Biasing JFETs in musikding katana kit

[jhob]

I've built a musikding katana pedal and used 20k trim pots for the JFET biasing.  However I can't get the voltage to 9v, most they go to is 8.6v.  Would switching these out for 50k trimmers give me more scope for adjusting the voltage?

Schematic: https://www.musikding.de/docs/musikding/katana/katana_schalt.pdf

I don't fully understand where, but I presume part of the circuit must be increased somewhere given that the input voltage is 9v.

I'm a relative newbie to pedal building so going through the basics is beneficial to me.

cheers!

[Clint Eastwood]

What is your actual V2 voltage? The schematic uses 10k pots, 20k should surely work with the right V2 voltage. I assume the ICL7660s is a voltage doubler making 18 volts for V2, so check that first.

[FSFX]

Typically a normal 2N5458 with an 18v supply and a 680 ohm source resistor should bias to about 9v on its drain with a 5.6k drain resistor, so a 10k pot should be fine if adjusted about mid-way.
If you cannot adjust the bias correctly then I would check all of the component for correct value, check the 2N5458s are correctly orientated and that you are supplying about 18v from the charge pump.
Where did you get the 2N5458 JFETs from ?? Are you sure that they are legit devices ?? 

[Rob Strand]

You might have a build problem but it's highly likely the JFETs are not real 2N5458's.
With real 2N5458's the drain resistor should be around 6k to 7k (I get similar values to FSFX).

If you keep increasing the drain resistor it might bias OK but then the circuit won't behave correctly:
Normally you have 6k to 7k driving the output caps (1n and 15n) when you increase the pot to 20k or more
the filter cut-off formed by the caps will drop by a factor of 3 or 4 making the pedal muddy - certainly not
as intended.

Another thing that will probably happen is the gain of the circuit will end up higher than the original circuit.

Ideally you would want to measure the JFET parameters,
https://www.runoffgroove.com/fetzervalve.html

If the JFETs aren't real 2N5458's it's likely the VP will be a low value like 0.3V instead of around 2V.   Also it's possible the IDSS parameter is a low value like 0.6mA instead of 4mA or so.  Alternatively it could be a high IDSS.

Suppose you do find the JFETs like this.  You will need to do a lot of mods to make the circuit behave like the original.

This idea can help bias the JFETs and preserve the gain,
https://www.diystompboxes.com/smfforum/index.php?topic=130258.msg1263261#msg1263261

Then you will need to decrease the 15n and 1n caps.

Another thing to tweak is the high-gain switch which shorts the 10uF cap across the 680 ohm source resistor.  Off hand you will need to add resistor in series with the cap to get the same gain as the original.

Lots of mods.

So,
- first check your parts
- measure your JFETS
- if you find your JFETs are not 2N5458's then either,
    - find new JFETs, or,
    - do a heap of mods, or,
    - keep increasing the drain resistor and put up with the fact it's not like the original (maybe decrease the 1n, 15n caps)

[jhob]

What is your actual V2 voltage? The schematic uses 10k pots, 20k should surely work with the right V2 voltage. I assume the ICL7660s is a voltage doubler making 18 volts for V2, so check that first.

If purchased as a kit it comes with 20k pots.

I did wonder what the ICL7660s was doing, thanks for pointing that out.  I'll test the pins on the ICL to see if one of them is giving 18v.

I'm unclear on where I would test v2 in the circuit, are you able to advise?

[jhob]

Typically a normal 2N5458 with an 18v supply and a 680 ohm source resistor should bias to about 9v on its drain with a 5.6k drain resistor, so a 10k pot should be fine if adjusted about mid-way.
If you cannot adjust the bias correctly then I would check all of the component for correct value, check the 2N5458s are correctly orientated and that you are supplying about 18v from the charge pump.
Where did you get the 2N5458 JFETs from ?? Are you sure that they are legit devices ??

I've been testing the voltage between 1 & 3 (or might be 2) on the trim pots.  Is that the right place to be testing it?   

Testing voltages in a circuit and making some meaning of them is fairly new to me.

Are you referring to teh ICL7660s as the 'charge pump'?

I bought the 2N5458s from bitsbox in the UK - https://www.bitsboxuk.com/index.php?main_page=product_info&cPath=140_151_154&products_id=1075

I didn't realise it was possible to get non-legit parts, what's the story with that?  I think I've got one other 2N5458 from musikding (only bought one cos I didn't read the BOM properly  ) that I could try which might confirm if it's different.  How would I tell?

I have them orientated the correct way as indicated on the PCB, but will try putting them in the other way and see if that makes a diffence.

[jhob]

You might have a build problem but it's highly likely the JFETs are not real 2N5458's.
With real 2N5458's the drain resistor should be around 6k to 7k (I get similar values to FSFX).

If you keep increasing the drain resistor it might bias OK but then the circuit won't behave correctly:
Normally you have 6k to 7k driving the output caps (1n and 15n) when you increase the pot to 20k or more
the filter cut-off formed by the caps will drop by a factor of 3 or 4 making the pedal muddy - certainly not
as intended.

Another thing that will probably happen is the gain of the circuit will end up higher than the original circuit.

Ideally you would want to measure the JFET parameters,
https://www.runoffgroove.com/fetzervalve.html

If the JFETs aren't real 2N5458's it's likely the VP will be a low value like 0.3V instead of around 2V.   Also it's possible the IDSS parameter is a low value like 0.6mA instead of 4mA or so.  Alternatively it could be a high IDSS.

Suppose you do find the JFETs like this.  You will need to do a lot of mods to make the circuit behave like the original.

This idea can help bias the JFETs and preserve the gain,
https://www.diystompboxes.com/smfforum/index.php?topic=130258.msg1263261#msg1263261

Then you will need to decrease the 15n and 1n caps.

Another thing to tweak is the high-gain switch which shorts the 10uF cap across the 680 ohm source resistor.  Off hand you will need to add resistor in series with the cap to get the same gain as the original.

Lots of mods.

So,
- first check your parts
- measure your JFETS
- if you find your JFETs are not 2N5458's then either,
    - find new JFETs, or,
    - do a heap of mods, or,
    - keep increasing the drain resistor and put up with the fact it's not like the original (maybe decrease the 1n, 15n caps)

Thanks for the detailed reply, I'll go through it all when I'm in front of the circuit later.  I've got this multimeter which I think can measure transistors, seem to recall I used it for that with some germanium transistors a few years ago.

Not made any pedals for a few years so relearning a lot of stuff!

What's IDSS?

[Clint Eastwood]

What is your actual V2 voltage? The schematic uses 10k pots, 20k should surely work with the right V2 voltage. I assume the ICL7660s is a voltage doubler making 18 volts for V2, so check that first.

If purchased as a kit it comes with 20k pots.

I did wonder what the ICL7660s was doing, thanks for pointing that out.  I'll test the pins on the ICL to see if one of them is giving 18v.

I'm unclear on where I would test v2 in the circuit, are you able to advise?

You measure V2 from the top of P1 and P2, the side that is not connected to the jfet drain, to ground. It might be that there is only 9 volts there, where it should be 18v.

[FSFX]

What's IDSS?

Idss is the drain saturation current. Vgs(off) is the gate cutoff voltage, otherwise known as the pinch-off voltage (Vp).

Those two parameters define the two limits of the normal operating range for a JFET as shown in this diagram for the 2N5457.
You need to bias the JFET to operate at some point between the pinch off voltage and zero.
The drain current will then be at a corresponding quiescent current between zero and Idss such as shown for the three examples shown on the graph.

[duck_arse]

I have them orientated the correct way as indicated on the PCB, but will try putting them in the other way and see if that makes a diffence.

never do this. the odds are always against you. why kill good parts? get the datasheet for the parts you have at hand, they will have pinout drawings so no errors. if you don't follow anything on the datasheet, or the board markings, come here and ask.

[jhob]

So looks like it's the 2n5458s that are the problem.  Popped in the one I got from musikding (long legs in photo) and it biases right where it should at about 10k.

I never even knew non-legot components were a thing, what's the story there?

Can anyone suggest a UK supplier I could get some legit 2n5458s from?



Thanks for the help with with one, never would have worked that one out by myself!

[idy]

Non legit parts are a thing, JFETS are among the most seen; they are no longer made in "through hole" models, so you're getting old parts, or parts re-labeled. Sometimes re-labeled parts are "close enough" for some circuits. Sometimes they are not.

There is some kind of industry, especially in China, of "blacktopping" ICs and relabeling them as something pricier. So it is buyer beware when you start shopping at random vendors...

[jhob]

Seems mad to me as they're not exactly high value parts.

I've always bought my components from bitsbox and they've been great, this is the first incident I've had with them.  I think I'd consider them a legit supplier unless someone wants to tell me otherwise.

[FSFX]

Bitsbox are normally very reliable but with many of these out of production JFETs like the 2N5457 and 2N5458 the supply from authorised distributors has dried up.

There are still devices like the J113 which often can be used and is still manufactured as both through hole and surface mount versions.
Reliable authorised distributors like Farnell still seem to have quite large stocks of them and they seem good.
I recently bought a few hundred of both the THT and SMD types and they measure fine.
As with all JFETs, they are not as easily interchangeable as many bipolar transistors are and as Rob says above, using different JFETs often requires changes to the resistor values to get them biased correctly.

You will see a lot of fake JFETs that are marked with an 'F' logo. I suppose that the fakers think people will believe them to be Fairchild devices. However the Fairchild 'F' logo has a strict form to it as shown in this diagram. That doesn't mean fakers can't fake the logo closer to the real one but they often are quite different.   

[jhob]

Looks like farnell are out of stock 😭
https://uk.farnell.com/on-semiconductor/2n5458/transistor-jfet-n-to-92/dp/1467938

I'll just get another from musikding, realised they do letter postage which is pretty reasonable cost.

Maybe need to think about stocking up on these rarer components

[FSFX]

Looks like farnell are out of stock 😭
https://uk.farnell.com/on-semiconductor/2n5458/transistor-jfet-n-to-92/dp/1467938

I'll just get another from musikding, realised they do letter postage which is pretty reasonable cost.

Maybe need to think about stocking up on these rarer components

Farnell are out of stock because the 2N5458 is not made by ON Semiconductor anymore. This applies to many other JFETs used in pedals so it is always going to be a gamble getting any of these. Even Musikding will need to get theirs from unreliable and untrusted sources. There may be some brokers that are sitting on a stock of these but eventually they will run out which leaves the market wide open for the fakers.

As I said above, Farnell have loads of J113 JFETs. They would work just as well, if not better with the resistor values you have.

[jhob]

Just ordered some more 2n5458, 2n5457, J112 & J113 from musikding

Interested in trying them all and seeing what the difference is, it will be a potentially interesting learning experience for me.

I did have a look at the specs and J112 appears a closer match on paper for the 2n5458.

[Rob Strand]

I did have a look at the specs and J112 appears a closer match on paper for the 2n5458.

FYI ...

For that circuit you need a JFET with an IDSS of more than say 3.6mA.

Here's why,
For an 18V supply and say 5k drain resistor the JFET current needs to be about 18V/5k = 3.6mA.
If IDSS is less than 3.6mA it probably won't saturate like a 2N5458.

From the datasheet the IDSS for 2N5458 is 2mA minimum and 6mA typical.  In general the IDSS for the 2N5458 doesn't meet that requirement.

However it's not that simple.  When you adjust the drain trimpot it might end-up with a different drain resistance so the precise minimum IDSS depends on the specifics of the JFET you have.  The lower IDSS JFETs tend to require higher drain resistance and that means the minimum IDSS (~18V/Rd) drops with the JFET.  So the 3.6mA is only a guide.
(Working out exactly when things are OK with different JFETs has too many details to post.)

Also, if you look a the ROG results,  the JFETs don't tend to sit at the extremes of the datasheet,
https://www.runoffgroove.com/fetzervalve.html

Out of ten (10) 2N5458
average Vp: -2.18V (spanning from -1.74 to -2.87)
average Idss: 5.22mA (spanning 3.62mA to 7.41mA)

Don't forget if you have to adjust the drain trimpot too far over 6k to 7k it starts to affect the output filter.