Intermittent Problems on Modded Pedals Question

[vmazz38]

Ive had great help with members on this board, so first off Id like to say thanks!

Ive been racking my brain debugging an intermittent Aron Sch-Z with EWS mods.
Its worked and then faded out intermittently probably 20 times, ive replaced all ICS and Transistors most electrolytics, dealt with lifted traces etc .
There seems to be no practical reason why it works or doesnt(obviously there is but..)

My question is not about this specific project, but in general has anyone dealt with chorus,flange etc of existing (factory produced) pedals that have such intermittent problems.

It will do 4 of the following things 1)function perfectly chorusing nicely 2)pass no signal when engaged 3) function partially with crackling and fading in and out of the wet signal 4) pass dry signal but no chorusing when engaged

Does this sound like a cold solder joint or faulty part? All IC and BBD have been replaced with known working parts.

Thanks for any advice.

[R.G.]

There is no particular weakness in delay circuits for intermittents. It's intermittent because there is an intermittent contact in there. If it was not intermittent before the mods, the mods or the modding processes are the most likely causes.

If it was a skills problem with soldering or removing/replacing parts, the chances of getting new problems introduced increase every time you replace something.

If you've had to replace lifted traces, chances are there are issues with solder joints and/or other trace defects you have not found yet.

I realize this sounds harsh, but it's realistic. I have been in similar situations before: modify something and either it gets worse or has a new problem. Sad experience has convinced me that what is most likely to be the problem is some of the most recent changes I did.

[vmazz38]

Thanks for the inpuit RG.
In this particular case the pedal was intermittent before the mods. I got the pedal cheap because it was passing signal but no effect .
Took it home reflowed the solder joints and it worked immediately. A few days later stops working fixed etc.
Then did mods. worked wonderfully ..then again not working.
It would work or not work if I had not even touched it. It made me suspect a BBD or IC was bad or going bad replaced them all.
Im just being persistent now because I dont wont to give up on it. Its just weird because now it will work and then just fade out to dry signal.
My solder skills are pretty good..but I have gone over the board so many times traces came up..Im pretty capable of fixing them just trying to figure out what it is going on.
Ive had several  builds or modded pedals that gave me problems, but not one that came and went so randomly and right in front of me ..And then work when I go over the board for a few minutes then go away again.

Could a bad mylar cap or electrolytic cause an effect to fade in and ou? The crackling noises it makes occassionally lead me to think a bad joint but I have really gone over it alot and it works sometimes too.
Just trying to check off all the possible causes.
Greatly appreciate your input though

[R.G.]

Sorry for having to start at the lowest common denominator of problems.

Given good soldering and repair skills, the problem then does shift to components. I'd say solder joints again, but you have reflowed them all. The crackling is an indication of a bad joint, as you suspect. The subtlety hidden there is that all joints are not solder joints.

Mechanically stressed components can develop stress cracks in the bond from the inner workings to the leads. It's uncommon, but does happen. I hesitate to mention these things here, just because beginners will seize on it as a bad component, when soldering and wiring problems are so very predominant; but it does happen. Electro caps forced into PCB spacing that's not the same as the capacitor package can stress leads just inside the cap body and make an intermittent; so can resistors and film or ceramic caps. Electro lead spacing issues are more common.

The problem with all intermittents is to identify something that you can do to make it happen. When you get to there, the problem is 90% fixed. Generally intermittents are activated by mechanical stresses or heating. Poking parts and wires one by one with a pencil eraser or chopstick can help find mechanicals; be especially mindful of the connections holding the board in the box, and of controls and wires that may short to each other or the box. I mention thermal because you say that sometimes it takes a while for the problem to happen. Long time delays are generally thermal.

Finally, yes, electros can be a problem. There was an industry-wide problem with electrolytic cap durability in the time shortly after 2000. It was a near-disaster to some small manufacturers, and a huge warranty problem to big ones. If you are skilled with a soldering, just yank out and replace all the electros.

[vmazz38]

The hardest part has been trying to make it happen, as you mention. I'm not really sure the cause and effect. Mark Hammer gave me some good pointers on this type of circuit , as to what components could effect wet and dry signal passing etc. I'll post pics when I get home.

[vmazz38]

Here are the schematic for the Arion SCH-Z and some pics of the modded board with repairs etc  (yes there are lifted traces that have been repaired ,all test good for continuity, the problem was happening before I ever put on iron on the board)

[R.G.]

From a glance through the schematic, I'd say to check the pins of the CD4027. Each pin should be solidly at V+ or ground.

Apparently this pedal changed out the bipolar flipflop from old Boss practice for a CD4027. That chip controls all of the signal routing through JFET switches. If it's not switching properly, the signals don't go the right places. Second issue - check the actual control signals at the gates of the switching FETs. An intermittent on any of them could cause similar problems.

It would be really great if you could catch it not working for these measurements.


Edit: Check the bias voltage supply too. Unrelated: check for pin 1 and 2 of the CD4027 going solidly to V+/ground and flipping when you press the engage button.

[Rob Strand]

You could short out the switches Q7 and Q4 with wire when the problem occurred, that would eliminate the switched.

Problem should isolate where the audio is dropping out by following the chain along:

- source of Q2
- source of Q3
- output of pin 1


If biasing of the delay chip isn't right it might be cutting-off the signal.  I would try adding a 470k across R15, and if that didn't work remove it and try adding a 470k across R16.  One of these might push the circuit back into the working zone.

[vmazz38]

So to clarify( what already should have done sorry). Part of the mods that I did where to make it True Bypass, so Q4 and Q7 are removed and jumpered and cd4027 is disconnected from the circuit. The latest strange symptom its giving me is sometimes I get a negative voltage reading -8.3volts at the dc jack, where I know the supply is feeding it 9.4 volts. Thanks for the suggestions Rob I will try to set the bias resistor.
Another weird thing is that on the schematic most of the trannies should be fets or mosfet, but when I pulled them they were c3198s most of them, so they where replaced with appropriate trannys and it definitely worked with the substitutions which where done very late in the debugging. Thanks for the help guys. 

[vmazz38]

One more idea..If the circuit heating up is causing the effect to come and go, what part of the circuit should I look at?

[R.G.]

So to clarify( what already should have done sorry). Part of the mods that I did where to make it True Bypass, so Q4 and Q7 are removed and jumpered and cd4027 is disconnected from the circuit.

OK, scratch that. Introduce wiring issues from the new switches. How did you true-bypass a stereo output?

The latest strange symptom its giving me is sometimes I get a negative voltage reading -8.3volts at the dc jack, where I know the supply is feeding it 9.4 volts. Thanks for the suggestions Rob I will try to set the bias resistor.

What does the voltage read from the input jack bushing to the Vcc pin on the opamps *on the PCB*, not at the DC jack.

Another weird thing is that on the schematic most of the trannies should be fets or mosfet, but when I pulled them they were c3198s most of them, so they where replaced with appropriate trannys and it definitely worked with the substitutions which where done very late in the debugging.

While 2SC3198 (if that's an NPN) would probably work for the four positions where the devices were used as source followers, this makes me very suspicious that the person you got it from did some boy-genius, BUMS mods that didn't work, then dumped the body. That moves everything to suspect.

Until you get the power supply to the board to be correct, you can't count on anything else to be. So go for getting the power from DC jack or battery to be correct first.

You're getting very close to me asking you to fill out the full bit of info in "Debugging: what to do when it doesn't work". It's a powerful debugging tool.

[vmazz38]

Thanks for the help. I will post some voltages per the debug list. later..there is definitely something going on with the voltages in the middle of the board.

[vmazz38]

So I only was able to test part of it but its giving me some clues. I hope all the transistor CBE are associated with the right legs as the schematic shows FETs but the board had npn trans
Q1 c-3
     b-6.6
     e-2.7

Q2 c-2.9
     b-6.8
     e-3.5

D1 a=-.7
     k=0

D3 a=6.5
     k=6.1

IC1
1=3.1
2=3.1
3=3.2
4=0
5=6
6=3.4
7=3.5
8=3.5

IC2
1=.08 to 2.5 fluctuating
2=2.8
3=1.4
4=0
5=6.5
6=4.8
7=.2
8=3

So the big clue was on Q1 the Base would drop to .2 intermittently and the Collector would drop to 0. If I pushed on the board it would go back and forth. But I have gone over those joints many times. Wondering about the D1 voltages. Is it normal to have negative voltage there?

[R.G.]

So I only was able to test part of it but its giving me some clues. I hope all the transistor CBE are associated with the right legs as the schematic shows FETs but the board had npn trans
...
So the big clue was on Q1 the Base would drop to .2 intermittently and the Collector would drop to 0. If I pushed on the board it would go back and forth. But I have gone over those joints many times. Wondering about the D1 voltages. Is it normal to have negative voltage there?

You're well on the way to solving it. You have figured out how to make the intermittent happen when you want it too. Well, you're on the way to solving one problem. Others might be hidden underneath it, but this *might* be the last one.

Let's think about the schematic and how the circuit should work. The collector drops to zero? It's hard-wired to the Vcc supply. Either the collector is coming loose from Vcc entirely, or Vcc is dropping to zero, at least out on the end of the trace that connects to that collector. One way to figure out which (now that you can make the problem happen!) is to measure Vcc not on that collector lead, but on the solder joint on the bottom under the collector lead. It will either go up and down with the collector or not. If not, there is a break happening between the solder you're measuring on and the collector lead. If it wiggles up and down with the collector when you make the fault happen, then there is an issue somewhere along the trace toward the power supply. Follow the trace back toward the power supply until it quits going on and off, and then back up one. The break is between where it stops and starts. Or start at the power supply jack and work inwards toward the collector lead. In either case, find where the Vcc behavior changes. Track it down, then you can kill it.

After finding that, look at the change on the base/gate. That may or may not still happen. Since the base/gate voltage is generated from Vcc, there is a chance that fixing the Vcc-on-the-collector problem may fix the bias too. Maybe not - it could be the next layer of the onion.

[vmazz38]

Greatly appreciate your help RG! Hope I can do some circuit sleuthing tonight!

[vmazz38]

So I think I cured the intermittent at Q1. But now im having the recurrent low voltage at the dc jack. At the junction of D1 and R1 it reads (neg)-.8   at the dc jack side that connects to R1 it reads  -5.2 , at the positive side of dc jack it reads 4.1. The supply is puting out about 9.4 volts. But the circuit is only getting +4.1 volts and not functioning in this state. Trying to figure out what would cause this. Dragging down the dc voltage 

[vmazz38]

(if anyone can help with the schematic in this post)
Question: Im trying to figure out how in the schematic the power supply gets to ground. Above the c3 node it says bias.
What are these bias points?
Do any of them go to ground?
If I ground out the D1 R1 node I get a full 9.4 volts where it should be at the jack.
Just trying to figure out what point is not going to 0 when it should.
Forgive my ignorance.

[R.G.]

Question: Im trying to figure out how in the schematic the power supply gets to ground. Above the c3 node it says bias.
What are these bias points?
Do any of them go to ground?
If I ground out the D1 R1 node I get a full 9.4 volts where it should be at the jack.
Just trying to figure out what point is not going to 0 when it should.

Let's assume for a moment that the circuit is really what the schematic says it is, and see what that means.

The DC jack has a battery clip attached to it, on the momentary-switch side. So the + side of the battery is normally connected to Vcc. The minus side of the battery is only connected to circuit ground through diode D1 when a plug is inserted into the input jack, by the "ring" contact. The negative side of the DC jack (the center pin) is connected to circuit ground through a 150 ohm resistor and D1 as well when there is a DC power plug inserted.

What is supposed to happen is that the + side of either a battery or a DC power adapter goes directly to Vcc, while the minus side is always one diode drop or one diode drop plus the DC current times 150 ohms below circuit ground. So the DC jack negative will and must be 0.6V or more below ground. The "or more" is the circuit current times the 150 ohm resistor (which suspicious minds might think is not really 150 ohms, but something else - the pedal has been modified).

This is the very old Boss single-DC-supply circuit, which assumes each pedal will have its own DC supply, and that it will not have to share both -DC and signal ground with other pedals. The diode gives polarity protection, and I'm not sure what they wanted the 150 ohms for. Modern pedals connect the -DC to circuit ground directly and do polarity protection differently.

The way to debug this is to use your voltmeter and measure the DC voltage across the pins of the DC jack, which should give the full DC adapter power supply, 9.4? 9.5? volts in this case. The sum of voltages in the loop from Vcc through the pedal, through D1, through the "150 ohm" resistor must add up to be equal to this measured DC in. This is a specific case of the loop voltage theorem - the sum of voltages in any loop has to be zero. Taking the voltages across each component in the loop separately will quickly tell you where the voltage is going. Once you know that, you can reason why.

[vmazz38]

First I am thankful for your help.Thank you for your time RG.
Im not sure what all is in " the loop" you dont mean every component in the circuit do you? Every one in the dc supply? I find that all the components in that general area add up to 9.4 roughly. When I jumpered across that 150 ohm(it was actuallyy 120 ohm  because I replaced it when I was troubleshooting) I got the full 8.5 roughly from dc plus to ground(more what I expect from it) and the dc neg was about .85 as you predicted. But with the jumpered across R1 the circuit didnt pass a signal.

[R.G.]

Im not sure what all is in " the loop" you dont mean every component in the circuit do you? Every one in the dc supply? I find that all the components in that general area add up to 9.4 roughly.

A bit of theory, then. Electricity travels in complete paths. This is why we call them "circuits" - the electricity comes out of a source of electrical power, like a battery or power supply, goes through one or more components (or lumps of components) and then flows back into the source of electricity, completing an entire loop.  If you take the voltage across each component in the loop and add them up, including the proper sign convention, the voltages will add up to zero. Which makes sense: this is just another way of saying that the sum of the voltage drops across the components receiving power must be equal to the power supply driving it.

This is kind of obvious for tinkertoy circuits of a battery, a resistor and a diode or two. However, it also holds true for every "hole in the middle" of a schematic ringed by components, which is not so obvious. In the schematic you posted, for instance, the path through R12 to Q2 gate, to Q2 source, through C9 and back to R12 is a loop. The voltage around it must be equal to zero. It turns out that this is a bad example, because C9 does not conduct DC, so there is no current flow in the loop. The voltage through R12 and Q2 may have current flow, but C9 will have that same voltage. But the sum taken around the loop (of which this is one) is still zero. If we replaced C9 with a resistor, some current would flow through that position, but it would be equal to the voltage across R12 and Q2 - which points up one of the beauties of the idea - we can compute component voltages and currents in local loops.

This theorem is a companion to the node current theorem: the sum of all currents entering and leaving and leaving a node muse be zero. A node is the junction of two or more components, and the theorem is just the pontification of the idea that the dot of solder joining two components can't store electrons. The loop and node theorems let you write loop and node equations and solve the whole mess for every voltage and current in the circuit however complex. I know this from sad experience, because after the professor told us about node and loop equations, he sent us home with homework to write and solve circuits for all the node voltages and loop currents.

But I'm wandering. A loop is any closed path in the schematic.

When I jumpered across that 150 ohm(it was actuallyy 120 ohm  because I replaced it when I was troubleshooting) I got the full 8.5 roughly from dc plus to ground(more what I expect from it) and the dc neg was about .85 as you predicted. But with the jumpered across R1 the circuit didnt pass a signal.

Then you've found that there are other issues. An interesting thing would be to simply measure the DC across the 150/120 ohm resistor. If that's more than a fraction of a volt, then something on the main circuit is pulling so much current that almost all the battery voltage is being dropped across the resistor.