Q-FETs
Hi,
If you mean FQA19N20 and FQP12P20, see: http://home.tiscali.nl/data.odyssey/AutoBias.html
Although this article only addresses bias issues, I hope it is useful for you.
Compared to the Toshiba devices, these parts were a bit more prone to HF oscillations.
Regards,
Edmond.
CBS240 said:Has anyone played with Fairchild's Q-FET line of vertical fets? The properties seem like hex type fets. I assume that's what they are.Anyway, my experience is that they are quite rugged and can take a good toasting.
I used them with HEC circuit.
Hi,
If you mean FQA19N20 and FQP12P20, see: http://home.tiscali.nl/data.odyssey/AutoBias.html
Although this article only addresses bias issues, I hope it is useful for you.
Compared to the Toshiba devices, these parts were a bit more prone to HF oscillations.
Regards,
Edmond.
> The higher the beta, the lower the Early voltage.
I was thinking of a discrete Class-B voltage follower stage for an op-amp, running off +/- 15v rails, and Vbe-biased with two 1N4148s. However, one would need a PNP complement as well, or go quasi-complementary, which defeats the purpose of minimalism.
Another application I can think off is a low-power 4-transistor classic JLH running off a single +12v rail.
Low-power single-ended Class-A cascode, maybe with an inductive load (as discussed a few days ago in the germanium thread).
LTP current mirrors, where the Early voltage doesn't matter, but the emitter resistance multiplier is very useful - even better on a Wilson Current Mirror, preferably as the lower two transistors. It may work well even without any emitter degeneration.
I was thinking of a discrete Class-B voltage follower stage for an op-amp, running off +/- 15v rails, and Vbe-biased with two 1N4148s. However, one would need a PNP complement as well, or go quasi-complementary, which defeats the purpose of minimalism.
Another application I can think off is a low-power 4-transistor classic JLH running off a single +12v rail.
Low-power single-ended Class-A cascode, maybe with an inductive load (as discussed a few days ago in the germanium thread).
LTP current mirrors, where the Early voltage doesn't matter, but the emitter resistance multiplier is very useful - even better on a Wilson Current Mirror, preferably as the lower two transistors. It may work well even without any emitter degeneration.
Hi Edmond
That is an interesting article, I haven't seen it before, thanks.
I like the idea of not having a pot to adjust bias.
To be more specific, I used FQP14N15 and SFP9540. They were certainly cheap so if it didn't work, oh well.
I used a circuit resembling Mr. Cordell's hexfet output stage, but with an entirely different driving section. The BJT's used as compensating devices tend to over compensate so I set the initial bias to a higher level. When it warms up, the bias reduces a bit, but still stays above 150mA. These transistors like a higher bias anyway, IMHO. I figured since the HEC circuit is geared to correcting the error of each transistor independent of the other, matching transconductance wouldn't be as big a problem as long as the transconductance is suitable for the application. The whole premise for me is, could I use these cheap, relatively non-linear devices for a high end audio output stage and utilize their phat SOA. You are right on about the HF oscillations with these, a bit touchier than the IRF's. Gate stopper resistors have to be chosen carefully......a bit of trial and error. It appears not many folks have tried these for OPS with great success.
That is an interesting article, I haven't seen it before, thanks.
I like the idea of not having a pot to adjust bias.To be more specific, I used FQP14N15 and SFP9540. They were certainly cheap so if it didn't work, oh well.
I used a circuit resembling Mr. Cordell's hexfet output stage, but with an entirely different driving section. The BJT's used as compensating devices tend to over compensate so I set the initial bias to a higher level. When it warms up, the bias reduces a bit, but still stays above 150mA. These transistors like a higher bias anyway, IMHO. I figured since the HEC circuit is geared to correcting the error of each transistor independent of the other, matching transconductance wouldn't be as big a problem as long as the transconductance is suitable for the application. The whole premise for me is, could I use these cheap, relatively non-linear devices for a high end audio output stage and utilize their phat SOA. You are right on about the HF oscillations with these, a bit touchier than the IRF's. Gate stopper resistors have to be chosen carefully......a bit of trial and error. It appears not many folks have tried these for OPS with great success.
Hi CBS,
You're right, the bias is a bit overcompensated. That's why Bob Cordell put only one of the two sensing trannies on the heat sink.
See also the discussion starting about here:
http://www.diyaudio.com/forums/showthread.php?postid=1486884#post1486884
and here for a schematic with a somewhat lower T.C.
http://www.diyaudio.com/forums/showthread.php?postid=1501955#post1501955
(BTW, you better ignore syn08, he is just mad at me)
How well the FQP14N15/SFP9540 pair will work in a HEC circuit, sorry, I can't tell you. Anyhow, good luck.
Regards,
Edmond.
You're right, the bias is a bit overcompensated. That's why Bob Cordell put only one of the two sensing trannies on the heat sink.
See also the discussion starting about here:
http://www.diyaudio.com/forums/showthread.php?postid=1486884#post1486884
and here for a schematic with a somewhat lower T.C.
http://www.diyaudio.com/forums/showthread.php?postid=1501955#post1501955
(BTW, you better ignore syn08, he is just mad at me)
How well the FQP14N15/SFP9540 pair will work in a HEC circuit, sorry, I can't tell you. Anyhow, good luck.
Regards,
Edmond.
Well, they (FQP14N15/SFP9540) do seem to work OK. I have an "ugly" vero board prototype amp actually built. Sounds absolutely wonderful. At this time, both compensating BJT's are on the heat sink. I may take one off and see how much bias drift occurs. I thought about doing this before, but wasn't sure if the EC action would be affected. According to Bob, it isn't. Actually, there are two separate HEC circuits driving the speaker from a balanced bridge VAS. Also the HEC circuits are not within a feedback loop, just the local EC feedback. Apparently it doesn't have to be within a global feedback loop.
The next prototype will be a 'real' PCB. I am planning to use a physical arrangement similar to how Bonsai constructed it here. Seems like it should work. The compensating device would be in contact with the drain lead as close as possible to the output fet. If it overcompensates too much, I could insulate one of the transistors. Hope this is plausable. A little bit of trial and error I suppose. Think I may try FQP33N10 & FQP22P10.
At this time, both compensating BJT's are on the heat sink. I may take one off and see how much bias drift occurs. I thought about doing this before, but wasn't sure if the EC action would be affected. According to Bob, it isn't. Actually, there are two separate HEC circuits driving the speaker from a balanced bridge VAS. Also the HEC circuits are not within a feedback loop, just the local EC feedback. Apparently it doesn't have to be within a global feedback loop.The next prototype will be a 'real' PCB. I am planning to use a physical arrangement similar to how Bonsai constructed it here. Seems like it should work. The compensating device would be in contact with the drain lead as close as possible to the output fet. If it overcompensates too much, I could insulate one of the transistors. Hope this is plausable. A little bit of trial and error I suppose.
Think I may try FQP33N10 & FQP22P10.
You can get these fairchild transistors from digikey. Pretty cheap too! The FJL4315/FJL4215 are a bit cheaper than Toshiba 2SC5200/2SA1943. Onsemi MJL3281/MJL1302 are a bit more expensive than the Toshibas and have much better SOA. They might be a bit slower though.
For high power it will probably be very hard to beat the MJ21195/MJ21196 pair in TO3 though. The cost for those is just a little higher than MJL3281/MJL1302. OTOH you get 2 Fairchild TO264 for the price of one TO3 Onsemi. The SOA of 2 fairchild/toshibas is pretty much the same as one onsemi TO3 with 70 deg heatsink as long as the voltage isn't too high. At high voltage the Onsemi wins! (Class H is a good solution though)
For high power it will probably be very hard to beat the MJ21195/MJ21196 pair in TO3 though.
The cost for those is just a little higher than MJL3281/MJL1302. OTOH you get 2 Fairchild TO264 for the price of one TO3 Onsemi. The SOA of 2 fairchild/toshibas is pretty much the same as one onsemi TO3 with 70 deg heatsink as long as the voltage isn't too high. At high voltage the Onsemi wins! (Class H is a good solution though)
I have those Fairchild and i like them a lot
They have better gain, more matched gain from NPN to PNP and seems to me has better heat transference to the heatsink.
I am always searching for them... and when i find them i prefer to buy them instead of Toshiba units.
It is interesting to see....i have used a 2SC5200 from Toshiba and a 2SA1943 from Fairchild..... both had 90 of gain..... but the Toshiba had the upper case hot when driving high power.... the Fairchild was not hot...only the heatsink was hot...so...the heat transference was better to the Fairchild unit.
Transistors change characteristics a lot under heat...better to use them cold...big heatsinks and fan blowers..... and to that purpose, the Fairchild was tested and proved to be better.
Carlos
They have better gain, more matched gain from NPN to PNP and seems to me has better heat transference to the heatsink.
I am always searching for them... and when i find them i prefer to buy them instead of Toshiba units.
It is interesting to see....i have used a 2SC5200 from Toshiba and a 2SA1943 from Fairchild..... both had 90 of gain..... but the Toshiba had the upper case hot when driving high power.... the Fairchild was not hot...only the heatsink was hot...so...the heat transference was better to the Fairchild unit.
Transistors change characteristics a lot under heat...better to use them cold...big heatsinks and fan blowers..... and to that purpose, the Fairchild was tested and proved to be better.
Carlos
I'd like to get both versions, but mainly interested in the 80W TO-220 version. That's very good for a smaller transistor, and would be great in numerous projects where a TO-220 is a better fit than larger transistors. A single pair would be great for amps up to 60W and take little space.
But the big TO-264 seems pretty good too. Always great to get a bunch of TO-264 cheap. TO-264 is the best for high power IMO because of their size and they transfer heat really well to the heatsink, and mount much easier than a TO-3. Most of the time when I need more than a TO-220, I usually go right to the TO-264.
But the big TO-264 seems pretty good too. Always great to get a bunch of TO-264 cheap. TO-264 is the best for high power IMO because of their size and they transfer heat really well to the heatsink, and mount much easier than a TO-3. Most of the time when I need more than a TO-220, I usually go right to the TO-264.
These from Toshiba looks quite nice for drivers in lower power amps or maybe VAS or cascodes:
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail?name=2SA1837FM-ND
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail?name=2SC4793FM-ND
Fast, quite cheap and looks like they are in an insulated case too.
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail?name=2SA1837FM-ND
http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail?name=2SC4793FM-ND
Fast, quite cheap and looks like they are in an insulated case too.
megajocke said:
Yes, If I could get some of those, please
2SC4793 and 2SA1837 has often been mentioned at this forum
as being good drivers.
( do a http://www.diyaudio.com/forums/search.php 2SA1837 )
I would say this
To anyone:
If you have the opportunity to buy 100 of each of those .. Then do it!
2SC4793, NPN
3SA1837, PNP
... they would walk miles around other drivers, like BD139/40 and MJE15030/31
lineup
lineup said:
One thing about these:
-> 2SC4793F
-> 2SA1837F
Anything with 'F' means: These are isolated versions. (insulated)
Which means you need no insulation to put them to heatsink.
Personally, I NEVER buy 'F' isolated versions.
They get hotter, as they can not radiate heat, as well as normal versions.
Technically, this means they can not take as much power. Without getting hot.
Using normal 2SC4793M and 2SA1837M is to prefer.
You can put elctrical isolation brick on, yourself, if you need.
and then, these versions will be a better choice.
At least when operated 100 mA -> 250 mA.
In my opinion, from reading datasheet, 2SC4793 and 2SA1837
have best and excellent performance when used at: 50 - 250 mA
NOS stock of these is still available, but there are comparable or better drivers at similar or lower prices:
Toshiba 2sa1930/2sc5171 (2A, 200 MHz !)
Toshiba 2sa965/2sc2235 (0.8A, 120 MHz, pre-driver)
Hitachi 2sb649/2sd669
NEC/Fairchild 2sa1220a/2sc2690a
The main advantage of the 2sa1837/2sc4793 pair is that there's a fairly realistic Spice model of it around. The 2sb649/2sd669 models are inaccurate and/or optimistic. I haven't seen reliable 2sa1930/2sc5171 models anywhere.
If anybody has successfully downloaded the Fairchild ksa1220a/ksc2690a models, please post them here - the email link at the Fairchild download page doesn't seem to work for me.
Toshiba 2sa1930/2sc5171 (2A, 200 MHz !)
Toshiba 2sa965/2sc2235 (0.8A, 120 MHz, pre-driver)
Hitachi 2sb649/2sd669
NEC/Fairchild 2sa1220a/2sc2690a
The main advantage of the 2sa1837/2sc4793 pair is that there's a fairly realistic Spice model of it around. The 2sb649/2sd669 models are inaccurate and/or optimistic. I haven't seen reliable 2sa1930/2sc5171 models anywhere.
If anybody has successfully downloaded the Fairchild ksa1220a/ksc2690a models, please post them here - the email link at the Fairchild download page doesn't seem to work for me.
Here we go, hope someone at least says thanks this time.
.MODEL KSA1220A PNP (BF=289.3 NF=1.0 BR=9.76 NR=1.006 IBE=4.7863E-13 IBC=4.7863E-13 ISE=5.2481E-12 NE=2 ISC=2.4909E-11 NC=1.5 VAF=98.5 VAR=6.7
+ IKF=2.7061 IKR=0.0759 RB=2.26 RBM=0.2308 IRB=0.001 RE=0.1908 RC=1.1748 QCO=0.02 RCO=3.9811 VO=11.078 GAMMA=5.01187E-8 CJE=3.4786E-10
+ VJE=0.9575 MJE=0.4694 FC=0.5 CJC=1.1224E-10 VJC=0.5761 MJC=0.4365 XCJC=0.4955 XTB=1.7978 EG=1.2255 XTI=3.0 TF=6.72889N)
.MODEL KSC2690A NPN (BF=132.5 NF=1.0 BR=8.495 NR=1.005 IBE=1.7783E-13 IBC=1.7783E-13 ISE=1.9953E-13 NE=1.5 ISC=1.5849E-9 NC=1.98 VAF=580.75
+ VAR=18.15 IKF=4.0271 IKR=0.0120 RB=2.98 RBM=0.001 IRB=0.6396 RE=0.0909 RC=1.4705 QCO=0.68 RCO=3.6239 VO=6.587 GAMMA=2.8216E-7 CJE=4.0082E-10
+ VJE=0.6696 MJE=0.3296 FC=0.5 CJC=6.0404E-11 VJC=0.5 MJC=0.4266 XCJC=0.4955 XTB=1.2590 EG=1.2277 XTI=3.0 TF=4.16856N)
Alex
.MODEL KSA1220A PNP (BF=289.3 NF=1.0 BR=9.76 NR=1.006 IBE=4.7863E-13 IBC=4.7863E-13 ISE=5.2481E-12 NE=2 ISC=2.4909E-11 NC=1.5 VAF=98.5 VAR=6.7
+ IKF=2.7061 IKR=0.0759 RB=2.26 RBM=0.2308 IRB=0.001 RE=0.1908 RC=1.1748 QCO=0.02 RCO=3.9811 VO=11.078 GAMMA=5.01187E-8 CJE=3.4786E-10
+ VJE=0.9575 MJE=0.4694 FC=0.5 CJC=1.1224E-10 VJC=0.5761 MJC=0.4365 XCJC=0.4955 XTB=1.7978 EG=1.2255 XTI=3.0 TF=6.72889N)
.MODEL KSC2690A NPN (BF=132.5 NF=1.0 BR=8.495 NR=1.005 IBE=1.7783E-13 IBC=1.7783E-13 ISE=1.9953E-13 NE=1.5 ISC=1.5849E-9 NC=1.98 VAF=580.75
+ VAR=18.15 IKF=4.0271 IKR=0.0120 RB=2.98 RBM=0.001 IRB=0.6396 RE=0.0909 RC=1.4705 QCO=0.68 RCO=3.6239 VO=6.587 GAMMA=2.8216E-7 CJE=4.0082E-10
+ VJE=0.6696 MJE=0.3296 FC=0.5 CJC=6.0404E-11 VJC=0.5 MJC=0.4266 XCJC=0.4955 XTB=1.2590 EG=1.2277 XTI=3.0 TF=4.16856N)
Alex
A lot of guys spend a lot of time and added circuit complexity designing their amps for better lineariality and better sound. I prefer to use finest transistors to achieve this instead of adding cascodes ect, ect, ect. Sanyo transistors have superior performance compared to any other manufacturer, although they are a bit more expensive.
Alex
Alex
homemodder said:A lot of guys spend a lot of time and added circuit complexity designing their amps for better lineariality and better sound. I prefer to use finest transistors to achieve this instead of adding cascodes ect, ect, ect. Sanyo transistors have superior performance compared to any other manufacturer, although they are a bit more expensive.
Alex
Hi Alex,
Cascodes do a great job in removing the Early effect, for example. You can't do the same with the "finest transistors". Every transistor has this f*king Early effect.
Cheers,
Edmond.
Hi Edmond
Yes i completely agree with you, but you should also see it this way, not everyone here wants a lot of complexity and strive or can afford for the ultimate complex design. By using very good parts performance can be enhanced this way. I have seen your designs, they are very good indeed and address most of amplifier shortcomings. I am scared to think how good performance you would achieve say in changing some of the parts you used in amps like the PGP. I see u have already used some outstanding parts there, like 2sa1407 but Ltp...........
Alex
Yes i completely agree with you, but you should also see it this way, not everyone here wants a lot of complexity and strive or can afford for the ultimate complex design. By using very good parts performance can be enhanced this way. I have seen your designs, they are very good indeed and address most of amplifier shortcomings. I am scared to think how good performance you would achieve say in changing some of the parts you used in amps like the PGP. I see u have already used some outstanding parts there, like 2sa1407 but Ltp...........
Alex
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.