Rinat

I've noticed that in later versions of the PSU board resistors R1, R2, R63, R16-17 and R64 have been changed from 400k to 300k. I wonder about the reason. 300k resistors would be hot: (475^2)*2/600000 is about 0.75W or 0.37W for every 0.5W-rated resistor. Who tried 300k or 400k ?

Are they from the same batch?

It seems that current is limited to 0.20 ma. Look at the voltage across the 50k R12 resistor. It is 10v - Vbe(0v min). 10/50000 = 0.00020. So, it looks like 2sc3675 will work perfectly here.

I've asked @ibuskiin PM and no - I've bought these transistors from another and more reliable source, the current official dealer of Onsemi. I've measured them with DY294 and tried to build curves up to 750v. Everything is ok. I've got them in a factory package with stamps, date codes, etc. Before I've made a payment they said that these transistors have some discoloration due to long storage. But actually, they look better than the ones I got from bdent.com 🙃

Ok, thanks. I'll calculate if the replacement could be done.

I feel for you and could only imagine how it feels to see such a lightshow ( 4. I would check if everything is ok with Q24/Q25, just in case. And a question about your version of PSU: could I replace KSC5026 with 2SC3675 ? I have hundreds of the latter. And there is nowhere to buy the first ones(. I did not make calculations yet.

Thanks again, I am going to buy this DE-5000. I've heard many times that it's accuracy in a lucky cases can be better than in specs. It gives a hope 🙂

@jamesmking, thanks for another awesome post. I remember that you made resistors/capacitors selection to reduce the tolerance. Did you use a LCR meter ? And which one ?

Yea, I've bought a few parts for T2 from arrow. They have not a great customer support for individuals, but all parts are genuine.

Thermal resistance 100k/w instead 80k/w..... I think it can be used, but I would make an accurate calculations before. And you will need a board with a full metal polygon and a way to mount it on heatsink. It is gonna to be cumbersome. Anyway I can't find IXCY too.

Lower capacitance is better. And in this context I want to ask another question to all: I want to use AlN thermal pads for the transistors with 1mm thikness instead 1.6mm AlO. It adds a few picofarads to a capacitance between a metal tab and the ground. Is it ok ?

I have a bunch of 2sk214 from bdent and their Vbr is measured as 220-240v by DY294. This is sudden for me, because I have the same result with 2sk216s. But I still afraid to use them as a replacenment because in other conditions, like temperature, Vbr could be lower than it was measured by DY294. But maybe I'm just too scared to burn something expensive 🙂

I know a man who replaced EL34 by KT77 successfully, with slighly changed sound. These tubes was not from JJ and I know that JJ's KT-77 are not suited for T2. I would'nt be so adamant as Kevin - In my opinion you can change anything, but only with a really careful recalculation of the scheme. If you are not sure that you can handle it yourself, don't do it. I am going to try 6N23P-EV, but only because it is a direct replacenment for 6922 and I've measured every tube with good results. And am going to check all again, maybe even twice.

I've measured 62s of mine 2sa1413 from bdent: https://docs.google.com/spreadsheets/d/1sadOLdhxStUQfL94wJkuKEtn7qOEWFBdP0m-BNzwVw0/edit?usp=sharing It is interesting that green transistors were worse than black. And thanks to Helium - I've found that 2sa1413 has a twisted pinout - BCE instead of ECB for 2sa1486.

Oh, I thought that TO-251 = D-PAK. Thanks

They have 25 pcs in stock, while the original scheme requires 26 pcs ) Personally, I would suggest to use 2sa1413-z if one wants to strictly follow the original scheme: it is the same transistor in TO-252 case (bdent says that it is D-PAK, but they have sent to me TO-252 version, which is even better for me). https://www.renesas.com/us/en/document/dst/2sa1413-z-data-sheet?language=ja But I think that STN9360 is even better option than 1486/1413....

DC current gain looks too high

I think about one of these connectors: https://www.ges-highvoltage.com/us/high-voltage-connectors-hv-connectors-multi-pin.html

I am going to just make a CCS using IXTP08N100D2. "8/9/2022" is too far for me to wait.

Digikey suggests to replace MJW21194G/MJW21193G with MJW21196G/ MJW21195G. The only difference I've found in datasheets is lower saturation voltage.

No, thanks for the option! But MBT6429DW1T1G is cheaper (It is in stock at arrow.com), has lower capacitance, higher hfe at 50nA (it is a current in the active battery), similar noise. BCV61 has matched halves. I will think, thanks. So - do you agree that the high breakdown voltage is not required here ?

BTW: A shipping date for IXCP10M90S on mouser was moved to August 2022 ((( I would try to use IXTP08N100D2 (

These ones for example: https://www.mouser.com/datasheet/2/308/1/MBT6429DW1T1_D-2315127.pdf https://www.onsemi.com/pdf/datasheet/bc846bdw1t1-d.pdf https://www.analog.com/media/en/technical-documentation/data-sheets/MAT01.pdf https://toshiba.semicon-storage.com/info/docget.jsp?did=19156&prodName=HN1C03FU https://www.digikey.com/en/products/detail/linear-integrated-systems-inc/IT120-PDIP-8L/13688035 https://fscdn.rohm.com/en/products/databook/datasheet/discrete/transistor/bipolar/emx26.pdf I would try the first one or MAT01AH if I were sure that I don't need > 80v of Vce at the battery ( And I am not sure that we really need here low output capacitance. If no then there are another nice options. But I don't think that using separate single transistors is a good idea for current mirror due to temperature drift. I prefer one crystal or at least one package.

A question for anyone who calculated the amplifier - what is the point in using high-voltage transistors in current mirror in the active battery ? The current through the each transistor is ~0.05 mA, Ur41 and Ur40 ~ 5 mV. It means that Uce of the right part of 2sc3381 should not be more than ~5v (It is the max Ube of 2sc3675). And the same is for the left part: Ube is 5V max -> Uce max = 5v. So - what is the point ? There are nice double transistors with lower maximum Uce, with the same capacitance, with higher hfe etc....

A few thoughts about replacing transistors to a contemporary parts: 1. I thought that stn9360 is just an ersatz for 2sa1486, but it seems that it is even better than the original one. 9360 has mush higher hfe and a little lower capacitance. It means that it is more suitable for the active battery and CCSs 2. KSA1220A has lower Vbr (160v instead of 200v for 2sk216) but it is ok even in original scheme. Does it mean that 2sk214 is safe to use ? I also afraid to get lower sound quality because of poor linearity of this mosfet (. 2sk216 is much better in this term. So, I am trying to find better substitution 🙂 . I don't like to use obsolete parts. It makes me slightly nervous. Now I am going to use 216/79 but I want to build someday a "fully-contemporary parts" T2 🙂. 3. 2SC3675: FJPF2145 has lower hfe and Vbr (. I have enough 3675s but also want to substitute them in future. Maybe cascoded CCS using JFET + DMOS ? It has much better resistance(look at the photos from the book) than single-bjt CCS. Cascoded CCS measured resistance Single BJT CCS measured resistance 4) 2SK246 -> 2SK208 5) 2SC3381 -> HN4C5J . What about matching of halves ? There is no info about max difference in datasheet. Also I want to build the amp with replaceable active battery - soldered to the main board in 2 points. I don't trust mechanical connection. UPD: It looks like KSA1220A is obsolete now (official status - "last shipments")