100W Ultimate Fidelity Amplifier

I made a quick sim of A Class PSU to find out if the SK devices will work.
I used SK117, a low noise jfet counterpart for SK246 (no model available in Multisim lib). All 2N devices for the small TO-92s and TIPs for the output regulator. It seemed to work without a problem, trimmer adjust is set to 80%.
At 100% it is outputting 32vdc.

Terry, perhaps one or two of the SK device is failing, have you tried replacing them with newer ones?

Not sure what you mean by "newer ones". I don't have a big stock of the 2SK246.a I do have some SK117s. I can try those.
 
I mean fresh and unused ones, I was just thinking that probably a defect may exist on one of the installed jfets.
I built the other Apex Preamp PSU regulator (+-15vdc) and I also used 2sK246gr. The minus rail output voltage is slightly higher than the positive rail, but not sure though if the fets is causing this voltage imbalance, and I cannot tell if the fets that I used are genuine Toshibas. My knowledge is that a lot of these TO-92 fets out there in the market are no longer genuine.

You could try those SK117s it might fix the problem. Hopefully :)
 
To be honest, I almost cringe every time I see a circuit calling for jfets. I know a lot of folks like them but they are getting harder and harder to find. I'm not sure what they add to this circuit and wonder if the circuit could be made to work with BJT CCS. I keep a large supply of various BJT's as they can still be bought new in genuine form. A lot of Mile's circuits are likely old designs because they use obsolete parts. Most of the time I can find substitutes with current parts but jfets are a problem. To be honest the little shunt PSU seems to work great so I'm not sure this PSU is worth the trouble.
 
I do have some SK117s. I can try those.

Mind you, 2SK117 has a different pinout, you have to rotated them for 180° in order to match 2SK246 pinout. Also, 2SK117 is not 2SK246 direct replacement, but it may work. Try to find some 2SK30 in yours parts bin, these are much better choice for the CCS role.

I'm not sure what they add to this circuit and wonder if the circuit could be made to work with BJT CCS.

The circuit will work with any type of CCS, be it J-fet, M-fet of BJT as long as they are adjustable, since your voltage reference is a resistor (Vref=IxR)


To be honest the little shunt PSU seems to work great so I'm not sure this PSU is worth the trouble.

Definitively, tell me your desired output voltage and load current and I will set your shunt reg with 1 zener diode and few resistors:)
 

Attachments

  • IMG_20160101_132348 (Copy).jpg
    IMG_20160101_132348 (Copy).jpg
    302.5 KB · Views: 1,411
Mind you, 2SK117 has a different pinout, you have to rotated them for 180° in order to match 2SK246 pinout. Also, 2SK117 is not 2SK246 direct replacement, but it may work. Try to find some 2SK30 in yours parts bin, these are much better choice for the CCS role.



The circuit will work with any type of CCS, be it J-fet, M-fet of BJT as long as they are adjustable, since your voltage reference is a resistor (Vref=IxR)




Definitively, tell me your desired output voltage and load current and I will set your shunt reg with 1 zener diode and few resistors:)

OK, so this brings up something I have been wondering about. Is there a way to determine the pinout of a jfet? I have been looking through data sheets and it seems some manufacturers used different pinouts for the same part numbers. My lcd meter only really determines the gate. Also, the schematic symbol doesn't depict which pin is source and which is drain. How do you tell which is which?

Thanks, Terry
 
OK, so this brings up something I have been wondering about. Is there a way to determine the pinout of a jfet? I have been looking through data sheets and it seems some manufacturers used different pinouts for the same part numbers. My lcd meter only really determines the gate. Also, the schematic symbol doesn't depict which pin is source and which is drain. How do you tell which is which?

Thanks, Terry
Hello Terry!

I use this module and you can find chinese clone on Ebay.
https://www.youtube.com/watch?v=NCSb3vqkbqA

:)
 
Some jFETs, particularly those intended for RF duty are symmetrical. With these, I am told that Drain and Source are interchangable. A few datasheets state this.
I have tested quite a few non RF jFETs and find that swapping D & S makes very little difference, they are not identical, but very close.

The ATmega328 is a great little (approximate) tool.
very good for identifying devices and leadouts.
NOT good at hFE and Vbe and all the other parameters that we sometimes need fairly accurately.

Worth having among your testing tools @ ~£10 ($15)
 
Last edited:
I have this meter. LCD Transistor Tester Diode Triode Capacitance ESR Meter mos PNP NPN LCR SZTM | eBay

It actually does a surprisingly good job at most things but jfet fool it. It will always find the gate but can't determine the source from the drain. I have also read that they can be inverted, though here some have indicated that in this circuit that may not be the case. I had thought to try inverting them just to see. I don't remember where I got these. I bought them for an older project. IIRC they worked fine.

The center pin is the gate so inverting them should be easy. Maybe I'll give that a try today just for kicks.
 
Turn it over and read the number on the big chip. Does it say ATmega 328?

Now try measuring a jFET.
Swap it around the various holes in the zero insertion force device holder.
It should consistently tell you it is a jFET and it should consistently identify the gate lead, even though you have moved it from 1 to 2 to 3.

It actually does a surprisingly good job
I don't agree.
I reported some months ago.
It consistently tells me that dozens of different BJT devices have the same hFE and the same Vbe.
Yet, when I test them manually, noting Ib & Ie I can see that hFE is different, sometimes by an enormous amount. In hFE respect it is worse than the DMM hFE test.
 
Last edited:
Turn it over and read the number on the big chip. Does it say ATmega 328?

Now try measuring a jFET.
Swap it around the various holes in the zero insertion force device holder.
It should consistently tell you it is a jFET and it should consistently identify the gate lead, even though you have moved it from 1 to 2 to 3.

I don't agree.
I reported some months ago.
It consistently tells me that dozens of different BJT devices have the same hFE and the same Vbe.
Yet, when I test them manually, noting Ib & Ie I can see that hFE is different, sometimes by an enormous amount. In hFE respect it is worse than the DMM hFE test.

As I said, it does a good job at most things. I find it agrees with my DMM fairly close I was surprised at how well it matched mosfets. I had a box full of irfp's that I had matched using a jig and it matched up with the numbers I had gotten to a tee. I was pleasantly surprised.

I have found reading hfe to be tricky anyway because temperature of the device changes it so drastically. I have often wondered how oem manufactures deal with that when they are building so many units. I wonder how close a tolerance they work with.
 
....................
I have found reading hfe to be tricky anyway because temperature of the device changes it so drastically. I have often wondered how oem manufactures deal with that when they are building so many units. I wonder how close a tolerance they work with.
They either buy in duals and charge the customer for the priviledge , or they don't design for needing close matching.
Very few actually match themselves, but some will match. A tweak of a pot on the testing line is about as close as they get.
Speaker manufacturers went over to computerised measuring & matching many decades ago. I suppose some could be doing similar with semiconductors.
IC are now laser trimmed to meet guaranteed performance targets. That is a form of matching. Resistors too.
 
They either buy in duals and charge the customer for the priviledge , or they don't design for needing close matching.
Very few actually match themselves, but some will match. A tweak of a pot on the testing line is about as close as they get.
Speaker manufacturers went over to computerised measuring & matching many decades ago. I suppose some could be doing similar with semiconductors.
IC are now laser trimmed to meet guaranteed performance targets. That is a form of matching. Resistors too.

I think the problem with these Ebay devices and most voltmeters is they only test at one preselected current flow. I'd like to design one that will test at a selected current that will match what is used in circuit.
 
I made a quick sim of A Class PSU to find out if the SK devices will work.
I used SK117, a low noise jfet counterpart for SK246 (no model available in Multisim lib). All 2N devices for the small TO-92s and TIPs for the output regulator. It seemed to work without a problem, trimmer adjust is set to 80%.
At 100% it is outputting 32vdc.

Terry, perhaps one or two of the SK device is failing, have you tried replacing them with newer ones?

Can we use BF244 instead BF245 they are very similar.
3 X BF244A N Channel RF Amplifier Transistor Free Shipping | eBay
Very likely if has the same pin out..
 

Attachments

  • psu.jpg
    psu.jpg
    148.3 KB · Views: 1,169