transistors in the JC120

[Dimitree]

hello
I almost finished building a stand-alone Roland JC120 preamp..
I'm stuck with transistor substitution.
The original 2SK177-GR are out of production, I don't want to mess with possible fakes, NOS prices, etc.

So, can I safely substitute them with 2SK30A-GR in this application? I have lot of them and their equivalent in SMD format (I'd prefer SMD).



Also, do you think the output of this schematic is appropiate as "line output"?

[anotherjim]

Rightly or wrongly I like to think "line output" ought to be able to hit about 1v peak-peak into a 10k load.
Lose R29 & R32. Make C12 1uF (neg to out). Keep R33 100k as an anti-pop. That should easily drive a line input - but won't be tru polarity.
To drive the input of a guitar amp, the output from C10 looks suitable - and is tru polarity.

You may have to tweak your bias resistors (such as R3/R4) with different volts & jfets. Aim for close to half supply volts on every FET drain.

[Dimitree]

thank you
so, do you suggest adding an unity gain inverting opamp on the output to correct polarity and make a proper line output?

what about the jfets? are they simple followers, so any jfet would work in place of 2SK117-GR?

[digi2t]

How about 2SK170-GR? They're still available, I bought a bunch of through-hole versions recently. Check the datasheets.

[Dimitree]

thank you, but I just checked on mouser and they are obsolete too

[PRR]

> can I safely substitute them with 2SK30

"Safely"? It isn't going to explode, or bite your face. Not even smoke.

"As designed"? Can't say. 2SK30 specs, like most JFETs, vary ALL over the place. Even the graded GR spec is loose.

Roland did put those Gate bias voltages (the 1Meg:22K etc dividers) in to reduce the variation with different FETs. Is that enough? I suspect some 2SK30s won't bias-up as Roland expected. You can sort. You can diddle the 22K (33K, 27K) resistors to get the Drain voltages in a zone. What zone? Can you find Roland voltage-check specs? 10V-20V is a ballpark guess where it will "work OK", though maybe not "just like original".

> they are obsolete too

Cars are "obsolete" the day they are sold, but you can get parts for a 2002 Honda, a 1986 F-250, and even 1967 Cougar or 1930 Model A. You only need a handful, a lifetime supply.

[Dimitree]

> can I safely substitute them with 2SK30

"Safely"? It isn't going to explode, or bite your face. Not even smoke.

I meant safely to say that I can be sure that it would give almost the same results, since I don't have the original to compare. I'm sorry for my poor english

> they are obsolete too

Cars are "obsolete" the day they are sold, but you can get parts for a 2002 Honda, a 1986 F-250, and even 1967 Cougar or 1930 Model A. You only need a handful, a lifetime supply.

Mouser says obsolete, and that for me doesn't mean that there are better specs ones (since I don't mind about that), but means that I cannot get those transistors from they (or digikey or similar) but I have to rely on other sources (and I don't like to)

[anotherjim]

I'd breadboard the circuit from C7 to C10 in order to check biasing values. You can see from the test voltages it wants about x8 gain, and as usual, you want about half supply at the drain. You might have to be happy with drain volts anywhere between 1/3 & 2/3 of supply for any particular FET - but how consistent was the original production? If you want to be fussy, you can fit 100k trimmers in the lower position (R's 4, 15 & 25) of each FET bias, there's only 3 of them and for a DIY one-off, that's no big thing.

An inverting output stage to correct polarity would be similar, but with equal or near equal source & drain resistors for near unity from the drain.