Analog oscilloscope repair / JFET confusion

[Tokipoki]

Hi!

I need some help with a faulty analog oscilloscope. Long story short, CH2 was working as intended, CH1 was dead. After poking around for a few minutes, I found the issue, it was the JFET Q101 at the input stage, it is a 2SK55D.
(Out of curiosity I moved Q201 from CH2 to CH1, and voilà, CH1 works, so everything's fine except the original JFET.)

Here is the issue: I need a replacement FET, but the one I bought doesn't seem to work. Sadly, at the store I usually buy my stuff this exact part wasn't available, and they only had like 4-5 different JFETs, and that's it, so I checked the datasheets to find something similar.

I chose the BF256B.
Basically CH1 doesn't work with the new FET at all: AC/DC switch does nothing, GND does switch nothing, I can't get that 1kHz square wave to show up, there is nothing on the screen, except at certain vertical and horizontal settings I get a really faint, wide vertical green blob moving sideways.


Am I missing something? Are those replacement FETs no good, or is it some calibration issue, or something like that?
I attached the two datasheets and the related part of the circuit. (Edit: I wasn't paying attention, so the schematic I uploaded is for the second channel, but they are identical, except the component numbers start with a "1" instead of a "2", so in this case Q201 is the part I am talking about.)

[Kim Christensen]

I think you should have chosen the BF256A instead because the IDSS spec of 3-7ma is a match to the 2SK55D IDSS spec of 3-7ma... Whereas the BF256B has a IDSS spec of 6-13ma.

I have to say that those two datasheets are a bit confusing with their typos in their IDSS specs:
2SK55 sheet shows Vds = 10A whereas I'm sure they mean Vds = 10V
BF256 sheet shows Vgs = 15V, Vgs = 0 ( They probably meant Vds = 15V, Vgs = 0 )

Note: You may be able to adjust it so the B version works, but you'd need to look in the manual for that procedure. Mark the adjustments with a felt pen before starting so you can get things back easier if you mess it up.

[MathWizard]

I'd put the fets on a breadboard with a couple of resistors and with a PSU and DMM, see what the JFET's are really like. They may just need a little mod to the original circuit to get working somewhat or fully even.

IDK those circuits good at a glance, but yeah it still could need re-calibration, and who knows if it would work the same over the whole operating range.

But surely not a lost cause just over a JFET

But yeah, what was was this scope, what year ?And people were probably solving by hand how to bias these things. So what year of EE do you learn to make complex multi-level potential ladders all over the place to fir together just right at DC and AC

[Tokipoki]

@Kim Christensen
Yeah I noticed that difference, but they only had the "B" version available. Is the parameter I_DSS critical?

@MathWizard
It is a Goldstar OS-9100D, probably from the early 90s. Also Goldstar is the old name of LG, and as far as i know they changed the brand name around 1995, so they made this a little bit earlier.



To be honest I didn't do a deep dive into the circuitry or the service/operation manuals, since I expected it to be "plug and play". And I didn't had many options so I just grabbed a fet that was reasonably close... Seems like it's not that easy.

I will probably order the proper replacement parts... The reason I didn't do that originally, because my only options are ali or ebay, and at this point of the year it will be like a 1 month wait. I guess in the meantime I will look up some FET theory  and I do some measurements to compare the two channels

[Kleinstein]

The Idss itself is usually not directly critical. It is more the gate source voltage for the intended operating current that matters. However the 2 parameters are correlated. So a high Idss also means a higher threshold. With somewhat different FET types there can still be a difference.  The BF256 in principle is a reasonable substitute and not a bad choice and not that much difference expected.
The circuit looks like it can compensate for a certain range of threshold, but the range is limited.
Even within the groups there is still quite some variation in the FETs. It may be worth checking the threshold range of the JFET - is also a crude function test, in case one has a bad one or wrong pin-out.

The test could be something like gate and source to ground with something like 100 K and drain to some 12 V.  The measured gate voltage should be near zero (the resistor is mainly for protection in case of a wrong pinout). The source voltage gives an idea on the threshold (depending on the resistor a bit lower, but OK for matching / comparing types).
Especially with parts from ebay a functional test can be a good idea - to  make shure it is at least a JFET and the pin-out is correct.

[TimFox]

I_DSS can be a very important parameter by itself, since most JFET linear amplifier circuits are biased to a given value of drain-source current.  If that current value is greater than I_DSS of the replacement FET, then the GD diode will tend to forward bias.  That's why it is very useful that many JFETs are available sorted into bins of limited I_DSS range.

[Wallace Gasiewicz]

Here is a link listing so,e similar FETs to the 2SK55

https://vetco.net/products/n-channel-jfet-transistor-vhf-amp-mixer-30v-50ma-nte312

Notice that this is a list of FETs that the NTE 312 replaces
2SK55 is on the list
but the BF 256 is not on this list (other BF trans are on the list)

So I would pick one on the list and give it a chance, most of these should be cheap.
Or just go with NTE 312
Some of these FETs might have different pinout however.

[Kim Christensen]

I_DSS can be a very important parameter by itself, since most JFET linear amplifier circuits are biased to a given value of drain-source current.  If that current value is greater than I_DSS of the replacement FET, then the GD diode will tend to forward bias.  That's why it is very useful that many JFETs are available sorted into bins of limited I_DSS range.

Hmmm... Looking at the OP's schematic again, it does look like Q204 is set up as a constant current source sink of 3.85ma (Assuming no influence from the "Op-Amp")
Looks like the "OP-amp" is in the DC negative feedback path and adjusts the current in Q204 to keep it in balance.

[srb1954]

I_DSS can be a very important parameter by itself, since most JFET linear amplifier circuits are biased to a given value of drain-source current.  If that current value is greater than I_DSS of the replacement FET, then the GD diode will tend to forward bias.  That's why it is very useful that many JFETs are available sorted into bins of limited I_DSS range.

Hmmm... Looking at the OP's schematic again, it does look like Q204 is set up as a constant current source of 3.85ma (Assuming no influence from the "Op-Amp")
Looks like the "OP-amp" is in the DC negative feedback path and adjusts the current in Q204 to keep it in balance.

And the op amp has a fairly limited range of adjustment of that constant current to adjust for the FET's Idss variuation. A rough back-of-the envelope calculation indicates that the maximum the constant current could be adjusted to is about 6mA, more or less consistent with the Idss range of the 2SK55. 

[Kleinstein]

Here is a link listing so,e similar FETs to the 2SK55

https://vetco.net/products/n-channel-jfet-transistor-vhf-amp-mixer-30v-50ma-nte312

Notice that this is a list of FETs that the NTE 312 replaces
2SK55 is on the list
but the BF 256 is not on this list (other BF trans are on the list)

So I would pick one on the list and give it a chance, most of these should be cheap.
Or just go with NTE 312
Some of these FETs might have different pinout however.

That list includes quite difference JFETs, like 2N4416 and 2N4117 (considerably smaller). The 2N4416 is actually very close to the BF256: in some datasheets of the BF256 they refer to the 4416 for curves.

The match of an BF256a to the SK55 is not perfect, but also not that bad. The choice of JFET in a TO92 or similar case is relatively limited now.

Chances are the FET is used with a current a little below Idss. With the 100 ohm resistor at the source one would have some -600 mV  gate to drain, which makes absolute senes, as in this range the TC of JFETs goes through zero and thus low drift. If not too far off one may be able to adjust a resistor to make it work even with a BF256B. I would still check the FET first, if it is about in the right range. There is also quite some variations inside the groups and on occasion parts from Alibaba turn out to be way out of specs.

p.s. I looks at the DS of the NTE part: it is not a good substitute, with nearly 2x the capacitance and a large range of Idss.
 

[Wallace Gasiewicz]

Is your new jFET really a JFET
I received come counterfeit ones.
I bought some from China that were really NPN transistors with the JFET number on the plastic package.