Stripboard ADA flanger

Started by uncle boko, November 21, 2003, 08:15:27 AM

Previous topic - Next topic

uncle boko

As I mentioned a week or so back over in Ampage, I decided to solder up an ADA flanger clone on stripboard of a size that would squeeze into sensibly sized stomp box. A proper PCB was out of the question as I neither have the time nor the patience to draw one up, although after reading Mark Hammer's procedure of drilling the holes first, I was tempted to try.

I reported earlier that the unit I built a couple of months ago took up a piece of stripboard 11" x 4", but everything was given plenty of space to allow for trouble shooting, which funnily enough didn't amount to much, apart from the couple of strip tracks I forgot to cut causing havoc in the input section. I drew up an area of holes on Excel representing the size of stripboard I wanted to use and overlaid component shapes to give me an idea of how small I could go. That seemed to indicate that I could go to 4" x 4", using 2 MC3403 for the audio path and the manual sweep amp, LM324 for the LFO and other CV processors, 4007, 4047, 4049 and BBD. The 3403s are on top, 4049 and BBD in the middle and LM324, 4007 and 4047 at at the bottom.

I have now built the input section, LFO and clock generator up to the 4047 and sod's law has already crept in because I hadn't allowed space for power lines and decoupling capacitors, so I'm now looking at a board 4" x 5.25" which isn't too bad. Now, I've also built this beast twice on breadboard so you would think that any potential problems would have reared their ugly heads by now - wrong! I had extreme difficulty in getting the LFO to work, in fact the first time I looked at the LFO output my digital multimeter read 0v. I already had the input section working OK at the top of the board, using incidentally the passive limiter from Rev3, so I decided to desolder the LM 324 and try another - but tired eyes managed to desolder the 3403 instead! It's easy to take out the wrong chip when the board is upside down - I tend not to use sockets for opamps although I do for cmos and BBDs. I then checked the 324 on another board and found that it was dead - a brand new chip. So with another 324 in place which I knew was working, I tried for LFO voltage again but this time a static voltage, so some progress. I then replaced each component making up the LFO and it was not until I replaced the 33uf capacitor (again brand new) with an old 10uf,that it worked. I then tried each of 5 other 33uf caps I bought at the same time recently and none of them worked!

Upward and onward, the clock generator came next and I copied the layout from my other board. The 4047 is positioned 1 row below the 4007 so that pin 12 of the 4007 lines up with pin 2 of the 4047. I also managed to get a 22p - 65p variable capacitor from Maplin Electronics (wonders will never cease!) which will allow me to play around with the max and min delay. This worked first time, giving out roughly 6.4v at each of pins 10 and 11 of the 4047. If the voltages at those output pins are not the same, the BBD will not work - believe me.

That's as far as I've got folks - one thing to remember when designing stripboard layouts is to allow plenty of space for electrolytics, unless they are very small you need to allow an area of at least 3 x3 holes, and preset pots take up more space than you think.
better to be in bad taste than to taste bad

uncle boko

....From the pins 10 and 11 of the 4047 I needed to get the 2 out of phase clock signals to pins 7 and 5 of the 4049 buffer inputs. Now, I realised that I hadn’t quite thought this out properly so I downed tools for a couple of minutes, thought of stray capacitance and stray clock noise, then forgot about it quickly. After all, most Ampagers chuck as much distortion and grit as possible into the sound, many Db above the clocking crap which the minimal filtering on this beast attempts to kill! The 4049 and BBD were planned to sit in the middle of the board, with the BBD to the right so that I could string the series 1k resistor from the input 3403 via the top strip without disturbing the output 3403 - which I was leaving until last so that I didn’t forget the 1k, otherwise I might have to start elevating components, which has got me into all sorts of problems in the past. (So far I had got away with just one elevated resistor strung over the 324 because one end needed connection to +v into its pin 4 and the other to the maximum clock trim, already soldered well to the right of the 324.)

So with the 4049 buffers in place, and the correct equal voltages showing at its pins 10/12 and 2/15, I put in all the components around the BBD and it worked first time. It’s always good to hear the nothing â€" distorted sound â€" clear sound â€" distorted sound â€" nothing sequence when turning the bias trim. No flanging yet, just delayed sound with plenty of hiss from the BBD output. I was happy with progress at this stage and tired eyes convinced me to leave the output section until after our dental checkup on Saturday. Incidentally, we always stop off at a restaurant called Le Piaf in Dulwich Village (in South London) on our return home from the dentist which I recommend to any one passing nearby â€" situated in an oasis among some of the nastiest areas of London. I digress â€"

The last leg of this journey to ADA flanging was not all plain sailing. I needed to put in decoupling capacitors which I did and lost the LFO for a while. I had used a tantalum capacitor due to shortage of space to decouple the LFO power which arrived via a 470R resistor. I eventually found the tantalum had been connected with reversed polarity and was dead. I left it out of circuit as the LFO worked again and gave no noticable clicks except on very fast speeds which I don’t use anyway. I put in the output 3403 and its components which in theory should have given flanging at its output. It didn’t, the reason was not apparent for a while and I initially suspected a duff 3403 and was about to start desoldering when I noticed on the strip side of the board that the leg of the 4k7 resistor to ground was not properly soldered in. I decided to power up again and I had flanging, but with a very uneven sweep. More investigation revealed that track cuts, which should have isolated 2 legs of the clock variable capacitor from the manual sweep section, had not been made. This surprised me because the clock still worked â€" better brains than mine could probably give an explanation! After the cuts were made I had a very smooth sweep indeed, so I moved on to the resonance section. This was the last section to deal with, and I hadn’t left enough room for the filter components. I had to rig a network of a 30k and 47k resistor joined above the board, with the 1.2n capacitor connected to that join with all other legs soldered strip side. This was a bit fiddly but I got there without any burnt fingers. Powering up again with resonance produced no output at all, so something was badly wrong â€" more forgotten track cuts seemed to be the culprit, but it was not all. Instead of connecting the output filter to the resonance trim, I had connected the BBD clock null output, just one strip away. I put that right and was in business with strong familiar sounding resonance.

There is so much variation of flanging sweeps available on this beast. I am going to have the max clock trim as a panel control because it provides continuous control of which part of the sweep area can be used. So, for instance, if a narrow sweep is set up, the max clock trim can fix whether that sweep is in the high or low end. Another point to remember is that the auto sweep level trim does have settings which (Stay by Maurice Williams and the Zodiacs just on Bob Harris â€" one of my all time favourite singles â€" 1min 45 secs!) STOP THE LFO FROM WORKING, so I just set it in a position where the LFO works and forget it.

Whether the 3403 quad opamps make any difference I don’t know and I don’t intend to desolder them in order to try others â€" perhaps a later re-breadboarding may bring about some mad scientisting in that area and even bounce experimentation perhaps. I will be trying different capacitors for C7 â€" big deal . This is stated as being 4.7n on the component list, and Mike Irwin suggested increasing this. Looking at my board I see that I’ve missed out this capacitor all together so I expect there are DC considerations to this (Mike?) I see that on my larger board I have used a 10n capacitor for C7. I still have a couple of TDA 1022s left from the old days which I might also apply to a later re-breadboarding of the ADA although I have always found them quite noisy.

Now I need an enclosure for the board and it does fit nicely in an old EH 5 pot case I have although I would prefer to have all the controls in a single row.  :lol:
better to be in bad taste than to taste bad

Mark Hammer

Great to finally see you here, Steve!

Paul Perry (Frostwave)

If a LFO doesn't work wiht 3 new electrolytbut does work with one OLD one, it makes me think either the ckt is wrong in design, or there is a bias fault!