Please help me understand and modify my Thrashmetal/Fat Cat/RAT clone

[ptkw]

At the beginning I should probably state that I'm fully aware that starting such a project without enough research and proper understanding of inner workings probably wasn't the greatest idea. What happened is that bored with waiting for missing parts for my another (better researched and understood) modification, I started browsing through my pedal collection and found an old Ibanez TM-5 Thrashmetal, of which I was never particularly fond of. Without giving it much thought, I opened it and started poking with it, simultaneously trying to gather some information on it.

What I found out was that TM5 was a circuit very similiar to other Ibanez pedals, namely LM7 L.A. Metal and FC10 Fat Cat (a RAT clone), differing only in some parts' values and the presence of three additional stages in the signal path. Basing on the guidance found here and on The Other Forum, I decided to convert it into something similar to Fat Cat and possibly apply some further modifications. I removed the stages that are not present on the Fat Cat and ended with something like this:



I stopped when it came to changing the values of resistors, caps and pots. I could have changed them according to the Fat Cat schematic, but thought it would perhaps be possible to find some more interesting ones. This, however, is rather beyond my reach, since I understand only partially how the whole pedal works.

The "distortion" stage (marked in my picture in green) is not that much a mystery for me; I know what a negative feedback loop is and what are the diode changing options (I'm going to replace the silicons with red LEDs). My question for you would be:

• The "blue" stage confuses and angers me (it is the only stage that has no direct equivalent in other RAT-based circuits). Could you please explain to me what exactly happens here and would it make sense to fiddle with it? Perhaps it would be reasonable to completely disable it as well?
• The effect in its present form doesn't work as distortion, but rather as a booster. I know that the values in the green/distortion stage are responsible for this. What would be the method of calculating the values of two resistor/capacitor pairs at the bottom of the schematic? What else here would be worth of changing?
• What about pot values? How should I find them?

I should probably add that my aim would be to make this effect as rich in the low end as possible, with lots of gain, and also as crazily/heavily distorting as possible (I'm not a guitar player, but rather something of a noise/experimental guy and I connect my stompboxes to many different things including synths and microphones). I reckon that this circuit doesn't allow that much craziness, but any suggestions would be welcome.

Many thanks for your help!

[Earthscum]

Blue stage is the buffer and switching. You can omit this whole stage by using a mechanical 3PDT true bypass. As a matter of fact, you can get rid of the output stage as well. Just take the output from the purple stage (unless you prefer to have the additional output buffer) to your jack (or switch).

Pot values are marked, usually, on the back, or on the face (gotta take it out to read it). Some are stamped on the sides, but if you don't see anything, it's probably on the face. Any wierd pot codes can be deciphered by quite a few people here for a nominal fee of "Thanks"    (cheap, ain't it?)

BTW, nice drawing.

[Earthscum]

Ya know what? The general circuit (minums switching) looks like a DS-1. If you take the tone control and replace it with a BMP stack, and replace that output buffer with the one in the (pink?), it's amazingly similar to a true-bypass DS-1 clone.

Try a 100k resistor in place of the "200E", and a 100k pot in the feedback. If you don't get "distortion", you could add a booster stage before it with the other half of that 4558. That would give you 2 stages of variable gain, if you so desire. Straight from a Rat to a DS-1.... kind of a cool idea!

[ptkw]

Earthscum, many thanks for your answer. Your explanation of the buffer stages was very helpful. As to using the other half of the 4558... that's an interesting idea for sure, and I would certainly like to give it a try, but first I would like to achieve just one tone that I would be satisfied with.

I have replaced the diodes with red LEDs today, and now the signal comes out as distorted, but only at very high volume threshold.

What I would like to understand is the mutual relation of all the elements in the feedback loop - I simply have no idea whatsoever as to how one thing there affects the others (not to mention the final tone).



I understand that the R17-C9 and R18-C10 pairs are of special importance here, and that they affect, respectively, highs and lows. I would be immensely grateful if someone would be kind to explain how their values should be chosen. (Explanation of all the other parts in this circuit would be also lovely, but perhaps that's asking for too much.)

Here are some sample values I have noted down, but I can't make much out of it:

• Original Thrashmetal (current state): R17=180ohm, C9=1uf, R8=4.7K, C10=1uf
• Original Fat Cat: R17=56ohm, C9=3.3uf, R18=1K, C10=4.7uF
• one suggestion on modiyfing Thrashmetal I found somewhere: R17=120ohm, C9=22uf, R18=10ohm, C10=10uF

Also, when I was asking about the pots' values, I meant the same: how new values should be chosen, and not how to read their current values (I know them from the posted schematic).

Again, I'm looking for the hardest (and preferably bassiest) distortion I can get from this circuit. Thanks for any suggestions!

[Earthscum]

I really don't know for sure how that arrangement works, exactly, but it may have a steeper response with some proper calculation and setup, not sure. In the DS-1, and TS, this is a single resistor and cap to ground. Easier to calculate, I think... Like I said, not sure about this one. The DS-1 topology, btw, is different... my mistake.

I'm kind of out of it right now,but basically your VR, R19, and C8 make up a variable high-pass filter in the feedback (resulting in a low-pass response at the output). The R17/18, C9/10 combo to ground forms a low-pass filter, resulting in hig-pass response at the output. You can simplify it by starting with 4.7k and 1uF, attenuating anything lower than about 33Hz at the output of the amp, and passing all other frequencies to the next stage, or the clippers. If you want some clean bottom end, say like I would do for bass, make that lowpass at around 500Hz (3.3k and .1uF = 482.5Hz).

Next part of this... the cap decouples the feedback from ground. Ovbiously. The resistor acts as a voltage divider, dumping part of the signal to ground, causing the amp make up for lost feedback signal in gain at the output. If you are doing 100mV P-P signal at the input, you would need 12-14 times gain to be able to even start to break over a Si diode. To get a LED with a Vf of 1.6V to clip the signal, you would need 32 times gain. That's just for the signal to begin forward conduction of these diodes. So, if you have the feedback dump resistor at 4.7k, and R19 at 100K, your gain is going to be around 47x. THIS will clip a Si diode very nicely. A LED will be louder, with not as much clipping and distortion, and Ge's will put out less gain at the output, but will have more fuzz. If you have another 100k on the pot, you can dial in 47-94x gain. Say ya want a clean sound on the bottom... that's where you can decrease the value of R19, but your total gain at top rotation will suffer. You can balance this out by changing the dump resistor, minding the fact that it is part of an RC network. If you make it smaller for some reason, your frequency response will go up, and you will want to make the cap larger to compensate. 4.7uF and 1k work good, for instance. Too much low end can sound flubby, not enough, thin and weak, unless you clip the highs and then boost the lows in a later stage, or cut the highs down later.

If you want full top end response, put in a small (56-330pF) cap in the C8 place, like is shown. If you have too many harsh highs, you can increase the size of the cap to tame them. I generally find that 560pF works best to my ears on bass, in most circuits like this, while 220pF seems to be my general preference for guitar (or bright bass).

[ptkw]

Wow, now that was an exhaustive explanation. Tremendous thanks! I will probably need to read through your post again at least several times in order to fully understand it (preferably when I'm not as much sleep-deprived as I am right now), but that's exactly an answer I was hoping for. I've just finished putting sockets in a good part of the "green" stage, so tomorrow I'll start experimenting with different values. Thanks again!

[Earthscum]

no problem... hopefully if I made any mistakes, someone will point them out. I've always tried to go with the idea that you have to empty your bucket before you fill it again... in other words, if you want to learn more, you pass along what you've already learned. Best way to get over "writer's block".   

[ptkw]

Thanks to advice provided by Earthscum and tons of additional information found in many different threads on RAT modding, I'm finally onto something... Although my poor PCB looks now like a military proving ground, full of jumper wires, sockets and sad barren areas (due to removed parts of the circuit), I've managed to accomplish more or less the sound I really like.

My most important to date modifications included changing all the pots to 100K and adding a silicone/LED switch.

However, some new questions arose and I came back here to ask for help.

• The idea of a dual R/C network coming from the FL is literally giving me headaches and I think I'm going to settle with just one resistor and one capacitor if I want to play with finding new values (instead of going for RAT's standard 2.2uf/47ohm + 560ohm/4.7uf). I think I understand how the resistor and capacitor pair is shaping the filter's cutoff frequency. What I'm not so sure about is the relationship between said resistor and the overall gain. 1Kohm/1uF, 100ohm/10uF and 10Kohm/0.1uF would all give us the same frequency rolloff; but would they provide the exact same boost?
• It's pretty elementary, but am I correct that the resistor marked in my schematic as R19, being in series with the gain pot, simply acts as this pot's extension? For example, if I use a 100K pot and R19 would be 47K, it would shift the minimal value of pot to 47K and maximal to 147K, right? (I just want to be 105% sure here)
• I've wired new pots in places of the old ones that were surface-mounted ones. Now everything works as it should except for occasional signal interruptions that can be immediately fixed by touching one of the pots. What does it mean? Is it a lack of shielding, a grounding issue or perhaps something else? (I've already resoldered wires to every pot to make sure that it's not just the bad soldering... or maybe the problem still lays here?)
• Would adding optional, switch-enabled clipping diodes to the feedback loop require some additional modifications in current parts' values?

[Earthscum]

1 - the 100 Ohm will give you more boost because more of the feedback (gain taming) signal is dumped to ground. It will give you 10x more gain than the gain with the 1k. The 10k will drop your gain by 10x because more signal is finding it's way to the inverting input to tame down the gain. That should 'show' you the relationship a bit better 

2 - Yep, you got it! Also, say you want a min value of 47k, and a max value of 100k, but you don't have a 50k pot, you can wire the pot straight in, and put a 100k resistor between the wiper and the lug... one way you have just the 100k, with the wiper and lug essentially shorting out the 47k, and the other way you have the 100k + 100k parallel = 50k. In between is kind of a strange taper, for instance at half rotation you have 50k + (50k | 100k = ~67k) = 117k. Fun to experiment with if you hit a block and just wanna play around a bit. A small cap from the wiper across to the other lug yields some tone shaping possibilities, more useful!

3 - Not sure about that one... I've heard of this being remedied by running a wire from pot case to pot case (soldering to the back like in guitars), and running that to the main ground termination works for problems like that, but can't say for sure. That may be why they ran them PCB mounted. Shielding may help... twisting the wire pairs or threes may help as well.

4 - Nope. I think Beavis Audio has a neat diode switching board idea. Just do a search on the site or on google for something like "switchable diode mod". This is a mod used in lots of circuits like the TubeScreamer, Big Muff, and others, and is done the exact same way as the ground clipping diode mods in DS-1's, MXR Dist +, DOD 250, to name a few. This is where ya get to really hear the difference in diodes.

If you are thinking of doing diodes in the feedback of the amp (adding them in), make a similar mod to the ground clippers. That way you can switch off the ground clippers and switch in , say, LED-Ge in feedback for asymmetrical clipping, or clean feedback with LED-Ge to ground (my pref for DS-1). OR, if you align them right, you do a LED-Ge in feedback with a Si-Si to ground. The side of the signal that gets clipped won't be big enough to clip the Si, but the one that clips the LED WILL clip the Si. You get, what sounds to me, like really decent clipping and some added dynamics. Slightly asymmetrical. If you can clip LED and Si in series with a Ge the other way, you start off with REALLY asymmetrical clipping, almost octavey, like half-wave rectifier, and the Si to ground (to clip the tall side of the signal... back to back doesn't hurt anyways), brings it down a bit more.

Basically, with clipping, the harmonics generated in one stage of clipping get carried through to the next stage of clipping.  The next stage's harmonics are partially generated and amplified by the previous harmonics. 'Soft' clipping (LED's... clips just the tops) first before doing a hard clip (Si small signal diodes) sounds better to my ears than just doing a hard clip. This is a mod in the big muff... use Si-Si in one stage, then Ge-Ge in the next. Even though each stage of clipping is amplified, the principle is essentially the same. That's why you clip soft in the feedback, otherwise it won't be big enough to overcome the next stage of clipping.

Hope I've babbled enough info to get you going on diode clipping. I've spent countless hours just on this alone... even got some new diodes in yesterday to try out! 

[mrdorbe]

:/