Speaking of SMD JFET's ...

[kothoma]

I got a chance to put together the ROG Eighteen on the breadboard and test the J620 in it. It works. I built it up with the J201's first and then changed the first position to a J620. It brings down the distortion a touch -- still plenty of drive left though and I thought it sounded good. I then put one in the 1st and 3rd positions. This cleans the circuit way up. It still overdrives, but much more into a light crunch sorta area. I could not get a J620 to bias correctly in the second position. Actually, it did bias to the rec'd 7V, but the audio was gone. Biasing to any other voltage didn't work either. Itnteresting, but I tried an MPF102 in this position and it wouldn't work either. So, I think it has something to do with the gain structure off of the source and this would need to be altered.

The conclusion so far is that they will work in distortion circuits. The drop in gain does not seem to be as much as would be predicted from the ROG calculator. All-in-all, I think the J310's would be worth checking out. I'm going to keep experimenting with the J620's and see what more they could be good for. I think I'll try a discrete op-amp input stage next.

Great discoveries! I'm not sure what this second stage (phase splitter) really does. Kind of negative feedback?
But it's good to know that they can be used in phasers and distortion circuits. The J310 is one of the last affordable/available through-hole JFETs. Thanks for your investigation.

[RobA]

I've got some J310's in an order I'm going to place today from Mouser. So, I'll be able to test them directly in a couple of circuits. I'll let you know how that goes.

Yeah, on the actual amp, the second section has a real job. I think they've included it here to simulate the frequency response of the section in the amp.  They note that they've altered the resistor (R17) for the JFET's. I think doing some further altering or replacing with a 25k trimpot could make this stage work for the J310. On the other hand, I think the circuit worked best leaving this a J201 and using only one of the J620's in the first stage.

I'm working on designing a phase shifter using the J620's. At this point, I'm pretty hopeful that the match is good enough in pairs to do this.

Which reminds me, I need to try this in an Orange Squeezer. You could make one really tiny OS using these and an SMT op-amp and diode.

[RobA]

I breadboarded an Orange Squeezer with the J620. It works great. I built it up first using 2N5457's and a 1N34a and then replaced the transistors with the J620 and the diode with a BAT41. In my opinion the J620 based version actually sounds better. Just for reference, I used an NJM4558L for the op-amp and used the second half to buffer the Vref.

So far, the J620 is looking pretty useful. I'd like to find some other JFET arrays, but I haven't found any others to try yet.

[kothoma]

I breadboarded an Orange Squeezer with the J620. It works great. I built it up first using 2N5457's and a 1N34a and then replaced the transistors with the J620 and the diode with a BAT41. In my opinion the J620 based version actually sounds better. Just for reference, I used an NJM4558L for the op-amp and used the second half to buffer the Vref.

More great news. Thanks!

I'd like to find some other JFET arrays, but I haven't found any others to try yet.

Dual JFETs from Linear Integrated Systems:
LSK389
LSK489

And there are dual JFETs with common source (5-pin)
- from Toshiba:
2SK2145
2SK3320
- and similar ones from On Semi:
MCH5908
CPH6904

[RobA]

Thanks for pointing out those other arrays. Mouser doesn't have any in stock. I'll have to look around elsewhere. The LSK stuff is hard to come by and their website for ordering is silly. I'll have to contact them at some point about getting some samples or something.

I tried the J310 in an amp sim of an old Ampeg amp that I built. It does work as expected. I only used it in one spot because of the pain in setting the bias (more below on this), but it does reduce the gain and overall tame the signal in what I think is a good way.

Here's the one thing though, it doesn't work all that well to just put this in an already existing board. That's because the bias pot that is there, 50k or 100k in those I've tried, are much too big for the J310 and the bias point for the J310 is narrow. It really needs a much smaller trimmer, 25k or even 10k, to get it to work right in the same circuit as a J201 or 2n545x.

All-in-all though, I think it's worth some experimentation.

[kothoma]

That's very interesting. So you'd say the source resistor can stay as in the original designs and it is enough to adjust the drain resistor?

[RobA]

Good question. It seemed to work out OK in this situation and the couple I tried the J620 in, but in the newer layouts I did, I moved to using a trimmer on both the source and the drain and on those I did move the source trimmer when I adjusted by ear. It would be worth experimenting both ways.

On the Ampeg sim I just tried it in, I used the source resistor values from the original amp, so that probably plays into the whole thing somehow. On that one, I had tried a whole bunch of different JFET's in the various locations and they all worked, but I did settle on the J201's everywhere for the overall sound even though the gain is just completely over the top. I did like the J310 in there though, so I'll probably go back and do some breadboarding with the circuit when I get some time.

[jubal81]

Hmm. Having to use trim pots kind of kills the cost/size benefits of SMD. Is it a matter of biasing the 620s correctly for each circuit or do the 620s just vary unit to unit too much?

You know what would be great - if there are some lead clips that could be wired to a pot and clipped into the board without soldering. A quick biasing tool would be really handy. I'm going to look around for connectors.

[RobA]

I haven't had to do any additional biasing other than what the circuit already calls for. But, it would be slick to be able to establish what bias you need and then just solder in a resistor on the amp sims. You might be able to do it using a rig with pogo pins.

[jubal81]

So I dove into the junk box and came up with something.
Single row header -snap away 4 pins. Heat the middle pins up with the iron and yank them out. Attach two wires with alligator clips. If i had some female-female headers I could make one for caps, too - like a temporary socket you don't have to solder in.