Hello everyone
I'm building a Pladask Elektrisk Bandorg Delay pedal on a PCB from Pladask Elektrisk. Everything works on it *except* it will only produce one single repeat regardless of what the feedback pot is set to. It is a PT2399 based delay, so I'm wondering what pin(s) on the PT23999 control the amount of repeats heard. My guess (from looking at other PT2399 schematics) is that it is something to do with the relationship between pin 14 and pins 15 and 16. I suspect that pin 14 should get the output of lug 3 of the feedback knob (in this case a 25K pot) and so if measured with a DMM would report 25K on that pin? Then assuming the feedback pot is set halfway around its travel, (say 12K) that value should be present at pins 15 and 16. For me I get 25K at pins 15 and 16 and I suspect that is wrong, and that is why I only get one repeat. Does that sound likely to anyone here? Presumably lower resistance values would enable more repeats?
I will post some images tomorrow; the circuit has an extra level of complexity due to there being a momentary stomp switch that generates maximum oscillation / repeats that I'm unsure how to wire - but I'll need to show images for that. I need to get some sleep now. Thanks for any help you can offer!
Knutolai has kindly agreed that I can post some images of the PCB so there'll be pictures soon :)
The chip can't organise repeats in any clever way. The first repeat is simply the input recorded into the delay being played back after that delay. It is always recording new input, even if that is silence and that always overwrites and replaces older recordings. So in its basic function, it can only ever produce one true repeat.
To give the illusion of multiple repeats, the playback, after converting back to an analogue signal, must be fed back to the analogue input. Losses in tone in and level in the whole process mean it can never really "hold" a repeat, despite being digital. It will either decay to silence or build up into flat out distorted mush unless the feedback loop had perfect fidelity and exactly unity gain. If it is expected to record new audio at the same time as keeping repeats going, there is a problem, since combining the 2 audio streams will boost the recording level by an average +6dB. To avoid any distortion from that, the signal levels must be reduced which means the repeats must inevitably fall in level on each trip so cannot be sustained for long.
https://www.electrosmash.com/pt2399-analysis
The feedback path will be from pin14 and mixed via resistors back into pin16. The pot is usually acting as a volume control but you might expect some additional resistors and capacitors in the way. I would need to see a schematic to help more.
Thanks anotherjim - I didn't realise the chip didn't actually keep the audio circulating inside itself... I guess that's because it has so little onboard memory (that link you posted shows it has almost nothing; kind of amazing it even works!). So presumably to stop the repeats from building up on top of each other into digital mush the repeating part of the circuit goes through resistors / capacitors to cut the high frequencies and also lower the volume of the signal so that it doesn't swamp the next incoming notes / music / audio... and this part is the feedback path from pin 14 to 16. So this makes me think that if I'm only getting one repeat, this might be due to a component in the feedback path being incorrect and lowering the volume of the repeats to below audible threshold? Does that sound feasible?
Sadly I don't have a schematic for the circuit. I was hoping it would be similar to a pedal that Knutolai did post a schematic for a few years ago, but it seems quite different (to me at least). I did take photos of the PCB before and after populating it with components, obviously if that's imposible to follow then I'll try to draw a schematic.
This is the bare PCB (obviously)
(https://i.imgur.com/DsKwrdv.jpg)
(https://i.imgur.com/ZBnIZHv.jpg)
Here is my populated version, note the DPDT momentary swith is detached at present because I'm unsure of the best way to wire it to the circuit (the original design needs a SPDT momentary and I don't have one and was hoping I could use this instead?)
(https://i.imgur.com/rULn7lV.jpg)
(https://i.imgur.com/JeK8AOR.jpg)
Finally the offboard wiring (which is where I'm confused regarding the Max, Pot and Feed cables and how they would connect to a DPDT - just using one half of a DPDT doesn't work - I did try connecting Feed to Pin 2, making a jumper from pin 1 to 6 and then connecting pin 5 to Max and that worked, but then I wondered what the spare cable at Pot is for?)
(https://i.imgur.com/jVwOAIw.png)
Any thoughts on where to go next would be great thanks :)
Edit: the images should be smaller now! And if you click them should get bigger.
Max/Pot/Feed switching is I guess...
SPDT on-on or on-off-on type.
Max = full feedback level, guaranteed to runaway/howlround. Maybe lug 3 of the feedback pot.
Pot = a variable feedback level for the usual echo feedback level control from zero to max. This is probably the wiper of the feedback pot.
Feed = the input for the selected feedback via the switch. An on-off-on toggle could be used so a zero feedback setting is instantly available.
Thanks AnotherJim, however I have solved the lack of repeats issue :) Getting hold of a momentary SPDT switch and installing seems to cure everything - I now have repeats, and the momentary effect slowly builds up instead of triggering instantly as it did when I used a DPDT switch. Now I have the circuit working I am planning some changes before I box it. The modulation pot is far too extreme; anything more than a third if the way around totally destroys the signal, so I will try a smaller pot size so that I can still get complete audio destruction, but hopefully have a wider range of not so destroyed tones first. I am also adding an SPST switch that connects pin 9 and 11 of the PT2399, becasue this turns the pedal into an extremely loud fuzz with lots of weird artifacts (octave down, some arpeggiating, extreme noise... it's a blast, if somewhat uncontrollable). To control it I'll add a volume pot to calm it down a bit, and also a high pass filter because the circuit is extremely bass heavy (especially on humbuckers) which I may connect with a dual pot to the normal part of the circuit so that the bass there can be calmed down a little. I'm off on holiday soon so it will be a while before I can post again, but hopefully will have a demo for interested parties to listen to. Thanks for your help!
Quote from: moid on August 21, 2019, 05:26:36 PM
The modulation pot is far too extreme; anything more than a third if the way around totally destroys the signal, so I will try a smaller pot size so that I can still get complete audio destruction, but hopefully have a wider range of not so destroyed tones first.
Depending how the pot is wired, you might find that going for a pot value of one-third of the original and then putting a resistor of two-thirds the original pot's value above it in series does the trick (this essentially leaves the total resistance exactly the same, but allows the pot to sweep only a 1/3rd of the range - the useful part). There won't be a pot that's exactly a third, but don't sweat small differences in the values. Go for the closest thing you can get - it'll be fine.
I built the stripboard of this, and also have one of those PCBs though I just tweaked my stripboard layout accordingly. I agree the modulation seems too much, but with super short repeats and low mix it's interesting.
Take a look at the Casper Electronics EchoBender.. that circuit gets into the same territory.
Quote from: ElectricDruid on August 21, 2019, 06:42:32 PM
Depending how the pot is wired, you might find that going for a pot value of one-third of the original and then putting a resistor of two-thirds the original pot's value above it in series does the trick (this essentially leaves the total resistance exactly the same, but allows the pot to sweep only a 1/3rd of the range - the useful part). There won't be a pot that's exactly a third, but don't sweat small differences in the values. Go for the closest thing you can get - it'll be fine.
Just got back from a holiday, thanks for the replies. I did try this technique with a 50K pot and a 220K resistor (with the resistor attached to lug 3 and then also to lug 2 of the pot) but it didn't help sadly. I did test a range of other pots and the best result was a 20K Log pot - this gave controllable modulation across almost the entire sweep, with the very end reserved for total destruction :) I used that one in the end. The only downsides are the momentary SPST switch is no longer as chaotic as it was when it was added, but I can live with that... sadly the awesome fuzz effect I found by connecting pins 9 and 11 no longer works. However in it's place I found that a cable from lug 2 of the modulation pot to a log 10K pot (lug3) with lug 2 connected to pin 11 of the PT2399 gives a lovely broken up chorus sound! By adjusting the 10K log pot I can blend between chorus delay and the original broken delays.
There is only one (major) flaw. When I do this and play for a while, every now and then the PT2399 chip locks up and stops working, and at the moment the only way to get it working again is to break the connection in that little sub circuit and then re attach (everything is all connected with crocodile clips at the moment). I tried reading the electrosmash PT2399 document and it mentions avoiding the lock up with a circuit. If I built that anti-lock up circuit onto a separate veroboard and attached it to this one (presumably to pin 6 of the PT2399) would that prevent the lock up from happening? The circuit is discussed here: https://www.electrosmash.com/pt2399-analysis (https://www.electrosmash.com/pt2399-analysis) Does anyone have any idea how I can stop this from locking up because the sound is quite beautiful and I'd like to keep this mod if possible.
I have tried putting an 82nF capacitor onto pin 11 but that removes the chorus effect. I also tried the 100K pot attached to pin 2 of the PT2399 for chorus but that didn't do anything (presumably because of the lack of anti-lock up circuit).
Thanks for any thoughts!
Quote from: BetterOffShred on August 23, 2019, 12:50:18 AM
I built the stripboard of this, and also have one of those PCBs though I just tweaked my stripboard layout accordingly. I agree the modulation seems too much, but with super short repeats and low mix it's interesting.
Take a look at the Casper Electronics EchoBender.. that circuit gets into the same territory.
Try switching the modulation pot to A20K - it's fantastic! You still get lots of broken fuzzy repeats but they are mostly controllable and I don't need to add the high pass filter; the pedal suddenly lost a lot of over powering bass (it's still very present, but not overwhelming)
That Caspar Electronics EchoBender definitely has some cool sounds in it - thanks. I hadn't heard of that one!
Some... progress? I built the anti lock up circuit (see vero layout below with schematic - maybe someone could check that I didn't screw the circuit up? Thanks)
(https://i.imgur.com/Q88MAGy.png)
I attached it to the main PCB and now I don't have any delay signal at all, just a fairly pleasant if gentle chorus / reverb / filtered sound and most of the pots have limited effect on anything. I can mix between dry signal and the chorus on the A10K pot I put in (the one the connects the Modulation Pot with Pin 11 of the PT2399), but before I added the anti lock up circuit I would not get dry signal, just the broken delay signal... On the good side the chip hasn't locked up at all, but it can't do both chorus and delay it seems :( If anyone has any suggestions I'd be glad of some hints. I will remove the anti lock up circuit and chorus if I can't get the delays back... it's a shame; I thought I had found a cool new chorus sound with delays but fixing the chip lockup seems to prevent the delays from working.
Oooooh extra bonus feature - the LED on the PCB no longer turns off when the pedal is bypassed - it loses half it's brightness (the sond bypasses correctly)
Tomorrrow I could try swapping PT2399 chips in case I wrecked this one by accident? I socketed it at least!
That circuit makes the delay range very short in order to do chorus effects. It needs resistance out of pin 6 to ground to get back into longer delay ranges. Try putting a pot as a variable resistor between pin6 and the transistor collector.
Quote from: anotherjim on September 04, 2019, 03:37:24 AM
That circuit makes the delay range very short in order to do chorus effects. It needs resistance out of pin 6 to ground to get back into longer delay ranges. Try putting a pot as a variable resistor between pin6 and the transistor collector.
and, if you wire a switch to short across the pot [or open, so pot in circuit], you can switch between extra short delay non-latch-upping, and ordinary pot select delay [within reason, my eyes glazed over a while back, might be working on the wrong scheme].
Thanks very much chaps, I will try that this evening and report back! Duck - I had forgotten Knutolai had put the schematic on the forum... just searched and found it; I will post it below and also if I get these mods working draw another version in case anyone fancies having a go themselves.
(http://i.imgur.com/Uf2ys0k.png)
ahh, well, see, on knut's drawing, you'd replace R5 with the C,E of the antilock, and sling a short/open switch across the 1,3 of the TIME1. I thimk.
...and here we come to where my brain has issues with complicated fings / stuff / tricksiness :)
Before reading Duck's reply, I went with AnotherJim's suggestion and stuck a B100K pot between pin 6 and the transistor collector of the anti-lockup circuit. This worked (mostly). With this pot fully anti clockwise there is a sort of chorus (a weird one at that, with a strange filtered out hollow sound and some fuzz, plus some vicious feedback if the feedback pot is turned up too much), and as I turn the pot towards clockwise, the chorus slowly fades and is replaced with the usual delay sound of the pedal, and a mixture of half and half sounds great! So that was the good news...
...the not so good news - if the pot is turned clockwise too far (near the end of its sweep) the PT2399 locks up... bugger... on the good side, it can now be unlocked by turning this pot fully anti clockwise and then moving it more clockwise, but combinations of changes of the strength of the feedback, time and modulation pots all cause it to lock up. I tried other pots; a 50K lin lowered the maximum length of delays (and locked up), and a 250K lin turned the pedal into a total noise destroyer as soon as the sweep was halfway round (which was fun for a few seconds and then it started to hurt) and at full extent the PT2399 locked up again. I've decided to stay with 100K for this pot as it sounds best.
Other not so good news in the above mentioned mode the Time pot's setting is overridden by the 100K pot - as I turn this clockwise I don't just get a blend of short chorus and delay time, the delay time changes as well, and the Time pot doesn't do much. The modulation pot also seems to work backwards at this point! More modulation anti clockwise, sometimes no modulation in the center of the sweep and some more modulation on full clockwise, but not as much as the anti clockwise modulation. Bizarre. The pot I put between the modulation pot's lug 2 and pin 11 of the PT2399 does very little - it modulates the texture of the repeats a bit, but seems rather redundant which is a shame because it was great as a mixer before I added the anti lockup circuit.
I am thinking that I need to possibly remove that mod lug2 to pin 11 pot (and maybe the connection?) and then add the Electrosmash chorus effect circuit part I am missing (a 100K pot on pin 2 of the PT2399) and then somehow blend this with the delays... hmmm this probably won't work... I am wondering if there is a way to have the short chorus sound blend with a longer delay, but I suspect this is a no way josé sort of issue due to the PT2399 not being able to run two delay lines at he same time for obvious reasons!
Duck - regarding your suggestion, do you mean I should remove the R5 (1K5) resistor entirely? At the moment the C of the anti lock circuit goes straight to pin 6 - I didn't remove any of Knutolai's circuit. Should I connect C to lug 1 of the time pot and E to ground. I think I understand the short / switch idea - it would either allow current to flow via the Time put to pin 6, or if flipped on would run through the anti lock circuit first, bypassing the time pot and then go to pin 6. Is that right? I could use a SDPT with the center lug connecting to where R5 (1K5) is now, and lug 1 (for example) would go to the C of the anti lock circuit and lug 3 would go lug 1 of the Time pot... that sounds good? Wouldn't I need the R5 for the original circuit though?
I will have to try this tomorrow; it's too late for me to do this without screwing up! Thanks for your help both of you :)
I like the sound I currently have if I could just stop it locking up! And for my excuse of a prime minister to stop screwing the country and it's future. Oh, and I want the moon on a stick too; just in case anyone has one of those handy. One can but hope!
That anti-lockup circuit is brutal so when the transistor is fully conducting it is effectively shorting pin6 to 0v unless there is always some extra resistance fitted in between the transistor collector and pin6.
I'm not sure that lockup was ever a cause of trouble with the original circuit, although there's a chance it might, in which case I would put it between the 1k5 resistor and ground.
If you want to add chorus, the transistor can go across the 1k5 resistor (which remains in place) but it would be wise to put a low-value resistor (100R to 220R) in series with the transistor collector so that when it fully conducts, it can't short pin6 to 0v if the Time1 pot is at minimum resistance.
let's vote on it!
sorrry moid's, I think we are working at cross porpoises and different schems. as Jim says, you need some stopper resistor even with the anti-lockup. I think, and I'm not paying enough attention. go with wot Jim sez. as for my across the pot switch, in Knut's circuit, it would short the time pot end to end for minimum delay, or open and allow the pot to set the time as usual.
If you consider that the Electrosmash idea true (and it is only conjectured that that is how the VCO speed is set since there is no detail in the chip data), then there has to be a limit to how low the pin6 resistance can go.
If pin6 is the inverting input of an internal op-amp with a fixed internal feedback resistor.
If that op-amp has its non-inverting input tied to the internal 2.5v reference (plausible because we know that forcing the reference voltage up or down also changes the VCO speed).
Then there will be a point where reducing the pin6 resistor ceases to raise the VCO speed because the opamp output has hit the positive rail and can't swing higher. Also, as the pin6 resistor is set to ground, then there is always a negative DC offset into that opamp inverting input which means the output can't swing below the 2.5v reference no matter what value the pin6 resistor is.
So I'm not sure I agree with the Electosmash explanation. It's possible, but it don't seem right.
At any rate, The chip doesn't like pin6 to be grounded at any time. Knut's scheme fits a 1k5 which the chip should be happy with without any anti lockup device. In fact, it is fine with 1k as well. But note the 1k resistance to "spare" opamp output pin14. When audio swings that op-amp to a low voltage, the 1k is added in parallel with the pin6 resistance. So it can already get pushed below the 1k "safe" minimum pin6 resistance and doesn't really need it going much lower.
Thank you very much anotherjim and duck - there is some small success to report :) a 150R resistor to ground from the anti-lock circuit did the trick - no locking up so far regardless of what setting I put the pedal to. I then got keen (I know, a worryingly dangerous thing to do late at night) and rebuilt the circuit so I could use a DPDT switch to switch between the original circuit and the more chorus delay... and it worked! Yay! Except... the additional modulation pot no longer seems to do anything. It was attached to the original modulation pot and worked in conjunction with it (sort of, I think) but now it does nada... it might be that it now needs to be a larger pot type to have an affect, or perhaps the effect I heard from it was actually some sort of short between the mass of wires attached to the build... or maybe I've detached something else and not re attached it? I will try a larger pot tomorrow and/or consider attaching that pot to another point on the PT2399 to see if that gives better results. If I get nowhere I'm seriously thinking of moving the 10K trimmer off the board and out as a pot that can be adjusted because it has an enormous affect on the way the momentary stomp switch behaves and changes the tonal value of the feedback that produces, and it can go in the place that is currently in use by the extra modulation pot. It's fun! The only other negative is the chorus delay at ultra short settings doesn't seem to be quite as weird as it used to... I'm assuming that's the result of the 150R resistor to ground? I'll live with it of course (maybe I will try a 100R instead... that might be better, but I thought I locked the circuit up at one point when I tested that value, but that might have been a short instead). Or it might be the lack of modulation to the chors delay circuit that used to be going on when the two circuits were more connected. Anyway, at least it's a positive end to the day! I hope your builds went well and your oscillations were all intended!
It is quite normal to have 7 or 8 pots on a delay isn't it? Asking for a friend...
QuoteIt is quite normal to have 7 or 8 pots on a delay isn't it? Asking for a friend...
Hmmm.... let's see...
1: Delay input level
2: Feedback level course setting
3: Feedback level fine setting
4: Feedback tone
5: Delay time
6: Modulation LFO depth
7: Modulation LFO speed
8: Modulation envelope depth
9: Modulation envelope Attack
10: Modulation envelope Decay
11: Modulation warp depth ( input pitch to delay time)
12: Delay output tone
13: Delay output level
14: Dry signal mix level
...
...and then there are all the switches you could have...
Obviously, you can go crazy with features in a delay. You can find many designs that cherry-pick a few significant functions simply to keep it practical.
OK It seems that i could've got a lot weirder then :) Thanks for the suggestions Anotherjim.
Some good news! I finished the pedal and it works :) My first finished and functioning pedal in two years (what a sad statement that is...). A lot of extra changes were made across the weekend, mostly for the better (I am now a huge fan of attaching random pot lugs to a low value resistor and then connecting that to ground). My only mistake was making a pot out of a trimmer that at one point modfied how long it took for feedback to fade in, but I must've got keen on changing something else and didn't test that function - it still slightly works but is not worth the effort of it being there... but the box is such a nightmare to assemble that I'm not making any changes in case I break the mojo / various cables and dodgy bits of insulation tape). Anyway my son said he is happy to record a demo of it, so watch this space! The pedal sounds great - lots of lo fi echo tones to be had, plus some impressive fuzz sounds, a little near ring mod feature and all the oscillating feedback you could desire!
Well it's eating my words time of course... I've since discovered that this pedal does not behave itself when connected to any other pedals, even if they share different power sources. The really cool sounds I made from connecting some of the lug 1's on some pots via resistors to ground, are, I expect, to blame. On it's own I can get some really interesting sounds, but when connected to other pedals the circuit works, but only at a very basic level and there are none of extra cool sounds that I like. Sooo, big question time... is there any way to somehow isolate those ground connections so that the pedal could work as intended with other pedals? At the moment those extra ground connections go to the ground lug on one of the audio jacks, or to the ground of the power jack. I know I need to draw a schematic to help, and I'll need to take the PCB out of its box somehow to do this, but is there a way to allow part of a circuit to go to ground that is not shared with anything else? I kind of have a vision of a strip of copper or aluminium that these extra connections attach to and then have that metal wrapped in insulation tape so it can't touch the enclosure or any of the circuit. Is this a daft idea? If it's a good idea, how large does this piece of metal need to be? Thanks for your thoughts.
I suspect a lot of circuit bending practices only succeed because the victim is usually some poor defenseless battery powered toy that lives on its own.
You might get away with making the delays ground virtual, but I suspect that will bring its own problems. It would definitely need its own isolated power supply and the 9v split so that the jack grounds are at the +4.5v split point and the circuit ground then left to float at -4.5v below the split. Actually, as there is a 5v regulator, then the +5v could be connector ground - the split doesn't have to be equal.
Otherwise, you could isolate the input and output with audio transformers.
I'm sure you're right about the circuit bending... it's certainly been something I've found when I customize something to make a weird noise, that the weird noise is not so cool as soon as you plug the device into any other pedal :(
Hmmm isolating the input / output jacks with transformers sounds easier (to my amateur ears anyway!). I did some reading about transformers that can do what I want and I think I should look for a 600:600 ohm type (doesn't change sound volume or impedance at all) Does that sounds correct? I've never used a transformer before, but if this works it sounds like it could fix an issue with a couple of other pedal designs I have abused... Presumably I need one for the input and one for the output jacks? Quite how I will get them in the enclosure is another story...
Does this one look OK? Bourns LM-NP-1001-B1L https://www.bitsbox.co.uk/data/transformers/LMNPLP-42750.pdf (https://www.bitsbox.co.uk/data/transformers/LMNPLP-42750.pdf) I can get hold of those fairly easily, and they don't look too massive... but I don't know if they will fit into the rather snug space inside the box. If you happen to know of a smaller one please say so.
Thanks for the suggestions, they're greatly appreciated.
Hmmmm I guess if I can't fit them in the box itself I could make an additional small enclosure with four audio jacks - two for sound in, two for sound out, with the transformers inside so I could connect this to the delay pedal and any other pedal around it? Would it require power? It looks to me that it wouldn't because I'm not changing the audio signal... I think? :icon_redface:
Those transformers will probably be ok.
They are rated as Modem isolators. That only requires telecom voice frequency response, 300Hz to 3.3kHz, but as we're not in Kansas anymore, that might actually be for the best :) However, bear that in mind if you make them up as a general-purpose ground isolator.
According to the datasheet, they are full response 200Hz to 3.5Khz with some useful response up to 10Khz.
The primary impedance of 600ohm won't work well driven by passive guitars, so you will want an active pedal of some kind with a buffered bypass (any Boss stompbox for example) or if true bypass then the effect must be active.