The Week on the Breadboard: Sabbra Cadabra

[Chuck D. Bones]

Holy Crap! This thing ROCKS!
I built it stock except I put a 250K trimpot in place R22 and connected pin 2 to C16. This is where the Treble pot would go in a Marshall tone stack. More than enough gain on tap. There is no clean setting unless you dial your guitar volume way down. Who wants to do that? So far, I've explored RANGE settings between 8:00 & 11:00. The only JFET selecting I did was I made sure that Q5 had a higher Idss than Q4. I used a BC109C for Q1. I'll try some mods and JFET subs tomorrow, but I'm pretty happy with stock circuit. It has a lot more gain than The Wizard. This circuit employs the relatively high drain-gate capacitance of the MPF4393 to roll-off the high end and smooth out the distortion. LTSpice tells me that the bass roll-off in the 1st stage coupled with the treble roll-off in the 3rd stage produces a broad mid peak around 500-600Hz and I think this is critical to getting that Tony Iommi tone. I expect other JFETs will sound different. More to come.

Of course the knobs had to be black!
L-R: VOL4, PRESENCE, GAIN, RANGE.
The treble trimpot is at the lower left.
See the two black jumpers and the 470Ω resistor below the two red WIMA caps? That's my Idss test jig.

 

[Laundryroom David]

Holy Crap! This thing ROCKS!
I built it stock except I put a 250K trimpot in place R22 and connected pin 2 to C16. This is where the Treble pot would go in a Marshall tone stack. More than enough gain on tap. There is no clean setting unless you dial your guitar volume way down. Who wants to do that? So far, I've explored RANGE settings between 8:00 & 11:00. The only JFET selecting I did was I made sure that Q5 had a higher Idss than Q4. I used a BC109C for Q1. I'll try some mods and JFET subs tomorrow, but I'm pretty happy with stock circuit. It has a lot more gain than The Wizard. This circuit employs the relatively high drain-gate capacitance of the MPF4393 to roll-off the high end and smooth out the distortion. LTSpice tells me that the bass roll-off in the 1st stage coupled with the treble roll-off in the 3rd stage produces a broad mid peak around 500-600Hz and I think this is critical to getting that Tony Iommi tone. I expect other JFETs will sound different. More to come.

Of course the knobs had to be black!
L-R: VOL4, PRESENCE, GAIN, RANGE.
The treble trimpot is at the lower left.
See the two bkack jumpers and the 470Ω resistor below the two red WIMA caps? That's my Idss test jig.
View attachment 55566

It’s a monster

 

[Chuck D. Bones]

The minimum gain setting is over 80dB. I'm thinking of cutting the overall gain by around 14dB.

 

[steviejr92]

Ohhhhhh im siked for this circuit....

 

[cooder]

Very keen to see what chuckified version you come up with! Sweet!

 

[drgonzo1969]

Hell Yes! This circuit was one of the first pedals I built that really knocked my socks off. I let my buddy keep the first build. I built second one and then I gave that one too my brother. They sound so good I can't stop giving them away. I am curious to see how this one gets modded too.

 

[mdc]

I've been meaning to give this one a try - thanks for the review! I built the DLSMKIII and didn't love it (to be fair, I subbed different FETs) - I had thought those two were almost identical, but it looks like the Sabbra has a whole extra gain stage.

 

[Chuck D. Bones]

Ladies & Gentlemen - Presenting The Black Box.

It took a lot of fiddling, but I'm very satisfied with the results. I kept the Sabbath Distortion ref des as much as I could. This has about 14dB less gain than the stock Sabbra Cadabra, which is to say it still has plenty of gain. I reshaped the bass response a little bit in the 2nd stage and the tone stack between the 3rd & 4th stages. I also rescaled the impedances in the tone stack so I could use a 100K trimpot. I think the balance between the bottom strings and everything else is just right. I set the Treble Trim around noon, but it's a matter of personal preference. I lowered the gain of the 1st stage by about 2dB. The original 2nd stage was an SRPP (similar to mu amp) with some local feedback. Since I didn't need all of the gain in the original circuit, I changed it to a simple common-source gain stage, adjusting the gain and the bass response. I left the GAIN control, 3rd stage and 4th stage as-is.

Q1 can be any high-gain, low-noise Si NPN. 2N5088, 2N5089, MPSA18, BC549C, etc.

A word about subbing JFETs...
Feel free to sub Q2, just make sure you hit the drain voltage within ±250mV and don't change R8 by more than ±25%.
Q4 and Q5 are critical to getting the right tone. Use 2N4393, MPF4393 or PN4393. Others will sound different.
Feel free to sub Q6, just make sure the Idss is at least 2mA and the source voltage is between 4.5V and 6V.

 

[jubal81]

Awesome stuff!
How noisy/hissy is it with all that gain?

 

[Chuck D. Bones]

I'm not hearing any hiss. On a breadboard, the noise is predominantly hum because nothing is shielded.

 

[spi]

The only JFET selecting I did was I made sure that Q5 had a higher Idss than Q4.

Is this true for any mu-amp arrangement? I've seen you mention this before in context of another pedal (I think it was the Golden Falk), but I don't recall it ever mentioned in Orman's articles on the miniboost.

And what is the issue if Q4 has higer Idss?

 

[Chuck D. Bones]

It all depends on how much current Q4 can sink in the specific circuit. However much that is, Q5 has to supply that current plus the load current and the most current any JFET can flow is Idss. In the Golden Falk, the bottom JFET has it's source and gate at ground potential (for one of the three MODE switch positions). It idles at Idss. The top JFET is supposed to regulate the average drain voltage on the bottom JFET. It can't do that if it current limits.

In this circuit, five parts limit the peak current flowing in Q4 and Q5: Q4's Idss, Q5's Idss, R16, R20 and C13. Let's assume that Q4 and Q5 have Idss = 5mA. The maximum DC current that can flow is limited by R16 & R20 to just under 1mA. Since C13 bypasses R20, the peak AC current in Q4 is limited to 1mA by R16. Q5 might have to flow more current depending on the load. In this case, it doesn't matter if Q4's Idss is higher than Q5's because neither JFET can get to Idss because of R16. Not all circuits have R16 to limit the drain current. We can ignore note 2 in this case because the min Idss for MPF4393 is 5mA. If we use JFETs with low Idss (like the J201), we have to pay attention to the peak drain currents or we might have an unpleasant surprise.

 

[lrgaraujo]

Looking for some higher gain pedals, so I'm definitely interested in this one!
What makes the 4393 so critical? A particular combination of Idss, Vgs and capacitance?

 

[Chuck D. Bones]

High transconductance for lotsa gain and high capacitance to roll-off the treble. We might be able to use other JFETs if we add a small cap from Q4-D to Q4-G. This is the gain of the 3rd stage, from pin 2 of the GAIN pot to Q5-S. All that treble cut above 400Hz is due to R14, R15 and capacitance in Q4 & Q5. Other JFETs are much brighter.

 

[lrgaraujo]

LTspice comes with some stock models for 2N4393 and PN4393, are you using those? I'll try to play a little with some other jfet models and try to match it a little

High transconductance for lotsa gain and high capacitance to roll-off the treble. We might be able to use other JFETs if we add a small cap from Q4-D to Q4-G. This is the gain of the 3rd stage, from pin 2 of the GAIN pot to Q5-S. All that treble cut above 400Hz is due to R14, R15 and capacitance in Q4 & Q5. Other JFETs are much brighter.

View attachment 55671

 

[Chuck D. Bones]

I'm using an MPF4393 model that either came with LTSpice or it came in a large collection of JFETs I found on the 'net. I get similar results with PN4393 & 2N4393.

 

[lrgaraujo]

A pair of 2sk117-GRs gives less gain, but a similar tone roll off (according to ltspice). I'll see if I manage to breadboard it this weekend

 

[Chuck D. Bones]

The schematic has been updated because I had C12 backwards. C12 should be tantalum or film to minimize leakage. It worked ok on the breadboard because the DC voltage across C12 was <300mV.

 

[Laundryroom David]

High transconductance for lotsa gain and high capacitance to roll-off the treble. We might be able to use other JFETs if we add a small cap from Q4-D to Q4-G. This is the gain of the 3rd stage, from pin 2 of the GAIN pot to Q5-S. All that treble cut above 400Hz is due to R14, R15 and capacitance in Q4 & Q5. Other JFETs are much brighter.

View attachment 55671

On the breadboard this sounds absolutely KILLER. Lower noise than the commercial pedal. Still an abundance of gain. This one is a keeper. Wow wow wow.

 

[steviejr92]

On the breadboard this sounds absolutely KILLER. Lower noise than the commercial pedal. Still an abundance of gain. This one is a keeper. Wow wow wow.

Dang i havent gotten around to this yet! You just amped me up! I cant wait to hear what this sounds like