Building a RAT but need substitute for 2N5458, in post I list my options

[jgoldkamp]

Building a RAT but need substitute for 2N5458, out of these transistors, which would be the best sub? BC337, C1815, S8050, 2N2222, 2N2907, 2N3094, j201, 2N697, 2N3704, OR 2N334A?

Also, is there a site where you can plug in a transistor code and get a list of compatible subs?

[Kevin Mitchell]

That's an N-channel Jfet. Try the J201. It being for a distortion pedal you may like the results with this one.
Or, https://www.mouser.com/ProductDetail/Central-Semiconductor/2N5458-PBFREE?qs=u16ybLDytRbDLInQLnKe%252Bw%3D%3D

If you google 2N5458 alternative you'll be greeted with the dozens of other posts on this forum asking the same thing.
Along with a list of suggested alternatives.

[antonis]

There are plenty of (old) schematics where 2N5458 (wrongly) appears as BJT..

In case of working on such a circuit (Gate/Base biased at Vref via a big value resistor), you can implement any high beta n-p-n BJT..

P,S.
Gord (GGBB) could enlighten you more about RAT mods..
https://www.coda-effects.com/2016/01/proco-rat-white-face-1985.html

[jgoldkamp]

That's an N-channel Jfet. Try the J201. It being for a distortion pedal you may like the results with this one.
Or, https://www.mouser.com/ProductDetail/Central-Semiconductor/2N5458-PBFREE?qs=u16ybLDytRbDLInQLnKe%252Bw%3D%3D

If you google 2N5458 alternative you'll be greeted with the dozens of other posts on this forum asking the same thing.
Along with a list of suggested alternatives.

Yeah, I'd googled and I'd see when anyone suggested a J201, someone underneath would say "don't do that." 

[Rob Strand]

Yeah, I'd googled and I'd see when anyone suggested a J201, someone underneath would say "don't do that." 

Off the bat J201's won't bias correctly because they have a much lower turn-off voltage (VP) but you can give them a nudge by adding a resistor between the gate and +9V.   This trick will work on most JFETs with VP's less than the 2N5458.   

Circuit 2 is fine.  Circuit 3 is just show how you can knock the gain back (no point).  I used standard resistor values instead of fine tweaking the resistor values to get the same DC bias points.  The voltage offsets can be seen at the top of the plots.  At this level of matching they can be ignored since the simulated J201 circuit is going to match better than different 2N5458's.   For this same reason I haven't been too fussy about choosing spice models for the J201 or 2N5458.



There's quite a number of old posts on the RAT buffer, some with GGBB.   There's discussions on the bias point and all these tricks.

You can see some of RAT variants on GGBB's site use the added resistor.   The trick is to get the input resistance the same, or at least close.

[Clint Eastwood]

You will find many good suggestions for a replacement, but the problem is that a lot of them are no longer manufactured or only as smd.
The J113, J112 and J111 still survive as through hole components. Although it is not on your list, for this application the J111 would be most suitable I think, it would bias at around 3.5 volts at the source.

[Kevin Mitchell]

Mouser's acting weird the passed few days. But the link I shared was for 2N5458 PBFREE which are currently available to purchace.
Rob's spot on about needing some tweaking to work more adequately as a sub. Some folks don't even bother and are happy as is. But of course that depends on the application, and the person 

[Rob Strand]

IIRC, it's not uncommon for Rats with 2N5458's to bias down to 1V on the source.   On average maybe 1.6V - I suspect GGBB has more precise numbers.   Still doable with that trick I posted.  You can lower the source voltage by increasing the added gate resistor (3M9) to +9V.

[antonis]

IIRC, it's not uncommon for Rats with 2N5458's to bias down to 1V on the source.   On average maybe 1.6V

Of course, that could be a problem for relatively high V_F clipping diode pair..
(IMHO, even for ordinary Si diodes, for 2N5458 current sinking mode - especially for output 10k shunt resistor, like in RAT 2..)

[Rob Strand]

Of course, that could be a problem for relatively high VF clipping diode pair..
(IMHO, even for ordinary Si diodes, for 2N5458 current sinking mode - especially for output 10k shunt resistor, like in RAT 2..)

The silicon diodes prevent the JFET bottoming out.  Notice on the the Turbo RAT, with the LEDs, they bias the source at a higher voltage to accommodate the increased swing.

[GGBB]

I suspect you can make most "typical" JFETs work (J201s are not "typical"). Just look for something with a similar cutoff range to the 2N4548 which is -1-7V. Lower (more negative) is better, because you don't really want the JFET to go into cutoff.

I know that FATRATs use J109 (-2-6V) - my unit - and J310 (-2-6.5V) - and possibly others - without any changes to the bias circuit. In circuit, the gate is biased at 0V and the source voltage ranges between about 1-2VDC, so G-S cutoff of the JFET needs to be below that difference. And since JFETs have a wide variance range it's wise to hand-select. Keep in mind these numbers don't include AC signal swing. J201 is -0.3-1.5V which makes it unsuitable.

Interestingly, my J109 FATRAT has G-S at -4.87V - but that is within the spec range of -2-6V.

My FATRAT-updated Multi RAT schematic is below.

Notice on the the Turbo RAT, with the LEDs, they bias the source at a higher voltage to accommodate the increased swing.

For the curious, that's done with R16. You can safely do this to any RAT to help with JFET cutoff due to signal level, but it won't magically make any JFET work.

[Clint Eastwood]

With the bias arrangement of the turbo rat, you can use your j201 without a problem. You could in that case  decrease the source resistor to 4.7k to boost the current a bit.

[Rob Strand]

For the curious, that's done with R16. You can safely do this to any RAT to help with JFET cutoff due to signal level, but it won't magically make any JFET work.

The main use of the gate bias trick is to help bias low VP JFETs when they replace higher VP JFETs.   You can use the biasing method for any JFET but for 9V circuits it might turn out better not messing with the gate voltage.

For a buffer circuit the source voltage can limit the swing.  The gate-source voltage swing is quite low for a buffer, even when the swing on the gate is large, because of the natural negative feedback.

With the bias arrangement of the turbo rat, you can use your j201 without a problem. You could in that case  decrease the source resistor to 4.7k to boost the current a bit.

For a J201 Turbo RT I'd keep the source resistor at 10k. With a lower source resistor Id_pk = (Vo_pk + VS_bias)/RD can exceed IDSS for the J201 on the positive swing.

To maximize swing you would bias the source so it is close to cutting out on the *largest* negative swing, and just hitting IDSS on the largest positive swing.    With a circuit like the RAT with limit input range you don't *have to* go that far but you at least have to be sure it's not going to bottom out on the -ve swing (Vs=0, Id=0) or hit IDSS on the positive swing.

Using the National Semiconductor J201 model, and tuning the gate bias voltage, roughly,
- For Rs = 10k the maximum input swing is +/-3V when the source is biased to 3.0V.
- For Rs = 4k7 the maximum input swing is +/-1.5V when the source is biased at 1.4V to 1.5V

In both cases (Vo_pk + VS_bias)/RD hit IDSS but the Rs=10k case does it at a higher swing.