Messages

On diodes, it can get a bit weird because there are three different footprints for them in SMD, SOT-23, SOD123, and mini-MELF. So, you have to find one that matches the type of diode you need and the footprint. The 1N4148 is easily available in a SOD123 format.

For voltage regulators, if you are designing your own circuits, there are better parts available in SMD than PTH, but you really need to read the data sheets because they'll have different capacitor requirements than the standard PTH regulators that are commonly used.

Thanks for info. What do you think what is good substitute for PNP transistors on the tayda? BC847 maybe..

The BC847 should be an NPN. It's the SMD version of the BC547. The BC5xx all seem to have an SMD equivalent that are numbered BC8xx. The BC860 should be a PNP that would be a good choice for audio.

I blame the Belgians for me liking mayo on chips...

I blame the English for the same. I was introduced to garlic mayo on chips at a little restaurant/pup place on the uni campus in Bristol. I generally hate mayo, but the combo was just incredible.

My real dislike of most things mayo comes from casseroles and such. It's lead me to my mayo rule. If something has mayo in it, it can't have more than four ingredients total.

[...]
Hoof schematic in question:
http://www.bigmuffpage.com/Big_Muff_Pi_versions_schematics_part4.html

Cody

Thanks! It does verify that I got the parts right. This sounds nice with the AC125's in it. I tried a couple of AC187's in the Q2 and Q3 positions of a different muff I built a while ago and, although it did sound good, it wasn't as good as it is with the AC125's. I attribute that to the nature of the AC187's, but I don't have any other NPN Ge transistors to test with. Is there anything special about 2n1308's?

The Hoof schematic (along with a shed load of other muffs) can be found on the Kitrae Big Muff website.

My search skills are lacking. I can't find the schematic on their site. I couldn't even find it using Google restricted to their site, but that's not too surprising anymore.

It's not that important though because I'm pretty sure I've got the trace through the tagboard version right.

Can you just "add" the ge trannies to a muff in the hoof locations? Or do you need to compensate elsewhere?

If you've got some NPN Ge transistors, you can just drop them straight in the Big Muff circuit. The Hoof does use particular values for passives in their Muff version, but nothing outside the bounds of other Muff variants.

If you need to use PNP Ge transistors, then you just need to invert the power and ground to the Q2 and Q3 transistors' resistor networks. It would be difficult on an existing PCB, but it's trivial on a breadboard.

This winds me up...

Why have you decided to make the 'L' another 'D'? It really isn't. A sod is a bit of a bastard, or a clump of earth (sodomy is something else entirely).

Please stop it.

I'm pretty sure that this is one of those cases where the Americans have preserved an older English pronunciation while it changed in the UK. So, either way you look at it, it's England's fault. 

I've done a bit more playing with this on the breadboard. I converted the Q1 and Q4 stages over to PNP as well so that I could test out using the Ge transistors in those locations. I actually preferred Si transistors in these two spots. And since I've got plenty of BC550C's and only a few BC560C's, I think I'll leave those as NPN on the layout.

I tried using a chain of two 1n4148 and one D9b in the clipping sections again. After a direct comparison to only the two 1n4148's, I think it's interesting enough that I'll make space on the PCB to make it possible to use that configuration of clipping diodes.

I found the Earthquaker Hoof on Tagboardeffects and traced that to get the part values for it. I think it will be possible to build that circuit on this PCB with a couple of off board mods for the tone control section and a couple of jumpers. The only real difference would be that this layout uses PNP Ge transistors.

Thanks for the input on this. I'll probably give it another day or so and then the PCB and send it off to OSH Park.

The earthquaker hoof uses ge transistors and it's a great muff variant.

Excellent! I'll check that one out. Thanks!

[...]
In that case, i think it's normal that you could'nt hear any difference as the compound of the semi-conductor remains the same aswell as the hFe range.

Yeah, that's what I was hoping for with the BC550 to BC560 move. That way, all of the change in sound would be down to the Ge transistors. I was surprised to learn that PNP Si transistors had been used in commercial variants and I'm wondering why they made the move. Was it just because of availability and cost?

Well setup Ge's in Q1 through Q4 yields stunning grunty/raunchy results

I kinda wanted to keep Q1 and Q4 as Si. That's because I was focused on the distortion from what's usually the clipping sources. But, I could flip those to PNP too and test with both the BC560's and the AC125's in there easily enough.

You've got me intrigued now, so what's involved in well setup? Bias points, Hfe selection?

Thanks,
Rob

Nice, does it sounds smoother? Did you try it with germanium diodes?

Hmm, don't know if I'd describe it as smoother. With the diodes switched in, the diode selection still dominates the overall sound. There is a difference though. It's got a somewhat fluffier edge if that makes any sense. It's also darker, but I countered that a bit by changing the feedback capacitor value. (I also changed these so the high cut would have more to work with.) With the clipping diodes completely switched out, the Ge transistors give you pretty much what you'd guess you are going to get by using Ge transistors.

I did try Ge diodes. I didn't really like what they did. I was trying to get the output levels of the two stages to be closer to what they would be with the diodes switched out. (They are still very different, but at least they are within controllable range.) That means that you either need to combine the Ge diodes with Si or use a pretty long chain (4) of them. I liked the 1n4148 and red LED's better anyway so I went with them as it would take a lot less space on the PCB.

PNP transistors has been done in muffs in the past, there are even a couple in the spreadsheet: http://juansolo.co.uk/stompage/Muffsv2.htm

Interesting. I tested with BC560's in the PNP positions after I converted them from BC550's so I could get an idea if the circuit were working OK before I went to the Ge transistors. I couldn't tell any difference with just the change to PNP.

I've been doing some playing with Ge transistors in a Big Muff like circuit on breadboard. I like it. I just flipped the middle two transistors to PNP configuration and it works well. In the schematic below, I've added some switching in the clipping sections and changed the tone control to be what I used in the Scramjet tone control and then followed by what's essentially a Rat high cut control. The standard Muff TC works perfectly well too, I just wanted to do some playing with this idea.

The diode switching is fairly critical though and needs an on/off/on switch so that you can get a fuzz purely out of the Ge transistors overdriving. I've been playing with some AC125 transistors in the approximately 100 hfe range.

Before I start laying this out for a PCB, I was wondering if anyone had suggestions for improvement or had seen something similar that I should check out first.

Schematic: http://rock.it-frog.com/Downloads/Graphics/ge_muff.pdf 

One thing that definitely happens is that as the speed of the LFO gets really high, the lamp can't respond as well. Certainly, by the time the rate is at 50Hz, no lamp is going to have any variation at all. You can see that just in the way a normal light works with an AC power supply. Where the drop off in response is going to hit and how much probably depends on the specific lamp used.

I modded mine to have a pot in place of R11 and R12. Actually, I replaced R11 and R12 with two 47k resistors and a 100kΩ pot. This lets you dial in the chorus notches to various depths. I think that the chorus is stronger with the preamp out of the circuit (modern mode). I still always use it in modern mode because I think it sounds better overall. I'd also note that my thinking that the chorus is stronger in vintage setting could be purely my imagination influenced by other tone differences. I also did the Geofex mod that subs in 10µF caps for all the 1µF caps in the phase shifter section. It's subtle, but I do think it sounds stronger. The mod I'd recommend most is the depth pot thing.

I have no idea if this means anything, but Gilmourish suggest that the EMS Synthi Hi-Fli might have been used to do the Leslie like sounds on Any Colour You Like.

I like Tayda's cheap machined SIP sockets. They're one of the few things I'll still get there. They're certainly not as good as high quality machined sockets, but they do work fine.

I have the cheap leaf pin sockets from Tayda and a couple of other sources. They've all worked fine. I actually prefer them to machined sockets in cases where I know I'm going to want to exchange the part lots of times.

The only socket I've ever had that failed or was defective was some expensive machined SIP sockets that were totally useless because all of the sockets were blocked part of the way down and the parts' legs wouldn't go in.

The formula I typed above has a typo in it. It should be V_out = 2 * V_in - (V_loss + 2.0 * V_diode).

I took some measurements on my Flintlock using a variable power supply. If I set the input voltage to something like 9.1V, then the voltage from the charge pump will oscillate a little bit around 16.85 to around 17.25 and the output from the regulator is a dead stable 15.1V. But, if I drop the input voltage down to 8.9V, then the regulator can't hold the voltage steady and the oscillations come through at the test point, going down to about 14.8V or so. I've used good, low ESR power supply caps on the charge pump too, so that will lower the losses in the charge pump. The 1n5817's I have in there measure at 167mV for the forward voltage .

5718 is the same diode as the 5817, but with a higher reverse breakdown voltage. It will have the same Fv. Potentially lower, actually.. They're used in the flintlock to rectify the switching in the charge pump.

In other words, yes you can use 5817, 5818, or 5819 with the same results in this design.

Hey Jon. The data sheet I looked at in the link above shows that the Forward voltage of these three diodes is a little different. Am I reading the data wrong?

Thanks for any clarification.

Steve.

The data sheet I have (Fairchild) shows them as different too. Looking at the plot of forward current versus forward voltage, they aren't very different at lower current draw, maybe 300mV for the 1n5817 and 350mV for the 1n5818.

I think it would probably still work, but it might be close. The LT1054 gives the output voltage (voltage doubler configuration) as V_out = 2 * V_in - (V_loss + 2 * V_diode) where V_loss is the voltage loss in the LT1054. This figure depends on the caps used and the output current, but it can range up to about 1.6V at 100mA. So, if you were right at 9V then V_out could be as low as 15.7V and that would be too low for the dropout voltage of the LM7815 (speced at 2.0V).  That would be an absolute worst case kinda thing. The output voltage from the charge pump section on my Flintlock reads high than that. But, a combination of using high ESR caps and the higher voltage diodes might possibly be enough to push you in to the danger zone on the drop out voltage for the regulator.

The difference in the forward voltages of the two diodes is pretty small, but there is some chance it could make a difference in marginal cases.   

Edited: Fixed the formula for the output voltage from the LT1054.