- enclosure
- both boards
- main board
- see-thru
- daughter board
- under board
Catalinbread - Formula No. 55 [traced]
Yep. I think this is true for most of the Catalinbread Foundation Overdrive series.
Quite a few commercial "you can't tell the difference between this and the real amp" pedals seem to be two MuAmps in series. Vertex Steel String is another.
In this case, it's: MuAmp -> Tone & Volume -> MuAmp -> Tone & Volume -> Buffer
Yet somehow it still does a great 5e3 impression; capturing some of the original amps imperfections as well. This pedal is somewhere between a RunoffGroove design and "Make your amp sound like a Tweed Deluxe using this one SIMPLE TRICK" like the Lovepedal Les Lius.
That RunoffGroove level of detail in recreating tube preamps with Jfets is very cool, but comes with its own set of problems; noise, finnicky bias (cannot switch to 18v without adjusting each trimpot), reliance on J201s for gain. IMO, the Catalinbread F55 is a nice middle-ground between simplicity and authenticity. The build quality is very good too, I'll get into that and more in future posts.
The carbon comp resistors must be NOS or something because the resistances I'm measuring are way off. This isn't my area of expertise, but I was shocked to find 10k resistors with 5% tolerance (gold band) measuring at 15k. Almost every carbon comp resistor I measured was around 20% off.
There are also some mojo capacitors around the Tone control.
- C8 is a tantalum marked 471k. I think it's safe to assume it's just a 470p.
- C9 is a 4n7 axial cap that looks like one you'd find in a ptp tube circuit.
The push-button switch is a latching 4PDT, but unlike the Formula 5f6, only 3 poles are used, and one of those is for the LED. That means if you were willing to forgo the bicolor LED, you could wire this up with a standard DPDT (or even DPST) toggle.
I'm surprised that there is no pulldown resistor, but it doesn't pop when switching. The Muamps could have been biased for lower noise. Compare the way Q5 is biased to Q2 & Q4. Q5 is the smarter way, which doesn't add any thermal noise. When I roll my own F55, I'll be feeding all of the bias points from a single voltage divider for less noise and a lower parts count (can save 2 resistors)
The rest of the circuit is pretty much what you'd expect. I'm going to leave this on the workbench all week in case anyone would like me to check anything else.
There are also some mojo capacitors around the Tone control.
- C8 is a tantalum marked 471k. I think it's safe to assume it's just a 470p.
- C9 is a 4n7 axial cap that looks like one you'd find in a ptp tube circuit.
The push-button switch is a latching 4PDT, but unlike the Formula 5f6, only 3 poles are used, and one of those is for the LED. That means if you were willing to forgo the bicolor LED, you could wire this up with a standard DPDT (or even DPST) toggle.
I'm surprised that there is no pulldown resistor, but it doesn't pop when switching. The Muamps could have been biased for lower noise. Compare the way Q5 is biased to Q2 & Q4. Q5 is the smarter way, which doesn't add any thermal noise. When I roll my own F55, I'll be feeding all of the bias points from a single voltage divider for less noise and a lower parts count (can save 2 resistors)
The rest of the circuit is pretty much what you'd expect. I'm going to leave this on the workbench all week in case anyone would like me to check anything else.
-
aionios
- Solder Soldier
Information
This is awesome, I actually just pulled one of these apart last week intending to trace it as well. I will probably still do my trace but do it from scratch just to get independent concurrence but at first glance this is really well done and it tracks with other Catalinbread design patterns.daeg wrote: ↑20 Jul 2020, 16:52 The carbon comp resistors must be NOS or something because the resistances I'm measuring are way off. This isn't my area of expertise, but I was shocked to find 10k resistors with 5% tolerance (gold band) measuring at 15k. Almost every carbon comp resistor I measured was around 20% off.
There are also some mojo capacitors around the Tone control.
- C8 is a tantalum marked 471k. I think it's safe to assume it's just a 470p.
- C9 is a 4n7 axial cap that looks like one you'd find in a ptp tube circuit.
The push-button switch is a latching 4PDT, but unlike the Formula 5f6, only 3 poles are used, and one of those is for the LED. That means if you were willing to forgo the bicolor LED, you could wire this up with a standard DPDT (or even DPST) toggle.
I'm surprised that there is no pulldown resistor, but it doesn't pop when switching. The Muamps could have been biased for lower noise. Compare the way Q5 is biased to Q2 & Q4. Q5 is the smarter way, which doesn't add any thermal noise. When I roll my own F55, I'll be feeding all of the bias points from a single voltage divider for less noise and a lower parts count (can save 2 resistors)
The rest of the circuit is pretty much what you'd expect. I'm going to leave this on the workbench all week in case anyone would like me to check anything else.
Do you have any of the measured resistor values written down?
I have huge respect for Catalinbread's design process, but I've always been bugged by their adherence to mojo parts. I get that there may be tiny differences in capacitors (greenies, film, ceramic, even the axial type like they use here) but the resistors just drive me nuts. When they do the A/B tests during product development and swear they could hear differences in carbon comps vs. carbon/metal film, the difference is in the resistance itself, not the composition of the resistor.
You're absolutely going to hear a difference between a 100k carbon comp and a 100k film resistor, because the carbon comp might actually be 70k or it might be 130k. So if you prefer the carbon comp in listening tests... then just measure the carbon comp and use a film resistor of equivalent value.
Not to mention that the value of carbon comp changes based on humidity, age, and temperature. Even desoldering a carbon comp is going to change its value, sometimes permanently, even after it returns to room temperature.
Well put. You'd think most people would be bothered by this idea that these 'mojo' pedals vary wildly from unit to unit, and we're right back at the problems of inconsistency that they had in the 60's and 70's. How many stories have we heard that so-and-so famous guitarist used to go through buckets of pedals to find the one that sounded just right?aionios wrote: ↑20 Jul 2020, 17:51 You're absolutely going to hear a difference between a 100k carbon comp and a 100k film resistor, because the carbon comp might actually be 70k or it might be 130k. So if you prefer the carbon comp in listening tests... then just measure the carbon comp and use a film resistor of equivalent value.
Not to mention that the value of carbon comp changes based on humidity, age, and temperature. Even desoldering a carbon comp is going to change its value, sometimes permanently, even after it returns to room temperature.
Myself, I am one to fall for the idea of 'magic' in certain pedals, but the magic I like can be recreated through topology, frequency shaping, bias, impedance, etc, not unobtainium NOS components that need to be chilled to a certain temperature.
When tracing this one, I started writing the measured value vs the labeled value, but got irritated along the way and said 'screw it' when I measured a 10k 5% carbon comp at 15k (50% off!!!). I'd be willing to remeasure the carbon comps and the caps if it would make a real difference. If you do decide to do an independent trace, one thing to confirm would be that only 3 poles of the 4PDT are in-use.
Anyway, glad that someone else is interested in building (and producing!) a Formula 55 clone. I'm a customer of yours btw.
Hi daeg,daeg wrote: ↑20 Jul 2020, 16:52 The Muamps could have been biased for lower noise. Compare the way Q5 is biased to Q2 & Q4. Q5 is the smarter way, which doesn't add any thermal noise. When I roll my own F55, I'll be feeding all of the bias points from a single voltage divider for less noise and a lower parts count (can save 2 resistors)
I was just reading through the link you posted about Muamps, and got to the bit about reducing biasing noise. Regarding your comments on the biasing, am I correct in saying the use of 10K resistors for the voltage dividing is better than 1M resistors, but an extra resistor is needed on the gate in the case of Q5 that is 1M?
And the noise is created by the voltage dividing, but you can do that just once and then feed that into a gate resistor for Q2, Q4 & Q5?
I've just looked at the Steel String and that uses 10K resistors for the voltage dividing and a 1M on the gate, and I’d say thats a pretty quiet circuit.
Thanks for posting the schematic, I need to get some SMD adaptor boards for my FETs, but I fancy breadboarding this.
You nailed it! When I go to build my F55 I'll be doing it exactly like you described. It'll even save 2 resistors.FlyingWild wrote: ↑22 Jul 2020, 08:11 I was just reading through the link you posted about Muamps, and got to the bit about reducing biasing noise. Regarding your comments on the biasing, am I correct in saying the use of 10K resistors for the voltage dividing is better than 1M resistors, but an extra resistor is needed on the gate in the case of Q5 that is 1M?
And the noise is created by the voltage dividing, but you can do that just once and then feed that into a gate resistor for Q2, Q4 & Q5?
Current passing through high resistance causes thermal noise, so we not only have the opportunity to scale those resistors down to 10k, but also put that voltage divider behind 1M resistors which will pass a microscopic amount of current.
One thing to not forget is the large cap (C15) to filter AC to ground. This allows us to use that voltage divider as a 'virtual ground' and tie all of the bias points to it (via 1M resistors). This isn't anything new, tons of opamp based circuits use it to cut down on all of the DC blocking / AC coupling electrolytic caps.
Maybe sometimes there are good reasons to do biasing the Q2/Q4 way, but in this case I just don't see one. If so, I hope someone more knowledgeable jumps in and educates us.
Since no one had any questions, this one is resassembled and back in the box. Overall I'm pretty confident in this trace and hope others out there find the time to build it and try it out. Myself, I'm debating whether I should spend the time making a vero layout for this, or just wait for PedalPCB or Aion to put out a board.
As most of you know, the Tweed Deluxes (and most other tweeds) are known for being right on the verge of too much bass. I've come up with a simple mod to turn the the Presence control (which isn't part of the 5E3 circuit) into a Bass control (which also isn't part of the 5E3 circuit), with the flip of a DPDT. I think most people will find the post-gain bass-cut more useful than the post-gain treble-cut.
When I get a chance to draw it up, I'll post it.
As most of you know, the Tweed Deluxes (and most other tweeds) are known for being right on the verge of too much bass. I've come up with a simple mod to turn the the Presence control (which isn't part of the 5E3 circuit) into a Bass control (which also isn't part of the 5E3 circuit), with the flip of a DPDT. I think most people will find the post-gain bass-cut more useful than the post-gain treble-cut.
When I get a chance to draw it up, I'll post it.
Nah it's Brown-Black-Red-Gold.
- Screenshot from 2020-07-30 18-50-16.png (80.96 KiB) Viewed 7134 times
Take a look at the bottom left corner at the orange band on the 2 10k resistors for comparison. You'll also see that Catalinbread uses the 1k value in other foundation overdrives.
Just going to add that pairing this with a Rangemaster is glorious. My choice was the RoG Omega for a FET only distortion combo.
You can use a Rangemaster as a bass-cut or 'underdrive' to tame the bass in regular playing, then bypass it for a double-barrel shotgun blast of Tweed tone.
You can use a Rangemaster as a bass-cut or 'underdrive' to tame the bass in regular playing, then bypass it for a double-barrel shotgun blast of Tweed tone.
-
Lithium_Grim
- Breadboard Brother
Information
Is the Hi/Lo a 4PDT or a 3PDT. The schematic drawn indicates S1d which suggests there is a S1c. If so is it unconnected or connected to something else? I guess it is just a 3PDT and so a potential typo.
PS I guess the S1d schematic is wrong as it suggests there is always a LED on even when the effect is off. I'd suggest to put the LED switch of the true bypass after R69 to make it on/off switchable.
I only give negativ feedback and y'all love it!