No. All mosfet can be matched NP, but isn't easy to find out perfect pairs or quads, unless you have access to hundreds of them.
I can not rule that out but have never seen any seller offer them NP matched either.If I understand correctly Laterals can't be NP matched
I believe so because at least there is a seller offer them NP matched.but 2SK2013/2SJ313 and J201/K1530 can?
For the J313/K2013 and Laterals, 2-3 parallel output pairs will deliver lower output impedance and higher current drive capability notable on most speakers, more may be needed for difficult ones. Of course there is the F7 approach usable when you want a single Lateral output pair.I don't need NN and PP matching because I will use only one pair per channel...
Papa created the F7 for complementary Lateral output pairs, you can see the unit reviewed by 6moons use Allfet parts, very similar characteristic to Exicon. You are on your own if you want to use verticals like the Toshiba.
Yes they are.aren't 2SK2013/2SJ313 and J201/K1530 vertical?
No. All mosfet can be matched NP, but isn't easy to find out perfect pairs or quads, unless you have access to hundreds of them.
You don't need hundreds to match Vgs to 100mV -- maybe 20 or so.
This is true with IRF/Vishay devices, Renesas and Toshiba's. I can't speak to Allfet or Exicon.
Bought some 8 pairs of ALLFETS. They were very close.You don't need hundreds to match Vgs to 100mV -- maybe 20 or so.
This is true with IRF/Vishay devices, Renesas and Toshiba's. I can't speak to Allfet or Exicon.
you will find out how close when you bias up the output stage.Bought some 8 pairs of ALLFETS. They were very close.
The errors across the source resistors show how far apart the devices really are.
Problem with an F7 build is that it is still in production, Papa will not explain all build details.Ok then an F7 with unmatched laterals and matched input jfets.
As a beginner, it is likely that you do not have distortion analyzer, oscilloscope, curve tracers, temperature controlled part testing jigs and the required knowledge to properly execute, evaluate and tweak your build. You can build your interpretation of F7, but it is not easy to help you when the outcome of your build is no good.
By now you should have realized that while the Firstwatt circuits are relatively simple, parts matching and selection, PCB design, PSU in use, proper heatsinking, execution and measurement all contribute to the intended performance.
If you have not realized this, I really doubt that you can successfully build and obtain the exceptional quality everybody expect from a Nelson Pass design. It could be better for you to build the ACA kit for now, easier to get right and sounds as good, just not as powerful as the higher powered ones. Read and learn a bit more, then select and build your dream amp.
After reading up for several years, I still find the accumulated knowledge posted in Pass forum staggering, and new information kept pouring in. So don't you worry, you will find ways to improve your current best. If you have read a bit more, you would have found that commercial products can not be made the best possible sonicwise, manufacturers have to compromise for various reasons. We are DIY, we do not have to compromise, the most important part is through the journey, we come to understand the taste and behavior of our prime customer.🙂
Bought some 8 pairs of ALLFETS. They were very close.
It is helpful that their transconductance figures are low and that they also
have a low temperature coefficient. This loosens up the requirement for
close matching of Vgs.
Right Andrew, I curve traced them up to 2.5 amps. I think this is a good base. Anyway, you are right, I'll see it in the circuit.you will find out how close when you bias up the output stage.
The errors across the source resistors show how far apart the devices really are.
Cheers
Good to know. Thanks.It is helpful that their transconductance figures are low and that they also
have a low temperature coefficient. This loosens up the requirement for
close matching of Vgs.
I finally got a chance to return to this project. I have both channels up and running with a few mV of offset and biased somewhere around 340-380mA, which I measured, as a voltage drop on 0.68 Ohm source resistors. One question I'd have is that when PSU is not loaded it gives +/-24V on cap multiplier outputs, but it drops to about +/- 20.7V when under load. This is 15% drop, a bit more than what I was expecting...
Also, I'm quite surprised how hot those MOSFETs are. I literally cannot touch them, however heat sinks are just only quite pleasantly warm after an hour of cooking.
So, it seems like I'll be finally listening to some music this weekend 🙂
Also, I'm quite surprised how hot those MOSFETs are. I literally cannot touch them, however heat sinks are just only quite pleasantly warm after an hour of cooking.
So, it seems like I'll be finally listening to some music this weekend 🙂
Last edited:
PiotrL,
high voltage drop in PSU is usually a sign of low quality/weak/insufficient transformer and/or filter caps.
Big temperature difference between MOSFETs and heatsink shows bad thermal coupling of MOSFETs to heatsink - you got to make it better if you care about long term reliability.
high voltage drop in PSU is usually a sign of low quality/weak/insufficient transformer and/or filter caps.
Big temperature difference between MOSFETs and heatsink shows bad thermal coupling of MOSFETs to heatsink - you got to make it better if you care about long term reliability.
Thanks Juma. I suppose that terms like "high", "big" etc are hard to quantify. I'm taking your response as 15% voltage drop is too much. Since I have 400VA Antex 2x20V + 2 x (2x27000uF) then either 400AV toroid is too small or 100kuF in total is no enough?
As for the mosfets, are you saying that the surface of the transistor should have the same temperature as the heat sink? I find it hard to believe, to be honest.
I'm using Kerathem 86/82 for thermal coupling, have both surfaces de-greased, and can touch mosfets for about 0.5sec so I'd say they are much below their maximum rating channel temp being 150deg C. Also previously I did have my chassis opened, and when I got it closed now and left for few hours turned on then heat sinks went noticeably hotter but Toshiba's were pretty much the same.
I'm just curios what is other people's experience, are your mosfets also hard to touch?
As for the mosfets, are you saying that the surface of the transistor should have the same temperature as the heat sink? I find it hard to believe, to be honest.
I'm using Kerathem 86/82 for thermal coupling, have both surfaces de-greased, and can touch mosfets for about 0.5sec so I'd say they are much below their maximum rating channel temp being 150deg C. Also previously I did have my chassis opened, and when I got it closed now and left for few hours turned on then heat sinks went noticeably hotter but Toshiba's were pretty much the same.
I'm just curios what is other people's experience, are your mosfets also hard to touch?
- Status
- Not open for further replies.
- Home
- Amplifiers
- Pass Labs
- F5 with 2SK2013/2SJ313