Red Llama - Discrete conversion

[somnif]

I'm currently in an "I'm bored, lets make random stuff" mood. Found a layout over at Effects Layouts for a Red Llama build with discrete MOSFETs that looked fun, so I've got some p-channel parts on order.

In the mean time, though, the comments make me curious if it is actually "correct".

The Llama:


And done discretely:

from: http://effectslayouts.blogspot.com/2019/01/discrete-llama.html

First question: Every llama I've seen has a 1k resistor on the power rail. I know its part of the filtering, but presumably also acts as a current limiter. Would its absence have an impact on the health of the FETs?

Next: So I know the CD4049 is just a bunch of NOT gates, basically a P-channel and N-channel MOSFET in parallel. Input goes to both gates, P-channels (BS250) source is power rail, N-channels (BS170) source is ground, and drains tie together to output. In the llama first would invert, second would revert, simple enough.

I see that in the Discrete Llama.... but in each stage it then feeds into another BS170. Is this part of the 4049 that I'm unaware of? Is it just acting as a booster for each stage? (are the discrete FETs colder than the 4049 and need some help?)

Its a simple enough circuit to breadboard, I just didn't have any p-channel mosfet's on hand. And while I wait on Arrow I have naught to do but think about this. Any thoughts what the purpose of those extra 170's is?

[Scruffie]

Extra 170 is an output buffer on each stage, could replace it with an NPN if you wanted and you should consider putting in static protection diodes for the MOSFET inputs.

No, the loss of the 1k will mainly just change the operating voltage which will affect the sound but so will (one would hope) using discrete inverters. Those MOSFETs are good up to like 60V IIRC.

[TFZ]

I might be wrong since the circuit was actually released that way, but I don't think this can work without the current limiting resistor. Both mosfets are biased on, this is an operating point a logic gate isn't supposed to have. They are both low resistance and a corresponding current will flow, only limited by the mosfets' drain resistance. That's why with any such effect circuit a resistor is included. But mosfets differ of course, so it might be the case that for BS170/250 the current is at a reasonable level without the resistor, you will have to find out for yourself. Also take note of the last comment on the page you linked: "the transistor go very hot".

[somnif]

Also take note of the last comment on the page you linked: "the transistor go very hot".

I'll admit that is what made me consider the current limiting. But MOSFETs are weird, so I wasn't sure either way. I intend to breadboard the thing once the parts arrive to see if it works as-is, then experiment out from there.

[Scruffie]

A BS170 & 250 aren't the same kind of MOSFETs used in a CMOS chip (nor are they going to be complimentary like in a CMOS chip) I can't remember what the ON-resistance is so TZF might be right but the 1k resistor isn't going to change that situation. There's plenty that would cause that circuit not to function but I assume it can if they released it.

[storyboardist]

Let me know how you fair with this. It was sent to me by Alec from Mask Audio Electronics who'd worked with John from Electronic Audio Experiments on the idea. Don't know if either of the had breadboarded it or not, and I haven't gotten much feedback on the blog about it (that's constructive anyway) other than "it doesn't work" and "the transistors get hot" 

[somnif]

Yep yep, I've got it all set up on my breadboard OTHER THAN the bs250's which are apparently coming from the Netherlands (damn you Arrow, taking a whole 4 days to ship something 5000 miles for free!   ). Will just need to shove those in and connect some jacks.

I've always liked the llama/tube sound overdrive-y circuit, so I'm curious how the discrete version compares.

[TFZ]

I actually tried this a couple years ago when I looked at several CMOS overdrives. I used a dual complementary mosfet (I think it might have been this one) in order to minimize differences between the p and n part. Without a limiting resistor the current was above 1A, that's where I had the limiter of my power supply set. With a resistor in place it amplified just fine, but I never put any guitar through or used it in any real circuit. Too much other stuff going on...

[somnif]

Ok, sit-rep.

BS250s arrived and put the last bits on the breadboard.

WITH a 1k resistor on the power rail, sound passes through. WITHOUT 1k resistor, I get nothing. (I also had some zeners on the MOSFETs in both cases).

Didn't note any temperature change in the components, but thing was only powered on for 30 seconds at a time.

BUT.... my crappy Tayda breadboard is so loose and wiggly in its grip on the parts that just me strumming the guitar was enough to make things wobble. I'm not sure I'm entirely confident in the opinions gleaned from its current state in either case.

Next step, I'll throw the circuit together on some perfboard and see how it behaves there.

[storyboardist]

Cool. Here's an updated layout. Essentially just replaced the 1N5817 polarity protection diode with the 1k resistor and added a 1N4001 diode from voltage to ground for polarity protection. Keep us posted!

[Scruffie]

Well what do I know about MOSFETs unless there's 128 of them in a row

[somnif]

OK! Confirmed-ish

Finally managed to get some grading done and had a few minutes to finish throwing it together on perfboard (word to the wise, buy good stuff, making tracks on this cheap board was a nightmare) AND confirmed my earlier findings, more or less. This was using the older layout, just with a resistor added at the power point.

WITH 1k resistor on the power rail, it behaves like a Llama. Maybe a bit cleaner/lighter on gain than I'm used to from the circuit, but functions just fine. (even at max volume/gain its only just starting to break up, may be different on a hotter amp than mine)

WITHOUT the resistor, it does pass sound, but its hissy and broken sounding, AND draws MORE THAN AN AMP OF CURRENT. After a few seconds I could smell hot electronics so I pulled the plug. The FETs were already painfully warm. The 170 and 250 are rated for 500mA, but they do seem to have survived their brief cooking.

So yeah I'd say the current limiter on the power rail would be a good thing for this circuit.

edit: just realized I was using a linear taper for the Gain pot, oops.

[Scruffie]

What voltage do you get at the output of the 'inverters'?

Edit: And the supply after the 1k, just wondering how it biases up.

[somnif]

What voltage do you get at the output of the 'inverters'?

Edit: And the supply after the 1k, just wondering how it biases up.

Huh, thats odd

Getting about 4.8v after the 1k (before the 5817 in my case), and about 4.25 at the drains. That doesn't sound right.

~9.1v coming out of the wall wart.