In comparing the Java Boost to Tonepad's Neg Ground Rangemaster, there is a difference I'm confused about. The signal out of the transistor comes off the emitter on the Java and the collector on the Rangemaster. I would have thought that the Java would act as a buffer (emitter follower), rather than an amplifier.
Also, are there any differences in tone when you build a negative vs. positive ground Rangemaster. Or put another way, why ever build a positive ground device?
Well there are a "ton" of posts about this, do a search!! , but a negative ground "PNP" Ge built is full of problems, which manifest themselves as
"noise" "hiss" "motorboating" etc....
I built ( quite sucessfully I might add ) a negative ground NPN Ge Fuzz Face, using AC127's.
There is a reason why these shouldn't work "right" due to the way they are made, "substrate" construction etc is diffrent from PNP and they are
supposed to not sound as good.
Mine did, so ..... ??
Someone else with more detailed experience than me my may "chime in"
with reasons, but all I can say is that my "NPN" version works and "sounds" better than both my other "PNP" circuits .... !!
I can also power it off a regular 9v centre negative PSU :D
Cheers,
Marty.
Yep, my Java Boost and a homemade Easy Face (with negative ground), made a hissing noise when powered with the same daisy chain with the Digitech Digidelay. So despite the negative ground, i still couldn't use the same power supply.
I'm not certain this applys to Rangemasters but the FF's I've tried on PS's ... for about two seconds....noise, bad.
QuoteYep, my Java Boost and a homemade Easy Face (with negative ground), made a hissing noise when powered with the same daisy chain with the Digitech Digidelay
That hissing sound is likely to be the audible artifact of ultrasonic or radio frequency oscillation.
I say this every time this comes up, yet it keeps coming up:
If your PNP germanium effect works nicely by just swapping it to negative ground, you are extremely lucky. I'm pretty good at messing about with electronic circuits and there are some instances I never could get to work in spite of some fairly sophisticated debugging, decoupling, and layout tricks. I don't know of any way you can be certain that it will work when you get done.
Sorry to start the PNP/negative ground discussion. I should have searched the forum.
Any thoughts on my other question about the emitter follower?
QuoteIn comparing the Java Boost to Tonepad's Neg Ground Rangemaster, there is a difference I'm confused about. The signal out of the transistor comes off the emitter on the Java and the collector on the Rangemaster. I would have thought that the Java would act as a buffer (emitter follower), rather than an amplifier.
I'd never looked at the schematic before. I see why you're confused.
I'll look at it some more but the first impression I have is that the schematic shown is simply not the same as the circuit they ship. That's obviously true in the trivial sense as the schematic states that the three way switch circuit is there, but it's not shown. But I think it's more than that. I don't think the circuit as shown works as other than a variable buffer, as you suspect.
I think a more correct circuit would be to cut R2 and the 10K audio pot loose from +9, and the 5K and 100K pots loose from ground and flip them so that the 5K and 100K pots connect to +9 and then connect R2 and the 10K pot to ground. At least then the circuit would work.
The schematic as shown may be an example of misinformation or an accident. I can't tell. I would not trust it until I compared it part for part and wire for wire to a working unit.
Amazing - I never knew exactly how close to my original work the Java boost was, never looked. Sigh. I guess I should spend more on advertising.
Quote from: R.G.QuoteIn comparing the Java Boost to Tonepad's Neg Ground Rangemaster, there is a difference I'm confused about. The signal out of the transistor comes off the emitter on the Java and the collector on the Rangemaster. I would have thought that the Java would act as a buffer (emitter follower), rather than an amplifier.
I'd never looked at the schematic before. I see why you're confused.
I'll look at it some more but the first impression I have is that the schematic shown is simply not the same as the circuit they ship. That's obviously true in the trivial sense as the schematic states that the three way switch circuit is there, but it's not shown. But I think it's more than that. I don't think the circuit as shown works as other than a variable buffer, as you suspect.
I think a more correct circuit would be to cut R2 and the 10K audio pot loose from +9, and the 5K and 100K pots loose from ground and flip them so that the 5K and 100K pots connect to +9 and then connect R2 and the 10K pot to ground. At least then the circuit would work.
The schematic as shown may be an example of misinformation or an accident. I can't tell. I would not trust it until I compared it part for part and wire for wire to a working unit.
Amazing - I never knew exactly how close to my original work the Java boost was, never looked. Sigh. I guess I should spend more on advertising.
Should the tone pot's wiper connect to corrector instead of emitter also ?
I think your article is responsible for a large majority of the Rangemaster clones out there.
This Java (http://aronnelson.com/gallery/Boosters/Java) schem has been confirmed (or at least laid out) by Torchy. Although this appears to be positive ground with the OC44 flipped. I have trouble reading vero layouts without the schematic, but that leads me to think the schematic on Keeley's site is not correct.
So Keeley basically adds pull down resistors, an input cap selector and a low pass tone control to the output. Oh, and brilliant marketing.
Ah - that layout has been built and confirmed by me and 3 others on the forum. I also have the corrected schematic which I'll post if you want. The schem released by Mr Keeley (on his website, its in the Javaboost manual pdf) is most definitely wrong.
Quote from: Torchy. I also have the corrected schematic which I'll post if you want.
Yes Please! :)
Hmmmm I did a bit of a search regarding this because I built a PNP BMP and I'd love to convert it to neg. ground. Is the coversion to +9v problematic for all circuits or just for those with germanium transistors?
QuoteHmmmm I did a bit of a search regarding this because I built a PNP BMP and I'd love to convert it to neg. ground. Is the coversion to +9v problematic for all circuits or just for those with germanium transistors?
I just keep telling people this and it just keeps getting ignored.
There is a problem with the trick of using circuits intended for positive ground on a negative ground system the simple way. This way involves tying the negative battery lead to the old ground and then tying the pulldown resistors to the new negative ground, trusting that the innards of the circuit will work fine with their old "ground" at +9V and the old -9V power supply at the new zero volts.
I'm not sure exactly where the problem lies. I have various theories, involving the impedance of the power supply now being in the ground reference lead, capacitor ESR, things like that. I have messed with some circuits that had problems with this conversion and while I could fix some of them, there were some that I never could get to work, and I'm actually pretty good at getting things to work. It may be that most simple circuits have good(-ish) rejection of noise on the intended power supply but poor or nonexistant rejection of noise on the intended ground wire. It may be a combination of all the oddities, and may well be different ones or combinations of the oddities to different circuits.
In theory, the power supply swap should be fine and dandy. The theoretical view of the circuits says that it will work fine, first time and every time. And I *like* theory...
But this is one place where I have bowed to Mother Nature's insistance that the situation is really complicated. She's trying to teach me something that I can't comprehend, even after long study and a really, really good background in solving things exactly like this. So until I can tell people what's going on and how to do it right every time, I don't recommend doing it.
What hooks people into this is that sometimes, maybe even most times it works. Many people do get a successful swap of the power supply, and in their experience I'm blathering nonsense. It worked the one or two times they tried it, so what's the big deal?
The big deal is that it doesn't work well sometimes, and a beginner's project is no place to do things that don't work sometimes. A beginner with one of the non-working exceptions is lost. The only thing to do is to tell them to put it back the way it was. This seems to be a 100% cure.
What follows is my personal opinion.
That being the case A NON-EXPERT EFFECTS BUILDER SHOULD NOT DO THIS AT ALL unless they're willing to just say "oh, well, it's busted, that one didn't work" and toss it out.
Experts (either certified, or self made) presumably know what they're doing, know the chances they take, and ought to do as they like, but being big boys, they should not whine about it not working, and ideally they should not tell beginners to do it until they can demonstrate how to make it work 100% of the time, every case, every instance, and preferably explain why the non-working cases didn't.
But I'm calmer now. Typing does that.
For your question - is it just germanium or all circuits? It's all circuits. Just *try* running a vacuum tube amp on -300V and ground.
Opamp circuits are reasonably immune. I think that this is because opamps usually have great power supply noise rejection. Their noise rejection is usually better for the + supply than the - supply for many opamps, but it's larger than no rejection at all.
No rejection at all of noise in the normal ground side is typical of common emitter and common source circuits. For PNP germanium, common source is the positive supply, as it is for silicon PNP. For compound circuits, mixtures of NPN and PNP, one or the other side is a preferred ground, and it's often the NPN side.
Case in point: the jordan bosstone. I built one of these for a friend, and do to a unique set of circumstances I needed to swap the power supplies. No problem - swap them, make some housekeeping changes, and voila!
Yeah, voila. It oscillated at 53MHz. So I set to work to fix the bugs. Three days later I gave up. When I reconverted it to negative ground it worked fine, first time.
The cursedness of this particular trick is that it requires that the circuit and the power supply meet a set of subtle criteria, and worse, that it's OK a lot of the time. It's like the guy who makes a zillion dollars buying real estate for no money down, then writes a book to tell other people how to do it because, look, it worked fine for him. The works-fine-mostly nature plays to the human tendency to think that if it worked for me it will always work.
And that's true.
Most of the time.
For your problem, do this: convert it the hard, but sure way. Replace the transistors with NPN's, flip the polarized caps and diodes end for end, then reverse the power supply. THAT technique works 100% of the time.
I've built a few things with Ge PNP transistors and negative ground - my Tone Engine boosters are simple common-emitter single-transistor full-range boosters, and they work great on positive ground. In converting from one to the other, you have to keep POWER ground separate from SIGNAL ground. It's not hard to do, but it does take some work. In my case, I was able to manage more easily because I designed it from scratch to work on negative ground.
There is a very easy way to make the switch: Maxim / Dallas and others sell a line of "charge pumps," which can act as voltage converters. I have used the MAX1044 (10mA) to make -9V from +9V in the past, and if you need more current, the MAX660 (100mA) will do nicely. The app note on the 660 says that max input voltage is 5V, but I've got one here running an MXR Flanger from 9V that's been fine for two years. :) I know R.G. has a schem for a 1044-powered inverter on Geofex, but I'd be happy to provide my PCB layouts for these if asked.
Matt Farrow
(http://img.photobucket.com/albums/v220/veroboard/Keeley%20Java%20Boost/Keeley_Java_Boost.gif)
RK here,
Didn't mean for the mistake in the schematic. It has been pointed out before and I neglected to change it. I will get that changed.
Sorry to cause RG and others some added stress on this subject. I too think that the noise present in some units tested before they are corrected is related to the fact that PNP transistors don't like the signal ground of the power supply. In theory it works, but not every time..so we have to work at it.
We do a lot of testing to make sure that units shipping don't have excessive noise when used conjustion with other effects and common supplies.
The smart folks here have brought up some great suggestions.
PharaohAmps, I get the idea of separating the signal ground from the power ground as concept but when I look at the Rangemaster circuit I can't really see how the two could be separated? Maybe you could elaborate?
Some ways to detach the audio ground from the supply's ground are using opto-couplers or differential input opamps and then floating the ground input. I was not able to get the same sound out of the pedal when using the opto-coupler. Transformers.
Also, I think you can just use the case for the audio ground, and through a pair of shield cables (with the ground connected at only the input and output side, not the board side) connect the audio. The power is not connected to the case, but the case still acts as an RF shield. Hope that helps.
(http://www.robertkeeley.com/manuals/Java-Boost-Schematic.jpg)
R.G. Sorry if you had already answered my question but for the life of me I couldn't find a comprehensive answer in the archives. It was usually a string of contradicting opinions. I should have taken a moment to think about what I was asking and I would have realized reversing power supplies was a problem for all circuits(that was a real slap the forehead moment). Thanks for your reply.
Some day I'll get the time to write up "The Technology of Grounds".
Grounding is confusing for two reasons.
1) In considering grounding, there are no "conductors". Every single physical bit of metal, resistor or semiconductor that the ground current travels in is either a resistor or has a junction voltage that shows up back up where you thought you had "ground". The idea that any two points can ever be at the same potential is fundamentally wrong when considering ground conductors with current flowing through them.
2) We all learned to think in terms of voltages, not currents. It's hard to retrain your mind to visualize where the current is flowing and why. But you'll keep having grounding problem until you do.
Just to set a couple of items straight here:
While it is possible to do a power ground net and a signal ground net and connect them at only one place so that you have some chance of keeping them separate, this is impossible to do in a simple common emitter or common source circuit. It's impossible because the voltage that makes the active device amplify signal is the base-emitter (or gate-source, or grid-cathode) voltage. The power path flows from collector/drain/plate to the emitter/source/cathode. So the two are inherently connected at the emitter/source/cathode, and can't be separated. This is why simple gain circuit have no, naught, nada ground noise rejection. They can't.
This is not so bad if you have only one stage running into that wire on its way to the great ground sewer grate in the sky. Because the signal at the input is a larger or smaller version of the current in the ground, there is no noise or feedback possible in this single stage. But what happens if you have on ground wire serving two inverting stages? The first stage inverts and amplifies, the second stage inverts and amplifies, and a miniature replica of the second stage signal is delivered right back to the input of the first stage by the resistance of the "ground" wire. And it is in phase with the incoming signal. With a little bad luck in impedances and gain, you have an oscillator, either full time or just at odd moments; often ultrasonic or RF.
That being the case, it is incredibly crucial that any OTHER power users not try to stuff their ground current down the same wire as a common emitter/source/cathode gain stage. That's why all these people have problems with clicking when their LED uses the same ground wire. That's why running all your ground current down a wire to the input jack for switch is a setup for ground problems.
Likewise, it is impossible to separate signal ground from power ground with optoisolators or transformers. Inside the amp, signal ground and power ground will ultimately have to be connected together. There is nothing you can do to prevent that. The critical thing is to make sure that where they connect, the physical wires that carry signal ground do not carry non-related stages' power ground current. Think about where the current is flowing!!
Differential stages solve this a different way (sorry...).They are almost completely insensitive to what is happening on their own ground. Transformers also are a way of ignoring their own ground, as they operate only on the difference between the two leads. A grounded centertap is all a transformer needs to be true differential and have a lot of ground noise rejection.
You really should not disconnect the metallic enclosure of an effect from signal ground. It should be tied to one and only one of: input jack ground, output jack ground, designated ground point. If it floats, it can capacitively couple noise in. If it connects to more than one point it can be the conductor of ground currents and cause problems. For low gain pedals and small enclosures, you get away with both input and output jack grounds tied to a metal enclosure, but that's because they're physically very close and the resistance is small. There's a wider window to get through. There are instances where this will not work and you will have to use at least one isolating jack.
Thanks for the advice and information RG, can't wait for a great Technology of Grounds paper!
rk
What I meant by "separating grounds" is that the audio signal gets referenced to ground (here tied to the negative battery terminal) and the power for the transistor is all referenced the other way. The transistor's collector is tied directly to the -ve (ground) through a resistor, and the biasing for its emitter comes from the +ve supply, also through a resistor. These biasing currents are all DC, and are basically the same as in a positive ground setup. Any AC signals are capacitively coupled in, and can be referenced to the -ve (ground) because the DC currents flowing in and out of the transistor are blocked by the capacitors.
I've got two designs that use PNP in a -ve ground circuit, one of which is probably suitable for release now. I'll put the board layout, parts placement, and schematic on my site later on today (hopefully) and link to them from here.
The other one is kind of still on the breadboard right now, although I do have a PCB layout for it - it's an op-amp input, PNP transistor stage, then an op-amp output. -ve power is provided by a 1044. I am probably going to be releasing this design in conjunction with TonePad, since they're really set up to answer questions and provide support to the DIY community. Stay tuned.
Matt Farrow
Does the GEO layout for the Netron use a separate signal and power ground?