Mongoose / Giant Hogweed queries.

[Beaker]

By pure coincidence, I was in the middle of building a Roger Mayer Mongoose, when Alex posted the Giant Hogweed layout up on the main page. I want to build the Hogweed, but I want to get the Mongoose right first.

Reading David Rolo's notes on the Hogweed, he says that it is important to get the bias right on the LM308N to get the best sound out of the Mongoose circuit, with the optimum bias voltage at 5.5V.
So I measured my voltages, and found that I was only getting 4.2V on the IC.

Looking at the schematic:



and the layout:



I changed R5 (47K) - the left hand one on the layout, to a 33K which got my voltage up to 5.8V, and yes it sounds much better.
So then I changed R4 (47K) - the right hand one on the layout to a 33K to match.

Have I done this right? Do I need to change both of them or not?

David's Hogweed schematic has both 47K resistors on as well, so do they need altering too to get 5.5V?

Help, I'm getting confused!!!

[nocentelli]

If you just changed R5 and the voltage on the non-inverting input (pin3) was 5.8v and it sounded good, altering R4 will shift the voltage again, and may or may not sound good. Measure the voltage on pin 3 again with them both at 33k, decide if you like the sound (the exact voltage might not matter, drolo suggests it is because of assymetrical clipping caused by biasing above 4.5v). You could briefly replace the left hand 47k with a 100k (leave the right hand as it is) and dial in the best sound, then replace with a resistor of the measured value of the pot.  


With the giant hogweed, david rolo has split the direct DC-coupling of the transistor boost collector into the lm308 (by using a coupling cap which blocks DC bias and a subsequent buffer stage), so the two 47k resistors on the hogweed schematic bias only the transistor: the lm308 separately receives non-inverting input bias via that 200k resistor from the "stability pot" within the power section on the bottom right of the schematic. This pot allows you to dial in a variable amount of gating by freakin with the bias. If you think it's worth the bother, you could replicate this section to give the mongoose variable bias for the 308 separately from the 47k pair:

http://www.diystompboxes.com/smfforum/index.php?topic=108767.msg992490#msg992490

[induction]

Beaker wrote

I changed R5 (47K) - the left hand one on the layout, to a 33K which got my voltage up to 5.8V, and yes it sounds much better.
So then I changed R4 (47K) - the right hand one on the layout to a 33K to match.

Have I done this right? Do I need to change both of them or not?

The Mongoose's 308 is DC coupled to the collector of the transistor that precedes it. As such, the 308's bias is determined by the biasing of the transistor, which is sensitive to the sum of R4 and R5. If you reduced R5 and got (actually overshot) your target voltage, then reducing R4 will make you overshoot even more. You can verify this by measuring it, of course. So if you're aiming for 5.5V, then no, don't change R4 to 33k. Otherwise, choose what sounds best to your ear.

David's Hogweed schematic has both 47K resistors on as well, so do they need altering too to get 5.5V?

Help, I'm getting confused!!!

The Hogweed schematic shows that the 308 is AC coupled (C19), so the transistor biasing network has no relevance to the 308. The bias on the 308 (Vb2) is set entirely by the stability pot. (Beware: that part of the schematic seems to be missing a ground connection.) Those 47k's are only used to bias the transistor (T7). If the biasing of the transistor is important (I have no idea), then you can play with those resistors as well, but they won't affect the bias of the 308, which gives the asymmetric clipping you're looking for.

[Beaker]

Thanks Induction!

I had a total lightbulb moment after reading your posts. At first I wasn't sure I understood - even after reading them a few times.

A few hours later, while lying in bath it suddenly came to me. It's genuinely the first time I have ever had reason to shout "EUREKA!" from my bathtub!

[Beaker]

Ok, thanks to Induction's invaluable help, I have finally got this to a stage I am happy with.

I have changed R5 on the schematic (the left hand one of the two 47k resistors on the layout) to 33k. The right hand one is unchanged at 47k.

I am now getting a steady 5.52V on pin 3 on the IC. It does sound better as a result.

I have used KD521V clipping diodes, which sound best out of all the diodes I tried.
I did try 3mm red LEDs as well. Some builders say they sound great in a Mongoose, but personally, I did not like them - seemed to rob it of character.

I tried a load of transistors, before settling on ME4003 as my favourite. Metal can 2N2222 sounded great too, but not as smooth, and took away a little of the bass.

I know it is a alternative take on a Rat, but it does not sound like a Rat at all - far less spitty and farty. IMHO it blows a Rat into the weeds on every setting.

It's also really versatile. It can nail that Gossip "Heavy Cross" driven funk sound, and at full drive is pure Kevin Shields (not suprising as he uses one), and makes a great Stoner/Doom pedal.

Don't like Rats - build one of these instead. It's awesome!


Now on the Giant Hogweed!