Neovibe issues

[Stevpo]

Hey all, just thought I'd share some issues I'm having with a brand new neovibe build in hopes someone might be able to help out.  It seems everything is normal with regards to voltages, and a properly oscillating LFO section, but I'm having some trouble with the input transistors.  I'm not sure how to get this part debugged, but here are my voltage readings:  Q1 - C: 1.31 V, B: 0.68 V, E: 1.34, Q2 - C: 5.72 V, B: 1.31 V, E: 0.75 V, Q3 - C: 10.14 V, B: 5.72 V, E: 5.1 V.

The light pulses normally, and responds to the controls.  All other transistors have voltages that indicate they are operating normally, and I verified with a meter all resistors associated with the input amplification section.  From the looks of things, Q1 looks to be problematic, but replacing it does not help.  Any ideas?  Thanks.

[jasperoosthoek]

The base voltage of Q1 is lower than the emitter voltage. That's weird, it should be around 0.6 volts higher for an NPN transistor. That doesn't mean it is the fault, it means it is not properly biased by Q2. Q1 might be just fine.

The emitter voltage of Q2 is way to low as it also has to bias Q1 through the 1.2Meg resistor. Possible causes: the 6.8k emitter resistor to Q2 might be the wrong value (680R??), the cap across it could be shorted, or the 1.2Meg resistor to the base of Q1 could be wrong/bad solder joint. Try looking at Q2 first.

[Stevpo]

Thanks for the reply.  Unfortunately, all the things you suggested check out okay.  I replaced Q2 as well.  Every transistor behaves the same way in that position.  I am stumped.  If it helps, I removed Q3 from the circuit and these are the voltages I read then - Q1: C: 1.59, B: 0.46, E: 0, Q2 - C: 8.15, B: 1.59, E: 1.03.

[R.G.]

First thing to check is whether C1 is inserted the wrong polarity.

Measure the voltage across C1's pins, and see if the more positive side matches the pin that is supposed to be (+), and vice versa.

[Stevpo]

The side connected to R4 is more positive than the side connected to R3.  There's about 0.46 V across the cap.

[jasperoosthoek]

R.G. found the problem. The voltage across the C1 should be the same at the emitter voltage of Q2. The cap is reversed and starts conducting. Consider it broken. Replace the cap and put it the other way around. No matter how you connect the cap, the side of R4 is always more positive than R3.

The negative polarity of the cap should be connected to R3. The negative side is indicated by the shorter lead wire (when new) and also the white arrow with minus signs on them:
http://en.wikipedia.org/wiki/File:Capacitors_electrolytic.jpg

[Stevpo]

Sorry, but that was not the problem.  I double and triple checked that cap.  It was definitely in the right way around, but turned it around just to see, and there was no change.

[R.G.]

Closely check your soldering to be sure there are no solder threads shorting adjacent pads.

Also , pull that capacitor OUT and see if the voltages change.

Which version of the PCB are you using? Is is by chance one you got off the internet? There was an early version (that is stored in zillions of places, so I can't fix it) that has C1 shown reversed. Is it a GGG Neovibe board?

[Stevpo]

The version of the board is a self-etched one from the plans on the GGG site.  Checked the soldering visually and with a meter, and there were no connections that aren't supposed to be there.

[R.G.]

The version of the board is a self-etched one from the plans on the GGG site.  Checked the soldering visually and with a meter, and there were no connections that aren't supposed to be there.

OK, pull that capacitor out entirely and see if the voltages change. The reason I harp on this is that the voltages you show match exactly with what happens if the cap is shorted or reversed.

[Stevpo]

Sorry.  Meant to mention this.  I did pull the cap.  Made sure there were no other shorts between the traces.  Still, the voltages I read on those transistors are the same.

[R.G.]

I spent some time on the simulator, trying to find a condition that would make those voltages come out. Closest I got was with Q1 shorted.

At this point,if it were mine, I'd pull out Q1...Q3 and the electros into/out of that section, and use the meter to check the PCB etching and soldering for shorts/opens on each trace and its neighbors, as well as metering all the resistors again.

There are only a few possibilities. Solder and wiring are by far the most likely, followed by wrong component value or orientation (what transistor type did you use, by the way?) and then on down into PCB etching problems and far, far down the list a faulty new component.

[Stevpo]

Thanks for spending the time trying to help me out here.  This is actually my second neovibe build.  My first one I had working and sounding really good.  After about a year, it developed a problem I couldn't figure out (voltages were spot on, and it worked, somewhat, but intermittently), so I decided on a complete rebuild with new parts.

That being said, I did as you suggested.  All resistors check out, caps were inserted the right way, there are no visible errors in the etching, and there is no continuity where there isn't supposed to be.  Transistors are 2N5088 except for the lamp driver which is a 2N3904.  But like I said, the oscillator is working fine and the lamp pulses.  It just seems to be the amplification section of Q1 through Q3 that's giving me problems.

[R.G.]

That being said, I did as you suggested.  All resistors check out, caps were inserted the right way, there are no visible errors in the etching, and there is no continuity where there isn't supposed to be.  Transistors are 2N5088 except for the lamp driver which is a 2N3904.  But like I said, the oscillator is working fine and the lamp pulses.  It just seems to be the amplification section of Q1 through Q3 that's giving me problems.

Here's the deal - it's not possible for all of the following to be true at the same time:
(1) All correct parts values
(2) All good parts
(3) All parts inserted the right way round according to pinout and in the right places
(4) Correct supply voltage on the board
(5) Correct off-board wiring and connections to other parts of the PCB
(6) No extra continuity on the board where it should not be
(7) No opens on the board where they should not be
( All good solder joints
(9) The board not to work.
The only thing that's mysterious about any of this is that there can be flaws that are hard to see with the naked eye. Solder threads and unetched threads of copper between traces can be almost invisible. The proof is in the meter: as I said:

At this point,if it were mine, I'd pull out Q1...Q3 and the electros into/out of that section, and use the meter to check the PCB etching and soldering for shorts/opens on each trace and its neighbors, as well as metering all the resistors again.

I didn't say that casually. I have found copper threads invisible to less than 30X magnification with a meter before. It also helps to NOT trust a solder joint. This can be checked by measuring resistance from the copper trace to the component lead, and not the solder joint. I have seen solder joints that look good, but do not make contact to the actual lead. It is also possible for a resistor or cap to have a crack in it, disconnecting a lead from the body. These can only be found by metering from copper through solder to lead, through the component, through the lead, through the solder and to the copper on the other side.

I'm back to the Mother Nature concept. Mother Nature insists on Her rules, and does not care what we think. But the good thing is that once Her rules are followed, She **makes** it work.