Thanks flg, but neither of those guys have 2sk1530/2sj201 for sale. And I'd still like to know if the two mosfets that I mentioned in post #9697 would be OK in this amp; they're cheap too.
almost 3000pf seems to be a bit high, but I really dont know
but I have looked at hundreds of mosfets the last few days, and every time I look at a nice strong one, capacity goes up 😱
why not the original IRFP9240/240
Im looking at maybe using the newer IRFP9140/140
someone even suggested 9140 with 240
hey, I can get all 4 outputs for something like 20USD
also easier and much cheaper to repair, should it be needed
who knows, maybe theres a completely new and much better mosfet in a couple of years
one should really hope so, eh 😀
anyway, I can live with the second best 😛
Regarding Zhou Fang, I think you would have to e-mail him, you could be right. Sometimes he takes awhile to get back to you. Just PM him and ask.
Regarding Blues, looks like he has 15-20 pairs on his spreadsheet???
Regarding the FQA28N15, I've used it in ZV4 and I liked it. I still have 10-20 😉
I don't know about the other one. The C you quoted is double the FQA28N15.
A typical sustitute pair would be the FQA12P20/FQA19N20. But as I'm sure you realize, few will be as complimentary as the Toshiba's.
Regarding Blues, looks like he has 15-20 pairs on his spreadsheet???
Regarding the FQA28N15, I've used it in ZV4 and I liked it. I still have 10-20 😉
I don't know about the other one. The C you quoted is double the FQA28N15.
A typical sustitute pair would be the FQA12P20/FQA19N20. But as I'm sure you realize, few will be as complimentary as the Toshiba's.
Lars... I'm ashamed to say 😱 ... I was looking for something on the cheap. I did notice the high capacitance, hence my question. In order to get a few good matches of irfp240/9240 I'd have to buy quite a few, no? The price is about $5 a piece, at mouser. In comparison, the fqa36n15 is $2.62 a piece, and the fqa28n15 is just $1.67 a piece.
flg, I bought stuff from Zhou Fang before, so I'll just ask what he's got. If I could get a couple of matched pairs at a reasonable price, sure. Regarding the fqa12p20, it seems to be an obsolete part, hard to find.
I got a chunky tranformer, a bunch of Sprague 33mF caps, two hammond reactor 102mH/2A chokes for the psu. Got tons of 2sk170 and a few of 2sj74. Have a case and large heat sinks as well, but still missing the output mosfets, 3W resistors, and and zetex parts.
flg, I bought stuff from Zhou Fang before, so I'll just ask what he's got. If I could get a couple of matched pairs at a reasonable price, sure. Regarding the fqa12p20, it seems to be an obsolete part, hard to find.
I got a chunky tranformer, a bunch of Sprague 33mF caps, two hammond reactor 102mH/2A chokes for the psu. Got tons of 2sk170 and a few of 2sj74. Have a case and large heat sinks as well, but still missing the output mosfets, 3W resistors, and and zetex parts.
For the matched toshiba I think that EUVL may be selling or can give you same pointers on where to get them at least in the EU may be worth looking at the balanced F5 tread.
For the Farchild I have 100 each of the FQA12P20 and FQP19N20C
I could not get the FQA19N20 those are reccomended substitute by Papa in his paper
Please note FQP19N20C is a TO220 case I can't see any differencein the speck apart from slightly lower current and power dissipation.
Is a bit to early for me to sell any but I can alwayse help a felow Diyer send me a PM if you get stuck
Maybe some one more experienced can point out the difference between the FQP and the FQA.
Hi Mellonhead
Agree 100% corectley vired equipment does not need the thermistor.
Agree 100% about the "cowboys"
One thing is for shure the noise generated by the ground loop is there because there is a potential difference between 2 point of the circuit.
If the potential difference is zero = no noise
The way I see this is either zero potential difference which is impossible or no loops for the potential difference to take place and so no ground loop
The way I was suggesting is to avoid making a ring circuit and use star or bus of star circuit as you say only one conection I was proposong to cut the screen cable as the reciving end if there is a ground loop
So just enjoy the noise one minute and then snip noise gone maybe big maybe as it can be a real pain to stop the electrons from getting boucing up and down the most silly places
CL60 to stop the noise sorry don't think so maybe reduce it a litle bit or maybe make it worst as we are back to zero potential difference = no noise
There is one thing I am not shure I understand IMO there shuld be no current at all flowing troug the Earth cable except when there is a fault condition.
now 0 current is impossible to get More realisticaly here in the UK there is a safe limit apcepatable for domestic systems they have RCD at 30mA and in Hospitals the maximum current is limited to 5mA
In regard to the CL60 TH1 could it be there to reduce the tumph on the speakers at start
up?
For the Farchild I have 100 each of the FQA12P20 and FQP19N20C
I could not get the FQA19N20 those are reccomended substitute by Papa in his paper
Please note FQP19N20C is a TO220 case I can't see any differencein the speck apart from slightly lower current and power dissipation.
Is a bit to early for me to sell any but I can alwayse help a felow Diyer send me a PM if you get stuck
Maybe some one more experienced can point out the difference between the FQP and the FQA.
Hi Mellonhead
Agree 100% corectley vired equipment does not need the thermistor.
Agree 100% about the "cowboys"
One thing is for shure the noise generated by the ground loop is there because there is a potential difference between 2 point of the circuit.
If the potential difference is zero = no noise
The way I see this is either zero potential difference which is impossible or no loops for the potential difference to take place and so no ground loop
The way I was suggesting is to avoid making a ring circuit and use star or bus of star circuit as you say only one conection I was proposong to cut the screen cable as the reciving end if there is a ground loop
So just enjoy the noise one minute and then snip noise gone maybe big maybe as it can be a real pain to stop the electrons from getting boucing up and down the most silly places
CL60 to stop the noise sorry don't think so maybe reduce it a litle bit or maybe make it worst as we are back to zero potential difference = no noise
There is one thing I am not shure I understand IMO there shuld be no current at all flowing troug the Earth cable except when there is a fault condition.
now 0 current is impossible to get More realisticaly here in the UK there is a safe limit apcepatable for domestic systems they have RCD at 30mA and in Hospitals the maximum current is limited to 5mA
In regard to the CL60 TH1 could it be there to reduce the tumph on the speakers at start
up?
Thanks, i have seen people use a second pair of Toshibas. Is it for sharing the load. How can i extend the standard circuit for te second pair?
Thanks bksabath, I think I'll try the ones I mentioned, and when I come across a good deal on the toshibas I can always replace the fqas.
Just for laughs I ran a quick fft test with the irfp and with the fqas. Only one mosfet (not a paralleled pair), input 1kHz 1V p-p, output about 6V p-p into a 6R load, with about 1.3A bias in the output mosfets.
The fft on the left is with the irfps, the one on the right with the fqas. Not so different, really. The frequency response is obviously not as good with the fqas, but still going well past 300kHz in a straight line.
Haven't run a stability analysis yet. Anyway, it's just for fun. I can always get the proper devices if this isn't working well.
Just for laughs I ran a quick fft test with the irfp and with the fqas. Only one mosfet (not a paralleled pair), input 1kHz 1V p-p, output about 6V p-p into a 6R load, with about 1.3A bias in the output mosfets.
The fft on the left is with the irfps, the one on the right with the fqas. Not so different, really. The frequency response is obviously not as good with the fqas, but still going well past 300kHz in a straight line.
Haven't run a stability analysis yet. Anyway, it's just for fun. I can always get the proper devices if this isn't working well.
Attachments
I used the modu 4U 300mm deep. Mine is biased to about 750mV (~1.6A per FET) and the sinks get hot but I can hold my hands on them for 30+seconds.
Looks like these are out of stock everywhere. I would use them if I had them.
Nationals in F5 sound fine.
That's neat. Did you mount the boards on the heat sink or just the mosfets?
I'm considering both options.
Ouch, mouser has them on order, ship date mid June 2011. I didn't notice this before 🙁 back to look for something else.
Where are people sourcing the CL60 thermistors in the US? Mouser and Digi-Key don't seem to have them.
Although I will happily admit that I may not be searching for the right thing on their websites...
Although I will happily admit that I may not be searching for the right thing on their websites...
24v supply:
Doubling the outputs allows to double the total bias current.
Doubling the bias keeps 25w class A into a 2 ohms load. (Or difficult).
33v supply + Jfet cascoding:
You can get 50w classA into 8 ohms biasing each leg at 0.85A.
http://www.diyaudio.com/forums/blogs/cviller/191-gb-f5-guide-pcb-version-2.html
Last edited:
Ikoflexer:
There are a number of sources for matched fets, IRFPs you can have from John (TechDIY) and from me, Toshibas are sold by Blues.
By the way, the original Fairchilds are obsolete already for a long time now. Certainly interesting to note that the original F5-schematics show IRFP240/IRFP9240 and not the Fairchild-fets.
Have fun, Hannes
There are a number of sources for matched fets, IRFPs you can have from John (TechDIY) and from me, Toshibas are sold by Blues.
By the way, the original Fairchilds are obsolete already for a long time now. Certainly interesting to note that the original F5-schematics show IRFP240/IRFP9240 and not the Fairchild-fets.
Have fun, Hannes