Gallien Krueger 250ML overdrive pedal?

[sm3000]

I was just thinking that since the Gallien Krueger 250ML has a great overdrive channel it would be cool to have a pedal that could replicate that channel. Since the amp is solid state it shouldn't be too difficult to extract the overdrive circuit from the amp schematic and draw a layout.
And the sound is great! I own one..Iron Maiden and Rush in a second 

[sm3000]

more on the 250ml:
Gallien-Krueger 250ML: The way how these amps combine an IC dual FET and an opamp in order to create a soft clipping gain stage is ingenious. Some other G-K amps use another embodiment of the FETs + OpAmp idea. What's best, they utilize the soft clipping gain stages practically throughout the whole preamp, not just in a single overdrive stage. Old G-K amps have plenty of unconventional circuits, like the soft clipping power amplifier of 212SC and 300GT amps. Trace Elliot had few soft clipping power amp circuits as well but they were realized in a different manner.

Here's the schematic =)
http://www.megaupload.com/?d=SEI9EVYU

[valdiorn]

I love those old Gallien Kruegers, especially the 250ML. The way they use the FETs for overdrive is really onorthodox, but once you think of it, it's so simple and straight forward
I'm thinking about building a 250ML channel into my amp project

[sm3000]

I'm trying to isolate the overdrive channel and see if I can turn it into a pedal. I followed the "GAIN B" on the schematic and I think I should be able to capture the distortion of the 250ML following the schematic inside the green perimeter. Would this makes sense?



here's the schematic in bigger size http://img.photobucket.com/albums/v210/bionuker/250over.jpg

[sm3000]

U 5,8,9... are so called J412 / J406 dual n-ch jfet 8 pin. Which I can't find anywhere on the internet. On the chip in my amp is marked M8841. What could be an equivalent?!

U1, 4... are dual jfet op amps LF353N

[tranceracer]

That tone stack looks interesting!

[anchovie]

Hmm...curious. J406 and J412 are indeed dual JFETs, in that they are two N-channel devices in an 8-pin DIP package. Your schematic shows each IC labelled GK00040 as having connections to six pins, which would make sense as there'll be two gates, two sources and two drains in the device. What doesn't make sense is that the schematic shows 4 of the pins on each shorted together, no external components around the IC, no connection from the PSU and no connection to ground.

There's a cross ref document that names the LS842 and LS845 from Linear Systems as equivalents, but the datasheets for those show the active pins as being 1,2,4,5,6 & 8, which would mean that according to the schematic the audio signal goes to a pin that has no function.

I suspect this is one of those schematics that has been deliberately created to disguise the "special" part of the circuit, in this case whatever funky wizardry they are performing with the FETs. Are you able to trace anything on the PCB in your amp?

[sm3000]

Sure, let me open it up and take a look. Yeah I think that part is a little fishy! I've never seen ICs wired up like that, apparently it doesn't make sense, but if you download the .zip file there are 2 schmeatics in the same .pdf (which is the official release by GK). Here's the second schematic and here's the layout. And they seem to connect the pins as drawn on the schematic! I'll open up my amp and take some pics later on today..



on the layout picture U5, 8, 9 are above the pots R39 and R46

[sm3000]

I found two interesting pages concerning those obscure ICs.
http://www.electronicspoint.com/8-pin-dip-pinout-jfet-t107627.html
http://groups.google.com.tj/group/sci.electronics.components/msg/1225642226c0a68f

in the first one they say the J412 is equivalent to the LF412 (which has the same pinout of the TL082) and the pinout would be this:
1 out a v+ 8
2 - in a out b 7
3 in a -in b 6
4 v- in b 5

So following the schematic pins 2, 4, 5, 7 are wired together...would it make any sense?!

[sm3000]

So I don't think it's a TL071 which has this pinout so 3 is the input and 6 is the output would work but in U8 pins 3 and 6 are swapped (?), and still no ground or V+.

And if it was something like this?? http://focus.ti.com/lit/ds/symlink/ucc3837.pdf or http://focus.ti.com/lit/ds/symlink/tle2161a.pdf

[anchovie]

I think the mention of the LF412 in that link was just someone thinking out loud and therefore a red herring. That's a dual op-amp whereas the J412 is a dual JFET, i.e. just 2 transistors in one DIP8. Those ICs can't be op-amps based on the schematic/layout as the lack of a power source would make them stop the signal dead in its tracks.

[sm3000]

can you think of any proper and available equivalent of the J406 / J412?

[anchovie]

There are the Linear Systems devices that I mentioned earlier, but I still don't understand why the schematic has the ICs hooked up in a way that suggests they don't do anything. The Linear Systems website says that their ICs are drop-in replacements for other units by Motorola/Fairchild/Vishay etc. so there is an industry-standard pinout for these dual JFETs but if you follow that pinout against the GK diagram then it's basically saying that it would stop the audio signal at that point.

What would be interesting to find out, if you have access to an oscilloscope, is if there is actually a difference in the waveform when you compare the signal at pin 3 of one of the ICs to pin 6. Maybe even an audio probe would give an idea of any change that occurs.

It all goes back to the lack of a V+ and ground connection to the chips, and all those shorted pins. Even if it had been set up to use the FETs passively as diodes I'd expect to see some connection to ground to clip the waveform. I'm very suspicious of the schematic given that one page gives the choice of J406 or J412 and the other names the chips as the mysterious "GK00040".

[sm3000]

Unfortunately I don't have an oscilloscope but just a Ohm meter. As I mentioned earlier the chip in my unit is marked M8841..
If you adjust those trimmers (in the schematics are called R35, 36, 37 the gain changes dramatically! they have a big importance in regards of the distortion texture)

[teemuk]

Hello all, my first post here. 

The "mystery IC" is indeed a J406 dual JFET IC, not any proprietary chip. The FETs are wired as shown in the schematic and it's a perfectly valid circuit.

When FETs are wired that way they become current regulators, which is a pretty common circuit but not used very much in audio circuits, let alone guitar amps or stompboxes. What this regulator does is that it limits current whenever it rises above a certain level. At low VDS the FET is operating in its "passive region" where its output current follows the input voltage. Above a certain drain-source threshold voltage the FET transitions to "active region" where output current IDS becomes constant.

Practically, that means clipping since current is not allowed to increase to its "fullest level" that would be required to produce the signal peaks without distortion. Instead the peaks will just "flat out". FETs transition to this limiting somewhat gradually. Since one FET can only handle one half-wave at the time there are two wired in series. Both are contained on a convenient IC package.

The OpAmp is not necessarily needed but it provides a nice virtual ground reference and also adds some gain to the otherwise passive circuit. The trimmers will alter the gain of this stage but if the gain is increased too high the OpAmp itself will begin to clip and its clipping will "mask" the softer clipping of the current limiting FETs.

[sean k]

Hey Teemuk, thanks for clarifying that. So how exactly do they wire them? would it be gate and source or gate and drain and do you know of any discrete jfets which would do the same?

Interestingly, finding a discrete jfet that would do the same would actually let the cat out of the bag and one wonders if the opposite didn't happen with GK who found a discrete to do the job then found a non-proprietary device to shround the use in mystery.

[sm3000]

Thanks A LOT for clarifying teemuk...Do you have any idea about getting some J406? mouser.com or other vendors don't even have it on their catalog..maybe an equivalent part might do the job? But could you write the part# of an equivalent...?
Maybe using two 3-pin FETs (which one though?) instead of one double FET in an 8-pin package.

[teemuk]

I understand the theory behind the circuit's operation but that doesn't make me an all-around expert concerning device parameters, part numbers and possibly substitutes. I have no time or interest to check that stuff out.

I think Google search could reveal a lot about those issues, though. Similarly it could reveal other dual JFETs in DIL package. You could also just breadboard the circuit and throw in a few arbitrary discrete FETs and see where you get with them. While you're at it, you can also experiment with adding some source resistance. Anyway, I'm not the one building that circuit and doing all the work. 

As for wiring, you wire the FETs as shown in the schematic. Download the datasheet, check pin configuration and draw the (discrete) circuit on paper as instructed by the schematic. That's what I did as well when trying to figure out what that circuit actually was. If in doubt, simulate or breadboard. It's not rocket science. You can "cheat" from here: Gate and source are connected together in both FETs and used either as the signal input or signal output. Likewise, the drain serves either as the input or the output. This "polarity" will determine which half wave gets clipped.

You can also Google search FET current regulator and find a bit of theory.

[valdiorn]

Here's a little test of what I *think* the JFETs are wired up to do, at least it creates a nice looking distortion

View Spice simulation

One thing about this circuit is that it depends on current to distort the signal, not voltage (like most JFET and Valve CE gain stages, diode clippers, etc). In the first circuit (to the left), a 2.5V sine wave goes through the non-linear resistor (the JFET) and a 400ohm resistor. Increase the resistance and you decrease the current through the JFET and it starts behaving linearly again.

In the second example, which is pretty similar to the Gallien-Krueger (I think ), the op-amp creates a virtual ground at the inverting terminal, therefore the output is equal to the current going through the JFET as if it were connected to ground.

PS: I wasn't gonna give this secret away, since I'm going to use it to for some, ehm, top secret stuff I'm working on , but put an op-amp summer/mixer stage in front of the distortion stage and you can add either a static voltage or... something else, and use it to vary the bias, creating asymmetric distortion

[sean k]

http://www.diystompboxes.com/wiki/index.php?title=Simplemods

""FETs as diodes

  You can connect the Gate and Source to make a FET diode. Connect a FET Gate+Source to ground and the Drain to signal. Then connect            another FET Drain to ground and Gate+Source to Signal. Try these where you might have your standard silicon diode clippers.

   Buzzwords: "More touch sensitivity", "FET sound", "Tube crunch"  ""

So there we go, it's actually in the wikipedia. The inline use isn't there as a current limiter but the idea is around.