AMZ Fat mod on tube screamer clone problems

[mr aize]

Hello all, warmest greetings. So I've just made my first stomp box, a clone of the infamous tube screamer. I got all the info from this excellent website. http://www.generalguitargadgets.com/projects/17-distortion/82-its8-tube-screamer-808-replica-project?phpMyAdmin=78482479fd7e7fc3768044a841b3e85a I printed the circuit board layout and etched the board (first time trying it so it wasn't perfect but I couldn't find any shorts with my multimeter and tinned it so pretty confident it's fine). I have used all the parts suggested, 3904 transistors and a 4558d opamp. However, I have used vintage parts for most of the resistors and capacitors (not the electrolytics) because I found a box of tropical fish caps in a load of electronic stuff I inherited. I have since changed several of the parts to the keeley 808 mod specifications on the mods page. It all works fine so far (although I have also done the diode lift mod and i'm noticing some popping (clipping?) sounds when the drive is cranked right up and the diode lift is selected) The only non spect part I've used, was to use a tantalum capacitor, instead of a film one at c2.

Anyways, some info on the real problem that I've noticed. I also did the amz fat mod (installing a switch to bypass c3 (the 0.047 (although now 0.1) capacitor). This is where the problems have started. (I need to mention that i'm actually using the pedal with a tb303 mono synth, rather than a guitar.) So, with the fat mod switch engaged and the resonance on the 303 around midway, the pedal quickly fails, with the sound getting steadily more muffled, until it completely goes away. It seems to have gotten worse over time (within the space of about half an hour of testing) (although I'm not certain of that) which made me think it was a short, causing a part to fail. I've since changed several caps (c10 and c1) and both transistors and the opamp but no improvement. There was a short (sadly for the life of me I can't remember exactly where it was), possibly around c8, but I found it and removed it before changing the transistors and opamp so don't think it was that.

So anyone who can suggest anything to check, that would be hugely appreciated. Many, many thanks.

[Seljer]

Even without the diodes, with a moderately loud input signal there is enough gain from the opamp to run out of headroom and clip directly on the supply rails.

Do you have a multimeter? Measure the voltages on the opamp pins and the Vref supply (the junction of R9, R10 and C7) with and without the switch engaged.


also, did you use the specified bipolar capacitors for C2 and C8?

[mr aize]

Hi, thanks for the speedy reply. For C2, I had to use a tantalum (polarized) capacitor, because I couldn't find a 1uf film capacitor, I've installed it with the negative pin facing the opamp.

[mr aize]

Just measured the voltages on the pins with a lead plugged into the input jack (in order to turn it on) but not connected to anything else. (Thought i'd mention it because there's no signal being passed through the circuit and no ground connected, other than the battery (I haven't installed a power jack))

With the positive lead on the opamp pins and the negative on the junction, pins 1-7 were all between -0.3 and 0.0v dc with the switch off. Pin 8 was 4.6v (This is with the switch off)

With it on, pins 1-7 were all between -0.7 and 0.0 and pin 8 was 5.6v



Also, yet another problem has just arrived, in that I've noticed a lot of crackle and hiss when the drive pot is fully open (think it's open, when it's most distorted) but there is no signal being passed through. When the bass boost switch is on though, this goes away.

[Seljer]

You chose a slightly unconventional way of measuring things, you measured all the voltages relative to the reference voltage. Typically you put the negative lead of the meter to ground to get all your measurements relative to ground.

Another measurement to take would be the voltage on both sides of C2 in case it's been fried by reverse polarity for some or other reason.


Craclke and hiss in my experiences usually comes from flux residue left on the PCB after soldering and tends to go away after cleaning the board with denatured alcohol and a toothbrush.

[Philippe]

ironic how the LDO (landgraf dynamic overdrive) created all of these perceived/ideal ts808 tonal options...IMO (meaningless at best) the tube screamer sounds great with symetric clipping & no doctoring of the frequency range. want an overdriven plexi sound? buy a marshall & don't expect to fake the tonalities by running assymetric diodes through an ss amplifier. besides, 95% of the listening audience wouldn't/couldn't tell the difference between a tele, strat or lp when run through a massive pa system with overdrive & distortion.

[mr aize]

You chose a slightly unconventional way of measuring things, you measured all the voltages relative to the reference voltage. Typically you put the negative lead of the meter to ground to get all your measurements relative to ground.

Another measurement to take would be the voltage on both sides of C2 in case it's been fried by reverse polarity for some or other reason.


Craclke and hiss in my experiences usually comes from flux residue left on the PCB after soldering and tends to go away after cleaning the board with denatured alcohol and a toothbrush.

Ah ha, as you may have guessed, bit of a noob, will recheck those voltages tomorrow and will try cleaning up the board as it's pretty well coated in flux. Thanks. Will let you know what happens.

[mr aize]

morning. So, rechecked the voltages and got the following,
with the switch off
vref 3.8v
pin1 3.5v
pin2 3.5
pin3 3.5
pin4 0.0
pin5 3.5
pin6 3.5
pin7 3.5
pin8 8.4
capacitor 3.8v and 2.6v (sorry stupidly can't remember which way round I measured)


with switch on
vref 3.0
pin1 1.8v
pin2 2.6
pin3 2.7
pin4 0.0
pin5 2.3
pin6 2.3
pin7 2.4
pin8 8.4

capacitor 2.8 and 2.0

Unfortunately, in the process of cleaning the board, I appear to have made it worse, think the brush I used had something on it, so now waiting on some acetone to clean it properly. It's still very crackly and its developed a ground fault issue I think (it's in a plastic housing and when you touch the drive pot it humms a lot) Was gonna change the pot to a 1meg anyway and will run a ground to it when I do, but unfortunately that's in the post now. Anyways, so if anything looks leary from those voltages, I'd appreciate knowing but for now I'm admitting defeat, until more parts arrive in the post, hopefully on tuesday.
                         

[duck_arse]

I also did the amz fat mod (installing a switch to bypass c3 (the 0.047 (although now 0.1) capacitor). This is where the problems have started.

here is the hint ^. when you say "with the switch on" and "off", is it this switch you mean, or the bypass switch? if you have the wiring wrong (and we have seen that happen before, I think once), you might be applying the Vr voltage, via R6, to places it is not supposed, like D1//D2//C4 etc. and your voltages all dip when the switch closes.

try remove the switch mod, replace with a single cap, value doesn't matter, and see if same problem. what cap did you put for C8, and what voltages on either end? and to disprove C2, remove the electro you have fitted, and stick in any value poly cap, 100nF will do, so you can measure the voltage either side without worrying if it's fried. replace it later.

[mr aize]

Hi there, thanks for the reply. When I say 'switch on' and off, I mean the bypass switch, so by switch on, I mean, bypass switch engaged. I have ordered some new parts, including a couple of 1uf capacitors, which I'm gonna swap for the tantalum and install in line with the bypass switch, to see what happens. Unfortunately, none of the parts will arrive till mid next week so may have to wait till then to do any experimenting.There shouldn't be any reason for the wiring of the switch to be wrong, since I have followed the layout on the website in the op exactly and I doubt they've put a faulty layout on the website. I will double check tomorrow, in case I've done anything ridiculously stupid and triple check for any shorts, but after that, it's just a case of waiting till the new parts arrive I spose.

[duck_arse]

.... and I doubt they've put a faulty layout on the website.

they do. they will. they have. they are. don't get caught with a faulty layout. do a search for a build report or someone's build notes. there is a current thread on bad layouts, worth searching for.

let us know where you get to w/ new parts.

[mr aize]

BOOM Fixed it! 

So did a bit of research and decided the problem could just be that the vref was too low, since it was less than half the live. A bit more research and figured I could probably change that by just changing a resistor, so I did. Changed R10 to 11k, the voltage of the vref came up to just over half an it works like a dream. Happy days. Turned out to be a bit of a balancing act though, actually tried several slightly different resistors because it made quite an impact on the tone of the pedal. The closer to being exactly half volts, the more grit it had, right up till it craps out, if it dips too low.

The pedal's still quite noisy, although not as bad as it was. Gonna try change the drive pot to a 1meg one, to see what effect that has (sposed to give it more distortion) and i'll run some grounds to the pots when I do (it's in a plastic case) see if that improves anything.
Many many thanks for all the advice here. I can pretty much guarantee I'll be back. Got the bug now. Thinking of making a compressor pedal next, maybe the color sound supasustain? Want to try run some drumloops through it and see if it does anything interesting.

So, to summarise though, in case anyone else has this problem, make sure the vref is exactly half or slightly over and thanks again guys.

[duck_arse]

it's always good to hear a boom.

can you measure the resistor you pulled out of the circuit, and also the one that was left behind? just outta curiosity.

I have a coloursound supa sustain, was given to me many years ago, and a roland as1. I don't think I ever heard it working, the case went to a double fuzz delux, still have the board right here (he gestures to the corner). search up compressers hereabouts, there are plenty current designs.

[mr aize]

R10 was originally about 9.8k and R9 was originally about 9.9k, after some mucking about, I ended up with about 11.6k for R10 and 10.2k for R9. I think the issues might have been caused during the etching process, when I first made the board, it might have made the vref rail a little thin? Incidentally, that's something I should probably mention for anyone thinking of etching their own board, using the guitar gadgets layout I linked to in the op. I would recommend making it a bit bigger than it is laid out, since it was very fiddly. (Of course I've been using vintage parts, because that's what I had lying around and they're all huge, so might work with brand new components, although the solder pads on the layout are very very small)