You need derating on critical parts. For low thermal distortion, you need substantial derating.
This is quite a measurable effect. The two feedback resistors must behave exactly the same,
and for this to happen, the larger resistor must be composed of several series resistors that are
identical to the resistor to ground.
This is quite a measurable effect. The two feedback resistors must behave exactly the same,
and for this to happen, the larger resistor must be composed of several series resistors that are
identical to the resistor to ground.
Last edited:
F5 bias not stable problem-plese help
Just put together another F5 amp with all parts from diyaudio store and I cant get the bias and offset stable. I biased the chanels to 0.6v bias and offset to 0 volts but the bias goes constantly from 0.595 to 0.605 volt up and down and the offset from 0volts to 10mV up and down. I let the amp warm up and readjust bot trimmers but with no sucsses.I have checed all parts on boards and ewerithink is ok. My heatsinks are 20cmx50cmx7cm for chanel, so good and big, the amp warms up nice and I have no problem from start up to set bias to 0.6v but like I say the bias and offset are not stabe not for a minute. Plese for help.
Just put together another F5 amp with all parts from diyaudio store and I cant get the bias and offset stable. I biased the chanels to 0.6v bias and offset to 0 volts but the bias goes constantly from 0.595 to 0.605 volt up and down and the offset from 0volts to 10mV up and down. I let the amp warm up and readjust bot trimmers but with no sucsses.I have checed all parts on boards and ewerithink is ok. My heatsinks are 20cmx50cmx7cm for chanel, so good and big, the amp warms up nice and I have no problem from start up to set bias to 0.6v but like I say the bias and offset are not stabe not for a minute. Plese for help.
I just conected the speakers just to see if the amp is working and the amp sounds verry good, but I have see that if I turn the volume up to mid loud point the bias and offset goes totaly crazy up and down, the bias jumps from 0.6v to 0.8v up and down and the offset also. So I checked the psu rails and I see that the voltage drop on each rail is from 24.1V to 22.5V when I listen to mid loud volume. My psu is CLC so 40000uf-2,5mH(0.12ohm)-40000uf.Could it be that the psu hase to low capacitence and I must incrase this? I have hiraga le monstre and hiraga 30w A class and bouth amps working perfect without this problems. Any ideas what went wrong?
Hi Zen mod, thanks for replay. In this moment I cant check for oscilation becouse I have no osciloscope at home. Im not home at the moment but I will post some pictures. My build is standard F5 from diyaudio store with all parts from store with matched mosfet and Jfets. Its strange to me that this is now for first time I have this problem and becouse I checked the parts and boards for second time today and I couldnt find any fault.I dont know why the bias goes so up and down when amp is playing I alway have thinking that A class amps have always fix bias. Or is Pass another design?
there is no other way to be familiar with dynamic behavior of an amp, than to observe one for which you're 100% sure is functional
A Class is, in short, constant current draw from rails, but effective/combined
of course that you'll see changes across source resistors and you don't have means to observe sum of positive and negative currents, without having scope and summing function, using two scope channels
with DMM , what you can is to observe uniformity and nature of biasing recovery, right after signal burst; ignore any reading while burst is still happening and observe right after burst stop
that speaking of dynamics
when speaking of steady state ( no signal) both offset and Iq must be stable ; flicckering a little up and down is normal, because rails are never steady ( mains fluctuation) in lesser or bigger amount
though, just small flickers are normal
now - did you made your observation without load and with input shorted?
while first is not so important, second is, at least in my book - ZM being chicken and all, always expecting someone to poke in input RCA, behind my back
A Class is, in short, constant current draw from rails, but effective/combined
of course that you'll see changes across source resistors and you don't have means to observe sum of positive and negative currents, without having scope and summing function, using two scope channels
with DMM , what you can is to observe uniformity and nature of biasing recovery, right after signal burst; ignore any reading while burst is still happening and observe right after burst stop
that speaking of dynamics
when speaking of steady state ( no signal) both offset and Iq must be stable ; flicckering a little up and down is normal, because rails are never steady ( mains fluctuation) in lesser or bigger amount
though, just small flickers are normal
now - did you made your observation without load and with input shorted?
while first is not so important, second is, at least in my book - ZM being chicken and all, always expecting someone to poke in input RCA, behind my back
Zen mod thanks 100x for your help. I have two DMM(multimeter for now) Like I say before I have mesured and set the amp without load but I didnt short the input. I will do this again with shorted input tomorow, today Im not home. I will the report the results from mesurments.But to say the amp, so bouth chanels sounding good even with this problem. So the only problem I was in worry was becose the bias and offset are not rock solid stady but they go up and down like I say in first post, without load. Thanks again Zen mod.
Hi Mr. Nelson, thanks for answer. So when I measure the bias and offset, i have left the amplifier to warm up, I have set first the bias to 0.4v bouth chanels, than after 3 hours to 0.5v and than after one day to 0.6v and always looking for offset to be as near as possibile to 0 volts.Like I say I have no problem for seting bias to 0.6v and without a problem the offset to 0 volts, but al the time to now after 4 days the bias and offset where never rock stable, so when I set the bias to 0.6v and offset to near 0 volts this are never totaly stable but goes up and down a little bit, like I say the bias from 0.595v to 0.605v and offset from 0-10mV.They go up and down constantly not even for a few second this are stable, so now Im thinking could it be becouse of Psu fluctuation becouse of power grid fluctuation or is there something wrong with Jfets or output fets? But like I say I have try the amp runing on my speakers and I have no hum, only tiny bit hiss but this is so small I must put my ear on the tweeter to hear that. And amp is sounding verry, verry good, better than bouth my hiragas. So thanks for this Mr. Pass!
Hi ZenMod, so I did now the mesurments with shorted inputs but no poad and the problem is the same with jumping bias and offset. I have shorted the inputs on rca conector is this good or must I do this on pcb where the signal enters. Another thing that I have see now is that my big heatsinks are when amp is power up for a hour or more this are verry, verry hot and I can touch this for maybe a 5 second. And my heatsinks are 50x20x6,5cm big, I have thinking that this must be ok becouse is 500 cm/2 pro mosfet.
audiorado
my F 5 heat sink was 30x17x4 cm and it runs happy .7
but I have glued a little kitchen cooking thermometer to top each heatsink for piece of mind so with your heat sink size you have no worrys but if still worry put a babysiter in standby mode so it only comes on and a high temp-
my F 5 heat sink was 30x17x4 cm and it runs happy .7
but I have glued a little kitchen cooking thermometer to top each heatsink for piece of mind so with your heat sink size you have no worrys but if still worry put a babysiter in standby mode so it only comes on and a high temp-
When building my F5, I bought parts at diyaudio. They were IRFP9240 / 240 and 2SK170 / 2SJ74. As I wrote earlier on this set, I had a THD of 0.017%. I did not replace the resistors with non-inductive ones as suggested here, because on Vishay-Dale and Panasonic metallized resistors, many DIY enthusiasts including Nelson Pass alone achieve results of 0.005% to 0.002%, so I did not see the need to spend at least about 10 times more on one e.g. Caddock resistor.
I decided to replace the Mosfets with Exicon, probably the best Audio transistors currently being produced. I bought 2 complementary ECW20N20 / ECW20P20 pairs. My THD increased to 0.022% (1.8A quiescent current). I got rid of the 0.47R source resistors and the thermal protection consisting of Q5 and Q6, TH1, TH2 and the rest of the resistors (heat sink temperature 53 degrees C).
Can anyone advise me on what parameters of Jfets for ECW20N20 / ECW20P20 should be noted that my F5 has less than 0.005% THD. The Diyaudio store for IRFP offers Jfety Idss 6-8mA, what does it depend on, which part of the performance characteristics should be paid attention to?
Thank you in advance for your help.
I decided to replace the Mosfets with Exicon, probably the best Audio transistors currently being produced. I bought 2 complementary ECW20N20 / ECW20P20 pairs. My THD increased to 0.022% (1.8A quiescent current). I got rid of the 0.47R source resistors and the thermal protection consisting of Q5 and Q6, TH1, TH2 and the rest of the resistors (heat sink temperature 53 degrees C).
Can anyone advise me on what parameters of Jfets for ECW20N20 / ECW20P20 should be noted that my F5 has less than 0.005% THD. The Diyaudio store for IRFP offers Jfety Idss 6-8mA, what does it depend on, which part of the performance characteristics should be paid attention to?
Thank you in advance for your help.
Last edited: