Possible damaged chip?

Hardware questions and issues with the FV-1

Moderator: frank

djmalan
Posts: 36
Joined: Fri Mar 06, 2009 8:22 am

Post by djmalan »

Forget the scope.what do you expect to see on scope? :D Make simple in out program like rdax wrax ADC to DAC and try the chip with external and internal programs .We can not solve the problem by ourselves with only datasheet ,need more information ,it s like a closed box.
But I solved my problem by starting the chip in internal mode pin13 to the GND ,like resetting ,this not a real solution but it works . :D
frank
Posts: 1244
Joined: Wed Oct 19, 2005 12:26 pm
Contact:

Post by frank »

Little out of it today guys, had some dental surgery this morning.

One thing a scope will help with is to look at how fast the power rails come up. If too slow that can cause problems to the internal reset.

One thing I find interesting and have never seen is the switching from external to internal mode fixing a problem. That is very strange, check the voltages on all pins and the EEPROM, make sure none are floating.
Frank Thomson
Experimental Noize
djmalan
Posts: 36
Joined: Fri Mar 06, 2009 8:22 am

Post by djmalan »

One thing a scope will help with is to look at how fast the power rails come up. If too slow that can cause problems to the internal reset.

Yes ,I dont have scope that s the main problem for me ,so I have to economize with the digital multimeter. :D
I had the problem , slow rails come with another chip but that was not a real problem ,because in normal operation in the beginning the led lits for 1second but with slow rail it lits app.for 3 seconds ,after that time it goes normal.
And I tried the good chip with lower voltage (2.4volts) with weak battery ,I got no problem but ,opening was alittle bit slow .

All voltages were the same as the good chip,exept the Led out pin and chip out pins (these should be normal ,not an error ,if I overdrive the chip the same will occur).The damaged chips behave like overdriven but without sound ,but so little with a (DINNNNN)sound :D like dc on the output ,but no dc on the output .
Starting with pin13 to ground solves the problem.
frank
Posts: 1244
Joined: Wed Oct 19, 2005 12:26 pm
Contact:

Post by frank »

What is the lot code on the chip? Should be a 4 digit number with a possible alpha after it. I.e. 0819A
Frank Thomson
Experimental Noize
differo
Posts: 22
Joined: Fri Feb 26, 2010 8:01 pm

Post by differo »

Hey guys, I've been sitting last night with the build and basically getting to know how to work with my new scope, and it's Great! I do understand know why everyone is recommending to have one because you can finally See whats going on and not just hear. Frank, first of all, I think the clicks and pops come from the 9V rail even thou it's filtered, because I can see the strange artifacts on the dc voltage which are consistent with the sounding (like 80-250 mV spikes?) but I'm not so sure if they are caused by bad filtering - I did however raised the filter cap values and it's a Lot quieter (almost none). This helped little or none in my understanding why the circuit isn't stable. I got rid of most of the pops and crackles but it's still hard to get a stable working circuit. My problem is different from djmalan's because I don't use eeprom (don't have any right now..). So my pin 13 Is grounded all the time. By the way is it only 32K eeprom that can be used (I2C) or maybe 64K can also work with the circuit?

What about voltages, I want to understand more. if it's 3.4V would that fry the chit? is 3.3 the Max limit? should I bias lm317 to 3.2 maybe? my DMM shows stable 3.3V but now with the scope I get the much better resolution and see the small variations in the supply voltage(s). I'm pretty sure now that the buffer part of the circuit prior to fv-1 works well, the guitar signal comes clean and steady to the pins 1,2 no noise or cracks. But now I get some white noise on the output pins. I took some time to analyze clock/crystal. It doesn't oscillate every time I start the circuit. Still, sometimes it does which doesn't necessary means that circuit will work. I played over a minute (sounds bizarre.. : ) ) with nice clean sounds and could change the different effects and then suddenly just stopped and gave me some noise and silence and noise..
Again, I test with 10, 12,15,22 pF on the X2. Also different types (multi/single layers ceramic, poly..). I bought 3 more crystals and couple more caps to swap, no significant change. I went threw all the solder joints on fv-1 seem to be nice and clean, even on the through-hole components. When the circuit is working, voltages on the pins on fv-1 look right - input/output/ref pins are around 1.65V and all the other ones compared to a another fv-1 that I 'discovered' in one commercial pedal that I own ;).
Well, that's what I got for now. Does the slow rail mean that circuit won't start oscillating/resetting/working correctly?
donstavely
Posts: 53
Joined: Thu Jan 07, 2010 2:29 pm
Location: Windsor, Colorado

Post by donstavely »

A couple of comments on hardware configuration:

First, it is usually not possible to put a scope probe on either side of a crystal and have it oscillate correctly. The probe has enough capacitance to load the oscillator to the point of stopping it.

Second, the FV-1 says that the absolute maximum rating for the supplies is 3.5V. The nominal voltage is 3.3V to allow for supply tolerances.

Third, you need a very clean power supply, since this chip derives the analog supplies and ground reference from it. Be especially careful about the virtual ground on any input and output circuitry. Any noise or ripple on it will look like signal to the FV-1. This is a likely problem, since it is a common practice to use a voltage divider to generate a virtual ground from a single supply in guitar pedals.

Fourth, note that there is (or should be) an RC filter on the REFP pin, which I think is the positive reference for the A/D converters. The data sheet says it only draws 50uA max. The "typical application" shows 100ohm and 1uF on this pin, the development board shows 10ohm and 0.1uF. The dev board values are way too small to filter any 60Hz. Try increasing both the R and the C on your board.

It might help if you post a picture of your schematic.
Don Stavely
differo
Posts: 22
Joined: Fri Feb 26, 2010 8:01 pm

Post by differo »

Thanks for the reply donstavely, does that mean that a battery should eliminate glitches? I was debugging for so long that I stopped using battery after a While to preserve it. Will try this when I come home from work. I don't think I'm allowed to post the schematics here but you (and everyone else) can found it on the tonepad.com site here is the link to the project file:

http://www.tonepad.com/getFile.asp?id=122

There are some errors in the schem like fet orientation, input labels and some more, no 15pF (I added it in my layout) if you or someone else notices some other errors or things to (re)think about shout out!
donstavely
Posts: 53
Joined: Thu Jan 07, 2010 2:29 pm
Location: Windsor, Colorado

Post by donstavely »

It looks to me like they don't have a DC path to Pin 26 (REFP), just a capacitor. This should not work at all!

There needs to be a resistor from Pin 26 to 3.3V. Anything between 1K and 100ohm should work.
Don Stavely
frank
Posts: 1244
Joined: Wed Oct 19, 2005 12:26 pm
Contact:

Post by frank »

donstavely wrote:It looks to me like they don't have a DC path to Pin 26 (REFP), just a capacitor. This should not work at all!

There needs to be a resistor from Pin 26 to 3.3V. Anything between 1K and 100ohm should work.
OUCH! I completely missed that when looking at the schematic. Add that resistor in there an it may just solve everything. Good catch!
Frank Thomson
Experimental Noize
differo
Posts: 22
Joined: Fri Feb 26, 2010 8:01 pm

Post by differo »

Good catch! thing is that Francisco missed that resistor in his schematics but there is one actually on the layout (that I etched with some modifications cap on X2 etc). I used 100 ohm and as you said any up tu 1k would work I changed it to 470 ohm just to try it, run it on battery, well first time I fired it up I could play 1 minute without any problem, then, slowly it started, clip led blinking from time to time scratches crackles etc... Still unsolved but feels good that you people have your eyes on this, I'll keep trying!
donstavely
Posts: 53
Joined: Thu Jan 07, 2010 2:29 pm
Location: Windsor, Colorado

Post by donstavely »

Having tried the most obvious stuff, I think you need to step back and look at everything. My suggestions:

- Redraw your schematic pin-by-pin, node-by-node, EXACTLY as you have built it. By hand on paper is OK.

- Power it up fresh, and while it is working correctly, measure the DC voltage on EVERY node of the circuit with your multimeter or scope. Write the voltages on the nodes of the schematic. Look at waveforms of all of the nodes of the signal path with your scope (which I am sure you have already done).

- Wait until you get the incorrect behavior, then measure all of the DC voltages again. Write them down an the schematic. Again look at waveforms of all of the nodes of the signal path with your scope. Look carefully for any wander or shift on the FV-1 signal input pins. Also look carefully at the analog reference pins: MID, REFN, and REFP.

- If you can, scan and post your schematic here. It looks like there is a button to do this above the message edit box.

Here is a story that came to mind:
Several years ago, I had an audio circuit that worked right for a while, then started wandering, popping, cracking, etc., much as you describe. It took me three days of debugging, resoldering, cleaning, replacing parts, etc. The culprit was solder paste! It turns out I had borrowed some from a colleague. He said it was an "aggressive" flux. What he didn't tell me is that it was plumber's flux for soldering copper pipe. This is caustic and actually conductive. Even though I cleaned the board with alcohol, some flux remained under chips and connectors. It caused some weird electrolysis effect on high impedance nodes, which wandered around chaotically. I had to literally take the key components off, soak the board in alcohol, and resolder them with good flux.
Moral of story: sometimes s**t happens in places where we don't expect it.
Don Stavely
differo
Posts: 22
Joined: Fri Feb 26, 2010 8:01 pm

Post by differo »

Well, to be honest, I've suspected flux also, as I used flux to help me solder the chip. But I did clean the board really well afterward and before too. I can not see any residue on the PCB (under magnifying glass). Still a possible cause thou. I'll try to measure All the voltages as you said (because I have no other clues) but that could be hard thing to do as it is impossible to know when and how long the circuit works ok. I don't know if am imagining stuff but it seemed to me that it worked a bit longer with 470ohm resistor (I mean worked without pops and noise) but that could be just me imagining and hoping. Also maybe this could be important, I changed the lm317 from to92 to a bigger version of it you know the one with metal back which is able to push out at least 500mA as I recall. But even this big version can be really hot!!?? And NOT?! always just sometimes probably when it's malfunctioning. I think/suspect that my power supply ON the actual PCB could be fishy (traces/flux bad joints- even thou I resoldered everything at least twice). I could put every single component out scrape their legs a bit with knife and solder back. Basically that's the first thing I'm gonna do when I come home today (in about half an hour - can't wait :) )
differo
Posts: 22
Joined: Fri Feb 26, 2010 8:01 pm

Post by differo »

I finally fixed this. IT was a tough one but well worth it. A have to say that I'm impressed with this chip in several ways. First of all it's sound and second, after a week of abuse and heating, soldering, poking and what not I made it work without any problem! What I finally did is that I got pissed yesterday and desoldered it. Yes, I desoldered it Smiley This is my first smd experience but I choose to do it anyway because I ran out of ideas. You know heat I found UNDER the chip? A broken trace. It was actually a tiny bit of copper that got chipped and touched two traces while drowning in flux underneath the chip. What I learned from this build is DO NOT use any flux even if you can see a lot of tutorials dealing with smd soldering that advise using a lot of flux in order 'not to fry the chip'. Just clean your board (specially if you etch it yourself as I did) and use rosin filled solder wire; and just a little bit of it. Clean your tip EVERY time between the pins. Surface tension under the pins will do the work. I'm seriously considering to throw away the flux can I have just not to get tempted using it with another pedal.. that is how frustration sounds this time of a year..
In the meanwhile I ordered two more chips and some eeproms - time get dirty with assembly!
frank
Posts: 1244
Joined: Wed Oct 19, 2005 12:26 pm
Contact:

Post by frank »

Glad it is alive and working now. No wonder we couldn't determine the cause and why it was acting so strangely, but these things happen.
Frank Thomson
Experimental Noize
Post Reply