For example if my JFETS are j113 & j176 have Idss about 35mA. So now How can I reduce it 8ma to 10ma bias with paralleling or in series. I read about cascoding article and honestly I don't have that much basic knowledge to process that information in that article.
thanks.
thanks.
majority of investment in amplifier is going to heavy hardware - xformer , caps , case with heatsinks ;
regarding how big decrease in performance and how small amount of saved greenies , if using any other JFets than prescribed , it's simply not wise to proceed with anything else than real thing
even taking in account that for someone obtaining them is complicated , not just or nothing at all matter of money
If you can't manage to have proper JFets , build some another amp , where you can substitute them much easily ....... or some other amp without JFets
regarding how big decrease in performance and how small amount of saved greenies , if using any other JFets than prescribed , it's simply not wise to proceed with anything else than real thing
even taking in account that for someone obtaining them is complicated , not just or nothing at all matter of money
If you can't manage to have proper JFets , build some another amp , where you can substitute them much easily ....... or some other amp without JFets
Yeah right. But some newbies want economy build to dirt their hands for first time builds. Cost is too high after shipping and customs for linear systems jfets. May be Mr. Nelson or someone has some secret circuit for guys like me. If I ask only then Members will provide different ways. I have another project PLH in mind because of absolute parts of Toshiba. any exicon lateral or irf available replacement for irf610 to build PLH ?.
Thank you.
Last edited:
build Babelfish J - using J271 for input LTP and pretty much any IRFP N- channel for output
you can make it even on veroboard or PtP ...... or you can make your own pcbs from pdfs posted in first (old ) thread
instead of IRF610 you can pretty much use any of these - IRF510 , 520, 530 , 610,620,630
higher the number , higher voltage and higher capacitances
you can make it even on veroboard or PtP ...... or you can make your own pcbs from pdfs posted in first (old ) thread
instead of IRF610 you can pretty much use any of these - IRF510 , 520, 530 , 610,620,630
higher the number , higher voltage and higher capacitances
I suspect there is a typo above, and Nelson actually means the J176 P-channel JFET whose datasheet (attached here) specifies
2.0 mA <= |Idss| <= 25 mA
The J177 has even lower Idss, but DigiKey doesn't sell it as a thru-hole part, only surface mount. Fortunately the DIY-focused vendor "Linear Systems," does sell it in thru-hole.
_
J176 has higher dissipation rate than J177(260mw). J113 & J176 have around 360mw power dissipation.
Actually, measuring my extensive population of these parts,
probably should have said J175.
The J113's come in at around 20-40 mA.
The J175's are 25-30, J176 are 8-15, and P1086 are 20-30.
Regardless of the spec in the old datasheet, this is what I'm seeing.
Getting matches for these is not easy, and for the F5 adjustment of Source
and Drain resistors will be necessary to get a balance.
so Nelson the J175 and J113 are the right combination? Still I agree matching is required.
What should be the source resistance change for the F5 for these?
It might be wise to (i) acquire the devices; (ii) plot histograms of measured IdssP and IdssN on the devices you actually own; (iii) find a pair that looks acceptable; and THEN, once you know the Idss you're actually going to use, (iv) choose the value of R1-R2 + R5-R8 which give your desired bias point and your desired transconductance.
In the population on my shelf, they are the most likely to have the same
Idss. They will still have slightly different transconductance, and also not
the same actual curves. This is where variable Source and Drain resistance
comes in, but nothing is ever perfect, not even with 2SJ74 and 2SK170.
For the F5 you want Ids in the region of maybe 8 to 12 mA.
Mark's answer is so obvious that it hardly seems necessary.
But I have to ask:
Why is someone who can't design trying to redesign this amplifier?
Accept as is, or find one that is good enough as is, until one has learned how amplifiers work and then progress to learning how to design amplifiers.
I have never progressed beyond how amplifiers work and find I still get stumped (cricket term) by unusual topologies. I will never become an amplifier designer.
But I have to ask:
Why is someone who can't design trying to redesign this amplifier?
Accept as is, or find one that is good enough as is, until one has learned how amplifiers work and then progress to learning how to design amplifiers.
I have never progressed beyond how amplifiers work and find I still get stumped (cricket term) by unusual topologies. I will never become an amplifier designer.
Because he does not want to pay for 2SK170/2SJ74.
But if you look at the total cost of a Class A power amp, the JFET cost is <10%, even with today's premium price.
And it is such a compromise to use anything else.
So might just as well build something totally with BJTs.
Patrick
But if you look at the total cost of a Class A power amp, the JFET cost is <10%, even with today's premium price.
And it is such a compromise to use anything else.
So might just as well build something totally with BJTs.
Patrick
In the F5 Heamamp thread, you can find how the basic F5 topology can be adapted to use different frontend devices,
including BJTs, 2SK246/2SJ03, and 2N5457/2N5460 :
F5 Headamp ?
Posts #1,2,5,7,25
It is of course possible to do the same for the power amp.
But it will just not have the same performance.
Patrick
including BJTs, 2SK246/2SJ03, and 2N5457/2N5460 :
F5 Headamp ?
Posts #1,2,5,7,25
It is of course possible to do the same for the power amp.
But it will just not have the same performance.
Patrick
There are many Members who won't listen to the advice they don't like.
But worse, there are Members who won't read advice!
But worse, there are Members who won't read advice!
Well, if they have the skill to design, or at least to simulate by trail & error, I have given enough hints above.
Then they can decide whether it is worth saving 40 USD for less performance.
It's a free world. They have free choice. 😉
Patrick
Then they can decide whether it is worth saving 40 USD for less performance.
It's a free world. They have free choice. 😉
Patrick
Of course, the Toshiba or Linear System parts are ideal, but hey, it's DIY. Why not try other things I ask.
I'd go with MMBF5459 and 5462, as EUVL hinted (or 2N5459/5462 if you can find them). You can get 2 x 100 pcs each for a total of 30 Dollars at Mouser. Yes, these parts are not through hole, but SOT23 is BIG.
Drawback is the max voltage of the N Channel device, max. 25V.
If it's an issue (what Rails do F5 have?) cascoding is needed, or alternatively a zener (eg 4V7) at the drain, to eat up a bit of voltage.
It's DIY, it's fun, and if you can't get it to work, get those LSK parts.
I'd go with MMBF5459 and 5462, as EUVL hinted (or 2N5459/5462 if you can find them). You can get 2 x 100 pcs each for a total of 30 Dollars at Mouser. Yes, these parts are not through hole, but SOT23 is BIG.
Drawback is the max voltage of the N Channel device, max. 25V.
If it's an issue (what Rails do F5 have?) cascoding is needed, or alternatively a zener (eg 4V7) at the drain, to eat up a bit of voltage.
It's DIY, it's fun, and if you can't get it to work, get those LSK parts.