F5 Cascode and parallel output device questions

texas11 · 12th August 2010, 08:16 AM

I'm new to DIY Audio and I'm building an F5 as my first project. I'm going to build it to spec (except changing R20 and R21 to 22K).

I'm curious about trying to get more watts into 8 ohms. Is the purpose of cascoding so one can increase operating voltage and get more watts? I see folks adding parallel output devices (multiple IRFP9240s and IRFP240s), but it seems that only helps at low impedance.

What is the advantage of cascoding?

Can one increase biasing with more output devices (provided one has the coolig to do so) to increase power?

I'm guessing these questions are fairly simple, but I was confused after a few dozen pages of the F5 thread.

juma · 12th August 2010, 10:03 AM

texas11,
just keep reading, that story has been told in this thread at least 3 times...

Russellc · 11:26 AM

Quote:

Originally Posted by juma

texas11,
just keep reading, that story has been told in this thread at least 3 times...

I dont blame him, I've read the whole thing and I'm still not clear, and I was one of the ones asking one of those 3 times! The story keeps getting told, and while there is good info in Cviller
s blog, as far as implementing it with his boards, not much info as to effect, short of stating more current can be delivered into difficult loads and so forth, but nothing as to power output other than stating it could be done....

Russellc

flg · 12th August 2010, 11:50 AM

Quote:

Originally Posted by juma

texas11,
just keep reading, that story has been told in this thread at least 3 times...

Simple logic would not explain how a story could be told at least 3 times in a thread with only 2 posts??? I'll bet juma means one of the F5 threads and not this one

Cascoding has advantages of allowing higher rails and at the same time to protect a lower voltage or lower power device from extreme conditions. It also controls some of the operating parameters of the cascoded device helping achieve more linear operation... You can find more at one of the Pass sites especially in Nelsons Cascoding paper.
Output power is mostly determined by the available output voltage and load impeadance. In the F5 you may acheive a particular power level with 8 ohm load purely based on those factors. But, the amp will run in class A mode only to a point, after which it begins to operate in class B mode.
More output devices may allow the amp to go to higher power in class A mode before operating in class B.
There are other considerations that will effect the sound of this amp if you decide to deviate from Nelsons original suggestions. Those details have been discussed in the F5 threads and this post is long enoug already. It has to do with the balancing of the top/bottom transistors and the input/output transistors and there gain and non linearities canceling each other. I believe Nelson said something like the devil is in the details.

AndrewT · 12th August 2010, 11:50 AM

what's difficult about linking the statements of "using double the number of devices in parallel" to another statement that the doubled up F5 is "capable of double the current output" and the "F5 will dissipate twice the heat" and "double the bias current"?

Sorry, but it seems so obvious.

Similarly for power comparisons and for voltage comparisons (cascoded devices)

Bill Fuss · 12th August 2010, 11:52 AM

texas11,

Basically, the amp has been optimised for an 8 ohm load, so increasing the available power would involve increasing the rail voltages AND available current. Cascoding allows much higher rail voltages, and multiple output devices allows more current capacity. Doing one without the other won't get you very far.

Are you sure you need more than what the standard F5 will supply?

Russellc · 12th August 2010, 12:51 PM

Quote:

Originally Posted by Bill Fuss

texas11,

Basically, the amp has been optimised for an 8 ohm load, so increasing the available power would involve increasing the rail voltages AND available current. Cascoding allows much higher rail voltages, and multiple output devices allows more current capacity. Doing one without the other won't get you very far.

Are you sure you need more than what the standard F5 will supply?

It would be nice, but it looks like it involves changing it from an F-5 to something else. I really like the sound of the amp, if all that changed, it just be another amp possibly. It will have to serve me as is, I'll just look for one of Nelson's more powerful amps to DIY.

Russellc

juma · 01:13 PM

Ok guys, one more time:

To get a higher power on 8 Ohm speaker we need more voltage swing at output - in standard F5 it's 20V peak, for 50W rms we'll need about 28-29Vpeak. To achieve that we'll need +/-33V PS voltage and about 3.6A peak through speaker (if we want all of the 50W rms in class A the bias current should be 1.8A and that means two pairs of output mosfet with 900mA running through each pair). Personally, I'd use a 500VA toroid in PS and double amount of filter Elkos, but it can be negotiated. Bigger heatsinks are a must, too (Pd=66V *1.8A = 119W per channel - almost the double).
Simple, right?
Now comes the cascoding - why do we need it?
With higher PS voltage comes the greater dissipated power in input JFETs (plus, the 2sj74's datasheet says that it can't take Vgds higher than 25V, although Nelson tested them and says it's safe up to about 40V). So, the first objective of cascoding is to protect the JFETs from high/risky Pd/Vgds. The additional benefit is that cascoding enables JFETs to work in constant Vds (or near-constant Vds) regime which decreases influence of parasitic capacitancies and lowers the distortion, extends freq. range, improves the sound and kills werewolfs.
Are we in clear now?

tinitus · 12th August 2010, 01:51 PM

http://www.diyaudio.com/forums/blogs...version-2.html

but for a first build, I dont know

I have been around this too
even bought some parts
and I probably will build a bigger one
be it BA-1 or 2

for optimal sound quality I have decided to better build original F5 design
I think its a nice size for classA, nice and simple
and if something goes wrong its not too hard to handle

realisticly, what you need to build a bigger F5 may actually be enough to build 2x F5 stereo amps of original design(4 channels)
they could be used for biamp as is, or maybe bridged to give more power
but expect it to get really hot on warm days

texas11 · 12th August 2010, 10:03 PM

I'll chime in again.

It's not obvious to me, never having built an amp before that one can "just double" the bias and get double watts. I would expect practical limitations of any design, and would hope someone could give specific examples of experimants they've run.

Further, You guys are a bit too insular. The last time I looked at a circuit was in 8th grade. I read the F5 thread a number of times, that's how I came up with my questions. Not having built any DIY audio before, cascoding and multiple output devices seem logical, but since few seem to do it, one wonders what the limitations are, and well as the specific benefits.

I'm still trying to understand how resistance relates to transconductance from a conceptual perspective - cut a guy some slack.

12th August 2010, 08:16 AM	#1
texas11 diyAudio Member Join Date: Jun 2010	F5 Cascode and parallel output device questions I'm new to DIY Audio and I'm building an F5 as my first project. I'm going to build it to spec (except changing R20 and R21 to 22K). I'm curious about trying to get more watts into 8 ohms. Is the purpose of cascoding so one can increase operating voltage and get more watts? I see folks adding parallel output devices (multiple IRFP9240s and IRFP240s), but it seems that only helps at low impedance. What is the advantage of cascoding? Can one increase biasing with more output devices (provided one has the coolig to do so) to increase power? I'm guessing these questions are fairly simple, but I was confused after a few dozen pages of the F5 thread.

12th August 2010, 12:57 PM	#8
juma diyAudio Member Join Date: Nov 2005 Location: Zemun	Ok guys, one more time: To get a higher power on 8 Ohm speaker we need more voltage swing at output - in standard F5 it's 20V peak, for 50W rms we'll need about 28-29Vpeak. To achieve that we'll need +/-33V PS voltage and about 3.6A peak through speaker (if we want all of the 50W rms in class A the bias current should be 1.8A and that means two pairs of output mosfet with 900mA running through each pair). Personally, I'd use a 500VA toroid in PS and double amount of filter Elkos, but it can be negotiated. Bigger heatsinks are a must, too (Pd=66V 1.8A = 119W per channel - almost the double). Simple, right? Now comes the cascoding - why do we need it? With higher PS voltage comes the greater dissipated power in input JFETs (plus, the 2sj74's datasheet says that it can't take Vgds higher than 25V, although Nelson tested them and says it's safe up to about 40V). So, the first objective of cascoding is to protect the JFETs from high/risky Pd/Vgds. The additional benefit is that cascoding enables JFETs to work in constant Vds (or near-constant Vds) regime which decreases influence of parasitic capacitancies and lowers the distortion, extends freq. range, improves the sound and kills werewolfs. Are we in clear now? Last edited by juma; 12th August 2010 at 01:13 PM.*

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Parallel chips - output resistor questions	sangram	Chip Amps	9	5th February 2010 09:05 AM
Best voltage reference device for a cascode BJT CCS?	ray_moth	Tubes / Valves	2	1st July 2009 09:38 PM
cascode output stage what device?	Stefanoo	Pass Labs	24	27th February 2009 03:35 PM
SE parallel output questions	Richard Ellis	Tubes / Valves	3	18th May 2008 12:33 PM
New mixed-device-type symmetrical folded cascode amp	Mr Evil	Solid State	31	15th August 2005 11:37 AM

New To Site?	Need Help?
Register to Participate Search Privacy Statement Contact Us	Frequently Asked Questions Did you forget your password? Mark Forums Read


	Please consider donating to help us continue to serve you. Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving

12th August 2010, 10:03 AM	#2
juma diyAudio Member Join Date: Nov 2005 Location: Zemun	texas11, just keep reading, that story has been told in this thread at least 3 times...

12th August 2010, 11:50 AM	#5
AndrewT diyAudio Member Join Date: Jul 2004 Location: Scottish Borders	what's difficult about linking the statements of "using double the number of devices in parallel" to another statement that the doubled up F5 is "capable of double the current output" and the "F5 will dissipate twice the heat" and "double the bias current"? Sorry, but it seems so obvious. Similarly for power comparisons and for voltage comparisons (cascoded devices)

12th August 2010, 11:52 AM	#6
Bill Fuss RIP Join Date: Oct 2005 Location: Central Florida	texas11, Basically, the amp has been optimised for an 8 ohm load, so increasing the available power would involve increasing the rail voltages AND available current. Cascoding allows much higher rail voltages, and multiple output devices allows more current capacity. Doing one without the other won't get you very far. Are you sure you need more than what the standard F5 will supply?

12th August 2010, 01:51 PM	#9
tinitus diyAudio Moderator Join Date: Nov 2005	http://www.diyaudio.com/forums/blogs...version-2.html but for a first build, I dont know I have been around this too even bought some parts and I probably will build a bigger one be it BA-1 or 2 for optimal sound quality I have decided to better build original F5 design I think its a nice size for classA, nice and simple and if something goes wrong its not too hard to handle realisticly, what you need to build a bigger F5 may actually be enough to build 2x F5 stereo amps of original design(4 channels) they could be used for biamp as is, or maybe bridged to give more power but expect it to get really hot on warm days

12th August 2010, 10:03 PM	#10
texas11 diyAudio Member Join Date: Jun 2010	I'll chime in again. It's not obvious to me, never having built an amp before that one can "just double" the bias and get double watts. I would expect practical limitations of any design, and would hope someone could give specific examples of experimants they've run. Further, You guys are a bit too insular. The last time I looked at a circuit was in 8th grade. I read the F5 thread a number of times, that's how I came up with my questions. Not having built any DIY audio before, cascoding and multiple output devices seem logical, but since few seem to do it, one wonders what the limitations are, and well as the specific benefits. I'm still trying to understand how resistance relates to transconductance from a conceptual perspective - cut a guy some slack.