they can be bolted to the heat sink with the diodes on the botom Mosfets on top
Heath rise
what is the question about mono blocks?
board fit to one sink make monobloks with just one sink
There you have it.
AS far as I know paralel imputs give a bit more current and maybe same gain.
Imput impedance is given by R1/R2 (as far as I know as well )
They need to be close togheter for termal coupling / Traking
easy done just move close togheter or just put another set of pads close to existing ones.
I like the board wery much by fitting the Mosfets / diodes paralel to the board plane they can be bolted to the heat sink with the diodes on the botom Mosfets on top
Heath rise
what is the question about mono blocks?
board fit to one sink make monobloks with just one sink
Attachments
Last edited:
Yes, I was thinking more about what would happen if the devices were badly matched. i.e. Would one pair conduct early and render its 'partner' inoperable'. It just doesn't 'feel' right to me. But hey I'm open to be convinced.
Last edited:
That is the difficoult part about building with mosfets in parralel
Papa pass published a paper about matching
I did same work by matching at given temperature.(VGS changes with temperature)
Papa had big advantagge of good batches from suplier
I got my 100 FQA12P20 and FQP19N20 from farnell and batch was all over the place.
You can adjust for a lazy mosfet by adjusting the source resistors As you got 2 in parralell it may be easy to keep a few low values ones just for purpose
Papa uses a 2.4 Homs in parralel in a picture posted here somewhere by migty Zen let me go and get link to Papa article
http://www.firstwatt.com/pdf/art_mos_test.pdf
On the sketch is what I meant by mounting board to parralel to sink
Papa pass published a paper about matching
I did same work by matching at given temperature.(VGS changes with temperature)
Papa had big advantagge of good batches from suplier
I got my 100 FQA12P20 and FQP19N20 from farnell and batch was all over the place.
You can adjust for a lazy mosfet by adjusting the source resistors As you got 2 in parralell it may be easy to keep a few low values ones just for purpose
Papa uses a 2.4 Homs in parralel in a picture posted here somewhere by migty Zen let me go and get link to Papa article
http://www.firstwatt.com/pdf/art_mos_test.pdf
On the sketch is what I meant by mounting board to parralel to sink
Last edited:
Yep. There certainly is. I know about that one. So, how about we:
1. Move the F5 Turbo stuff from this thread to the F5 Turbo is posted thread, or....
2. Start the F5 Turbo BUILD thread. With folks already discussing build components, and build PCBs......a thread for BUILDing would be nice. Months from now this will be a thread with a lot of inputs/variations/comments/etc.
How about it guys?
CanAM Man
IMO start new tread
we can leave theoretical discussion on that one.
UKToecutter
Matching mosfet paper also good for imputs no source resistors there but easy enoug to get god matched Jfets from Blues (got my lot from him) or other members.
Jfets need matching for current (Ids Ithink)
Got to go and catch same Spatcha
Later Alligators
IMO start new tread
we can leave theoretical discussion on that one.
UKToecutter
Matching mosfet paper also good for imputs no source resistors there but easy enoug to get god matched Jfets from Blues (got my lot from him) or other members.
Jfets need matching for current (Ids Ithink)
Got to go and catch same Spatcha
Later Alligators
Last edited:
In the F5 Turbo article, Nelson stated that with V1 that "We are not in a good position to ask the original heatsinks to dissipate more, so if you increase the supply to 32 volts you may need to operate at a lower bias point"
And that's only to go beyond V1. The First Watt heatsinks appear larger than most everything I've seen for DIY.
If the output devices are doubled you would need a massive heatsink to have enough dissipation. Doesn't seem practical.
Last edited:
I've been running the 3497/681 in a balanced cascoded amp for about a year now...
F-5 Turbo Circuit Boards moved here:
http://www.diyaudio.com/forums/pass-labs/206756-f5-turbo-circuit-boards.html
http://www.diyaudio.com/forums/pass-labs/206756-f5-turbo-circuit-boards.html
Quote:
Originally Posted by labjr
........Doesn't seem practical.
hehehehe, when has that stopped anyone around here?
Well you still have options.
1)Find out how much power you realy need (See Pano tread)
2)find out impedance of speakers and how much current you realy need
F5 current out = 2X bias current to stay in class A above that it will go in to class B
(this is as far as I understand things)
you can have for example 2 mosfets at 1.5 amps or 3 at 1 A
I think that altrought the sink dissipate same wattage the Junction themperature on the mosfet be lower. Let the sink go to 55 C
(Onother thing I have noticed is that after a 3 4 hour realy loud session sink felt some what cooler than idle Maybe insted of dissipating the Bias current as heat they where producing blifsfull sounds and running colder so Play loud )
Consider that Papa already said "they sound better at hi voltagge and Bias"
Papa also said Get heat sinks as big as you can aford.
Run at less than 1.3 A mosfet and you can hear it.
No point to go low voltagge your speakers will not ask the amplifier for uge current .
Say 16 V out on 8 homs speakers no mather what you will never ask more than 2 A
with 2 mosfets at 1.3 A on each rail you got 2,6 X2 A avvaliable
Yeah speakers are not resistors and impedence drop so you got a bit of head room
Any point going for 4 mosfets and 10 Amp current output.
So what options you got
P&P may be the killer but if you like same 300 X 400 x83 heat sinks like maine just PM
OR you could use a fan litle CPU coolers are capable to dissipate 100W of heat
Bolt one on the sink and run at low voltagge al you need is a spare mosfet a spare thermistor and a pot to make a themperature speed controlled Fan.
Even at very low speed difference is uge
Use wires to place mosfets on 2 sinks just make shure gate stoppers are soldered to gate pins duble cek you got no hi frequency oscillation or go balanced to double the power
OR like me wait for F6
Originally Posted by labjr
........Doesn't seem practical.
hehehehe, when has that stopped anyone around here?
Well you still have options.
1)Find out how much power you realy need (See Pano tread)
2)find out impedance of speakers and how much current you realy need
F5 current out = 2X bias current to stay in class A above that it will go in to class B
(this is as far as I understand things)
you can have for example 2 mosfets at 1.5 amps or 3 at 1 A
I think that altrought the sink dissipate same wattage the Junction themperature on the mosfet be lower. Let the sink go to 55 C
(Onother thing I have noticed is that after a 3 4 hour realy loud session sink felt some what cooler than idle Maybe insted of dissipating the Bias current as heat they where producing blifsfull sounds and running colder so Play loud )
Consider that Papa already said "they sound better at hi voltagge and Bias"
Papa also said Get heat sinks as big as you can aford.
Run at less than 1.3 A mosfet and you can hear it.
No point to go low voltagge your speakers will not ask the amplifier for uge current .
Say 16 V out on 8 homs speakers no mather what you will never ask more than 2 A
with 2 mosfets at 1.3 A on each rail you got 2,6 X2 A avvaliable
Yeah speakers are not resistors and impedence drop so you got a bit of head room
Any point going for 4 mosfets and 10 Amp current output.
So what options you got
P&P may be the killer but if you like same 300 X 400 x83 heat sinks like maine just PM
OR you could use a fan litle CPU coolers are capable to dissipate 100W of heat
Bolt one on the sink and run at low voltagge al you need is a spare mosfet a spare thermistor and a pot to make a themperature speed controlled Fan.
Even at very low speed difference is uge
Use wires to place mosfets on 2 sinks just make shure gate stoppers are soldered to gate pins duble cek you got no hi frequency oscillation or go balanced to double the power
OR like me wait for F6
Last edited:
Mind you, this is just a simulation -- with lateral MOSFETs you don't get the "burden" as vertical MOSFETs so the rails can be lower, you also get lower transconductance so distortion should be worse. I simmed this with 4 devices in parallel -- looks like the f3 is about 500kHz.:
An externally hosted image should be here but it was not working when we last tested it.
Apart of checking here and there, you can have your own math here.
Let's say we want the F5 turbo work at 50w@8ohm
P=50
R=8
Iout = √(50/8)
= 2.5A
Ipeak = √2 X Iout
= 3.5A
Ibias = Ipeak/2 = 1.8A
//F5 turbo is using 2 pairs of output drivers//
Each driver's current:
Ieach = Ipeak/2n
n = 2
Ieach = Ipeak/4
= 0.875A
Vo = K x Ec = √(P x R)
Set K = 0.6
Ec = √(P x R) / k
= 33.3V
when supplying +/- 35V rail, the current bias for each driver is roughly 0.8A
So you can have your own calculation to decide how to set the voltage supply and
current into the F5 turbo.
ps. Correct me if i have did the wrong calculation.
Let's say we want the F5 turbo work at 50w@8ohm
P=50
R=8
Iout = √(50/8)
= 2.5A
Ipeak = √2 X Iout
= 3.5A
Ibias = Ipeak/2 = 1.8A
//F5 turbo is using 2 pairs of output drivers//
Each driver's current:
Ieach = Ipeak/2n
n = 2
Ieach = Ipeak/4
= 0.875A
Vo = K x Ec = √(P x R)
Set K = 0.6
Ec = √(P x R) / k
= 33.3V
when supplying +/- 35V rail, the current bias for each driver is roughly 0.8A
So you can have your own calculation to decide how to set the voltage supply and
current into the F5 turbo.
ps. Correct me if i have did the wrong calculation.
I too get lost with the introduction of k.
Prior to that K, everything is correct.
Use peak output current and nominal load resistance.
3.54Apk * 8r0 requires an output voltage of ~28.3Vpk.
This is where the fudge factor k comes in.
What do we allow for the voltage losses through the output stage from Vrail to Vout?
I would guess a fixed value of 5Vlost rather than a proportion of the peak output voltage.
28.3Vpk + Vloss of 5V comes to Vrail of +-33.3Vdc. (for a ClassAB amplifier the rails need to be ~3V to 8V higher than the ClassA amplifier supply rails)
Once you have built the amplifier, you can measure Vlost and use that actual value for those devices and that peak output current from that PSU configuration for any future design or clone.
BTW,
setting the device bias to 0.8A will result in the maximum ClassA output current being less than 3.2Apk.
Increase the bias to ~0.9A to achieve your desired 3.54Apk of maximum ClassA output current.
Prior to that K, everything is correct.
Use peak output current and nominal load resistance.
3.54Apk * 8r0 requires an output voltage of ~28.3Vpk.
This is where the fudge factor k comes in.
What do we allow for the voltage losses through the output stage from Vrail to Vout?
I would guess a fixed value of 5Vlost rather than a proportion of the peak output voltage.
28.3Vpk + Vloss of 5V comes to Vrail of +-33.3Vdc. (for a ClassAB amplifier the rails need to be ~3V to 8V higher than the ClassA amplifier supply rails)
Once you have built the amplifier, you can measure Vlost and use that actual value for those devices and that peak output current from that PSU configuration for any future design or clone.
BTW,
setting the device bias to 0.8A will result in the maximum ClassA output current being less than 3.2Apk.
Increase the bias to ~0.9A to achieve your desired 3.54Apk of maximum ClassA output current.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.