Obsidian Overdrive problem - still not giving up

[robbiemcm]

Okay, yes.. I don't like to be beaten by things. I've been trying to do this thing for like 8 months now, I've made it three times. It works on breadboard, and I've had exactly the same problem with both PCB builds. So I decided I'd try to fix this pedal again, and I still have no idea what is wrong with it. It squeals and the knobs can vary the pitch at which this happens. I built the Obsidian Overdrive with 2N7000's and switched the battery round as it says so its now negative ground. I also flipped the electrolytic capacitors. The transistors are the correct way round. I took out the switch (in case you were wondering what the stray wires are in the picture below) because I got the pedal to work on breadboard and I did that without switches or with the pots, but i havn't bothered to take them out.

I tested my circuit with an audio probe and the squeling noise appears to be throughout the circuit, not at any particular point, even at the point where the signal enters the PCB.. so that isn't very helpful. The only possible information leading me to a suggested problem is that Joe Davisson couldn't get the same voltages as me for the 2N7000's. I can't tell if I fried them or what, so I'll supply what I can below in the hope that you can find out the problem.

If I did fry the 2N7000's then I can't do a better soldering job than what I have done, I would like to have some sockets to just push them in to (this theory would explain why the circuit worked on breadboard) but yeah.. I couldn't find anywhere to buy something like that, I'll need to try and find something to make some myself, any suggestions? That's if that's the problem, I don't even know if it is. Jebus, I just have no idea what I did that made it so this thing woudn't work. Nobody seems to know. Sorry I couldn't get a picture closer to the PCB, the only digital camera we have in our house doesn't have a manual focus and it couldn't take anything of use.





In order from left to right according to http://www.diystompboxes.com/analogalchemy/pedals/obsidian.gif

S: 5.43
G: 3.20
D: 8.62

S: 5.43
G: 3.26
D: 8.62

S: 5.43
G: 3.19
D: 8.62


If you can solve this problem you will actually be my hero, if I don't fix this one I don't think I can face building another one. Chances are I screwed up and did something stupid, but I can't find that problem. If you spot anything odd, please say so, and please interrogate me. This was my first attempt at building a pedal (actually second, but it was the same pedal because the first one didn't work) but no matter how stupid something seems, how bad I'd have to be to make that mistake, please make sure I didn't do that one and just ask.

Thankyou for your time if you bothered to read that and thanks also if you decide to help me out.
Robbie

[Joe]

The gate voltage is near where it should be. That's a good sign. But it shouldn't be lower than the source voltage. I'm inclined to think the drain resistor value is too low for your particular batch of transistors. Try raising the drain value in steps (4.7k, 5.6k, etc.) until the drain is somewhere around 4 - 5 volts. If one stage works, just use the same value for the others, since the transistors seem consistent.

[robbiemcm]

Thankyou, Joe. I'll try this today if I can find the correct resistors. So it seems building these things isn't as easy as following the schematic, I would have assume all the transistors would be similar enough so that you didn't have to do this. One thing that I must ask you, I will still try doing this anyway, but could that be causing the weird squealing/oscillation noise I'm getting?

Thanks again for your help, I'll make another post to tell you how it went
Robbie

[EDIT]

I just thought of one problem with this possible solution. The circuit worked on the breadboard, although I didn't test the voltages on that, surely that would rule out the odd batch of transistor theory? I used the same transistors on the breadboard and it sounded perfect. Back to the drawing board unless I'm wrong with that statement? I would have assumed fried transistors would do that, but the voltages for SGD are still so consistent, would I not have to have fried them in pretty much equally to achieve that?

Gah, I have absolutely no idea of this problem.

[Joe]

Forget the transistors, I can tell from the gate voltages that they probably aren't blown. Breadboards aren't perfect, and maybe there was enough leakage bias to prevent the transistors from oscillating. Actually, it's not uncommon for circuits to work on the breadboard, then oscillate on the PCB.

[robbiemcm]

Alright thanks, I put off trying it today as I wanted to make sure of that first. I'll give it a try tomorrow though, for the rest of tonight I need to accomplish my goal of getting through another 50 pages of Lord of the Rings I'll tell you as soon as I can how it went.. and come to think of it, I'm not sure if I have the appropriate resistors. May have to go buy some in the morning.

[EDIT]

Oh, one last question to leave here. In the future when I'm hopefully making circuits of my own (in other words, in a long time) if it works on the breadboard, but not on PCB.. would you have measured the voltages when it's on the breadboard so that you could adjust the resistors on the PCB version if that happened to be the problem. Because you would have no point of reference as to what the voltages should be if you're designing it yourself. Just wondering..

[robbiemcm]

Alright, I'm pretty sure I didn't do anything weird to it, I bent a few capacitors to get a look at the resistors, surely this didn't do anything? Anyway, it seems the voltages have now changed dramatically for the transisotrs, battery doesn't appear to be flat. The gate voltages are now considerably different aswell, and considering you said that they were around where they were supposed to be, I don't think this is fixed. It's midnight now, and my sister is studying for exams, so there is no hope of me going up and testing it out. The source is now lower than the gate (as you said it should be) even when it's varying, it always remains lower. So I would have thought this would definitely be a turn for the better if you hadn't said the gate was meant to be around 3 point whatever. The drain is now also between 3 and 4, which now puts it on the opposite side of the 4 - 5 mark. This is just confusing, I'll test it out tomorrow morning and tell you whether it's working, although I highly doubt it. Probably just oscillating differently now.

S: 1.18
G: 1.27
D: 3.77

S: 1.15
G: 1.28
D: 3.85

S: 1.08 - 1.21 (Gain turned down to zero - Gain turned up to full)[Discovered this by chance, others transistors don't vary.. should this happen?]
G: 1.28
D: 3.61 - 4.00 (Same format as with the other thing above)[Again, should this happen]

I just can't think why it would have changed, unless there was some bad solder joint that has moved [in to/out of] place now. If there is, then I don't know how you are meant to avoid such a thing, my soldering seemed fine. Wow, I can't believe how incredibly frustrating it is.

[Joe]

Cold solder joints are maddening, I bet that was the problem. In the end you may want to lower the 3.9k's to 3.3k's to kick it up a bit.

[robbiemcm]

How does one avoid the outcome of or test for cold solder joints? I didn't think I did anything wrong when soldering Unless the problem is I havn't been using that flux stuff or whatever it is? I'm off to test it now.

[EDIT]

Briefly, in other news: my volume pot doesn't do anything. I'm sure I must have done something stupid there aswell though, I'll see if I can fix it. It works more like a tone pot right now

Alright, the squealing noise has gone. It has the most background noise I've ever heard though, equal to as loud as the guitar is heard, although it seems to die away when the gutiar is played. But I'm pretty sure it wasn't like this in breadboard. Although it seems as though we have the problem. As soon as I find out how to test for and fix these bad solder joints (and prevent them hopefully for the next project) it might work better. Plus I'm going to replace one transistor which is quite bent, the gate is really hard to test a voltage for further up the metal bit, but I can get it fine down the bottom. All the rest are fine, so if I replace this and fix the solder joints we may have solved the problem. Good thing the voltages made a weird change when I wiggled a few components around.

[robbiemcm]

I'm finding it hard to identify these cold solder joints. Do you think it might be the capacitors, considerin they are the ones that I moved out of the way? And to fix these cold solder joints I just reheat?

And the last thing, would someone mind telling me how your supposed to avoid cold solder joints in the first place.. or is it just a fact of life?

[Steben]

1. Try making movement of the lugs and legs impossible when soldering.
2. Heat the PCB first for a few seconds then make contact with the legs.
3. Heat the POT/JACK lugs for a few seconds first then make contact with the wires.
4. Check for glossy shine after the soldering action. If it is dull, resolder!

[johngreene]

Let's see.....

For starter's, the Gate voltage should be around 2.8V. 1.5/4.8*9 = 2.8125
A fresh battery will result in a slightly higher voltage so your first measurements of the Gate voltages looked 'more right' than the second.

The source should be around .8V lower than the Gate, or slightly more. You measurement shows it higher in voltage which indicates to me the MOSFET has to be in backwards. If the Gate and Source were equal the transistor would be off so this state just can't happen unless it's in backwards and is conducting current through the protection diode, or the device is damaged. However, if the circuit is oscillating, you will not get accurate readings with a meter. Check the leds of the MOSFET. You should get conduction in the direction of the protection diode but not the other way around. But I would probably guess that these measurements were inaccurate due to the oscillation...

The second set of measurements you made where the voltages look 'relatively' correct, just lower than expected. I'd check the gate bias resistors to make sure that they are the correct values and wired correctly. Again, isolating just one section will help narrow down the problem. If the input capacitor was shorted you would get a gate voltage of .75/4.05*9 = 1.67V.

This little monster has a lot of gain and when you get past the second stage, the gain is non-inverting and can very easily oscillate. My advice for bringing this circuit up would be to do it 1 stage at a time. Since the 3 stages are pretty much identical, you can test each stage individually and then connect them together. Connect the output of the gain pot to the output jack and get that working. Once you get there, connect the output of the last stage to the output jack and test it. You would probably want to connect the volume pot to the output stage rather than just straight to the connector. Once that is working, then connect your tone control. This way you are incrementally connecting known good things together so if anything stops working, you know the section where the problem is rather than trying to tackle the whole thing at once.

--john

[robbiemcm]

Thankyou for those very informative posts. First I will try the cold solder joints, but if that doesn't do it I'll move on to isolating the individual parts of the circuit to try and get it working. Hopefully I will be able to fix this afterall. Before I build my enxt pedal I think I'm going to brush up a little more on some of the theory of transistor voltages and such, so that I don't have to effectively get you to troubleshoot for me. I should probably start with something easier though, this one has proved to be beyond me.

One thing I must ask you johngreene:

Check the leds of the MOSFET. You should get conduction in the direction of the protection diode but not the other way around. But I would probably guess that these measurements were inaccurate due to the oscillation...

Protection diode?

[johngreene]

Yes, MOSFETs usually have a diode between the Drain and Source with the cathode connected to the drain(N channel). Under normal operation this diode will always be reversed biased as the source is always at a lower potential than the drain. Makes it easy to check the source and drain pins. Look at the symbol on this data sheet http://www.fairchildsemi.com/ds/2N/2N7000.pdf it shows the diode.

When troubleshooting, divide and conquer is always the way to go. 

--john

[robbiemcm]

Does that mean I'm supposed to have a diode in my circuit, or I'm supposed to have a diode in there when I'm trying to measure the voltages or something? This has now got me even more confused. The schematic doesnt say there's a diode in the circuit

[WGTP]

First, we've all suffered thru this kind of thing.

The earlier Obsidians that used N channel 2N7000 MOSFET's had diodes to protect them from static electricity that can fry them.  The later one used P channel BS250's that Joe said were more quiet and less prone to fry.  It did not have diodes for protection.  (There were also some Obsidians using 2N5089's)  It appears your trying to use the newer one with the battery and electros reversed (which has been done succesfully).  You might try finding one of the older schems that used the 2N7000's if your problems continue.  I have built most of Joe's stuff (they are great) and it is not uncommon for me to have to change some resistor values to get the bias correct.  Joe probably uses new batteries and I like 8v.  Don't give up on Joe's stuff.  Try the Vulcan if you can't get the Obsidian cranking.  Hope that helps. 

[johngreene]

No, it doesn't mean you should have a diode in your circuit. It means you already do. The transistor has one already inside although it's rarely shown in the schematic symbol. My point was that using a meter in diode mode, you can determine which is the source and drain by testing this internal diode. Nothing more.

--john

[robbiemcm]

Okay thanks, I really will try the things tomorrow. I've just been extremely busy lately.