I understand the importance of matching Mosfets within an N channel or P channel but I am not sure whether it is important to match between N and P channels. Obviously this leads to a difference between N and P output bias voltages which you adjust to zero the offset but do the two numbers have to be within a range say 10% and why?
This is for a F5TV3 with 4 N and 4 P devices per channel.
Also when a supplier provides a Vgs value and a Yfs value which is the more relevant of the two to consider?
Thanks.
Nash
This is for a F5TV3 with 4 N and 4 P devices per channel.
Also when a supplier provides a Vgs value and a Yfs value which is the more relevant of the two to consider?
Thanks.
Nash
Thanks both of you.
So for example lets say N channel bias voltage is .30V and P channel is .26V to obtain zero offset, is this much of a spread still acceptable? In a mono with N on one side and P on the other, wont this result in different heat sink temperatures?
Could someone please explain Yfs measurements. I cant find much on this anywhere.
Nash
current goes (by nomenclature) from positive PSU to negative PSU
that means that upper and lower halves of amp are on same (and only ) route (disregarding slight current through feedback net) ........ so difference in voltage across source resistors is coming from difference in their value ( tolerance!) , not difference in current
so - read tutorials for F5 , several times and it will click eventually
as long you're in dark regarding more basic issues , there is no use of knowing how to measure Yfs
however , at FW site you'll find articles for everything , except 42
search for articles having test and mosfet in title
Thanks Zen Mod. I am learning. Understand the concept about the current going thru the two halves. Using the example of bias voltage 0.26V and 0.30V over 0.5ohms source resistance for zero offset gives me bias current of 0.52A and 0.6A; so is it correct to say that the bias current is an average of the two at 0.56A?
irrelevant
choose one side Rs to monitor Iq (with first DVM) , put second DVM at output (ref. to gnd) to monitor offset
you can put 3-rd DVM to adjacent side Rs , but just to confirm that physics is still valid in Universe ;
as long that's the case (physic's ruling) and you didn't dreked something substantial , amp will be functional and Smarty Current will find it's way where Pa intended ........ and difference in reading between 1-st and 3-rd DVM-s will be probability sum of tolerances of two Rs-es and two DVM-s , combined
inputs shorted , no load at output
choose one side Rs to monitor Iq (with first DVM) , put second DVM at output (ref. to gnd) to monitor offset
you can put 3-rd DVM to adjacent side Rs , but just to confirm that physics is still valid in Universe ;
as long that's the case (physic's ruling) and you didn't dreked something substantial , amp will be functional and Smarty Current will find it's way where Pa intended ........ and difference in reading between 1-st and 3-rd DVM-s will be probability sum of tolerances of two Rs-es and two DVM-s , combined
inputs shorted , no load at output
Last edited:
your two resistors marked 0r5 are probably +-10% tolerance.
If you want or need to measure currents through them, then YOU need to measure them before assembly.
Now I know that I can't measure 0r5 accurately using a DMM set to resistance.
I know that I need to use a method that let's me get closer to the actual resistance.
Now you need to ask yourself:
"is it worth my effort to research the methods available to match, or measure, low value resistors?"
Alternatively,
you could use 9off 4r7 1% resistors in parallel and then add 12r 1% in parallel to give a final value of 0r5004 +-1%
you could use 9off 4r7 1% resistors in parallel and then add 12r 1% in parallel to give a final value of 0r5004 +-1%
"is it worth my effort to research the methods available to match, or measure, low value resistors?"[/QUOTE]
I wonder what questions you've done to yourself this morning to find out what the value philosophically your intervention in this thread, making little professor of donkeys. My dear Scottish tempered
I wonder what questions you've done to yourself this morning to find out what the value philosophically your intervention in this thread, making little professor of donkeys. My dear Scottish tempered
Thanks Andrew, but they are 1%.
Yes, I have researched some of the ways of obtaining a finer reading on small value resistors including your own contribution on the subject, but with 1% resistors the max bias voltage error is also 1%.
Just to reiterate my original query it was about the relevance of the difference in bias voltage between the N side and the P side and not between devices on each side.
Nash
with 1% source resistors your current estimates are going to be accurate to roughly +-1%.
You probably don't need any more than that.
That allows you to compare the upper current/s to the lower Current/s.
When you have no idea what tolerance your source resistors are then you are blowing in the wind.
You probably don't need any more than that.
That allows you to compare the upper current/s to the lower Current/s.
When you have no idea what tolerance your source resistors are then you are blowing in the wind.
Looking for a little help 🙂
I am building a stock F5 and had initially placed the order for the F5 trans.kit with Hannes (h_a) Unfortunately he is out of stock and his scheduled delivery is taking a global detour.
Are there anyone here who happen to have a complete F5 transistor/fet kit lying about? Part.# reference Q1-Q6 (2 of each) PM me with details if you are able to help.
Thank you for looking 🙂
Haldor
I am building a stock F5 and had initially placed the order for the F5 trans.kit with Hannes (h_a) Unfortunately he is out of stock and his scheduled delivery is taking a global detour.
Are there anyone here who happen to have a complete F5 transistor/fet kit lying about? Part.# reference Q1-Q6 (2 of each) PM me with details if you are able to help.
Thank you for looking 🙂
Haldor
Here's the definitive word: http://www.diyaudio.com/forums/pass-labs/121228-f5-power-amplifier-1449.html#post4136971
I understand this as: "A slight difference in current to correctly balance the output is normal".
My first F5 used 5% resistors as per jackinnj's kit, and the currents were off by a small amount but it didn't affect performance, the amp sounded great.
I wouldn't worry about it.
Just to be totally safe I am now using the 9140 instead of the 9240 as it is a slightly better match.
Thanks. I read NP's comment "The larger problem is that they temperature coefficients of the N and P are
not identical, and we depend on feedback to take care of that with regard to
DC offset."
What is 9140, 9240?
Nash
irf do the 240 and 9240, upto 200Vds.
They do not make good complements for linear audio.
They also do the 140 and 9140, upto 100Vds.
They too do not make good complements for linear audio.
They were designed for switching, i.e. off, on, off, etc.
Many point to the 9140 being a better complement to the 240, but this limits the amp to 100V rail to rail. Not a problem for the F5.
There is some discussion and test results posted very recently comparing all these devices and from different manufacturers.
They do not make good complements for linear audio.
They also do the 140 and 9140, upto 100Vds.
They too do not make good complements for linear audio.
They were designed for switching, i.e. off, on, off, etc.
Many point to the 9140 being a better complement to the 240, but this limits the amp to 100V rail to rail. Not a problem for the F5.
There is some discussion and test results posted very recently comparing all these devices and from different manufacturers.
- Status
- Not open for further replies.