DIYstompboxes.com

Project: The Hamlet Delay and Preamp

The Hamlet Delay is my take on a PT2399 delay with tails. The dry path is an always-on preamp with available boost, very high input impedance and low output impedance. The delay line is switched on separately in parallel with the dry signal. A tone trimmer allows the user to fine-tune the repeats, going from bright digital to tape-like to darker narrow-bandwidth analog-style repeats. Unlike many PT2399 delays, there is very little treble filtering otherwise, and combined with some progressive bass loss, the repeats take on a shimmery, singing quality that is not readily found in other delays. The circuit also contains what is, to my knowledge at the time of this design, a unique method of limiting the signal in the PT2399 via a single LED, making it impossible to distort the delay chip.

My goals with this design:
1. Bright repeats without too much noise
2. Tails and a transparent dry path that people won't mind having on all the time
3. High headroom.
4. Low parts count with common parts and values where possible (for the most part ...)
5. 1590B-sized or smaller

Background:
Not too long back, I built a friend of mine the Joshua Tree Echo, a multipedal with the Multiplex, Fatpants, and Rub-A-Dub. I had also promised another friend a delay for Christmas, but I didn't want to build him just anything.

While I was working on the JTE, I had breadboarded a delay circuit so I could work on some of the mods. I also had a couple transistor stages already sitting on the board, so I decided to use them instead of breadboarding an op amp, and that led to working out a way to actually get the delay to work with transistors ...  and I realized immediately that this meant I could come up with something, well, a little different. So after the JTE was done, I started tweaking the delay, working on the filtering and coming up with ways to make the repeats as bright as possible without the PT2399 distortion, then adding a charge pump to increase the headroom of the preamp, and eventually I had something that solves a lot of the issues I've had with PT2399 delays.

Today I verified a prototype PCB. It fired up the first time with no issues, and it only took me an hour and a half to populate, including the pots. Yay! The final version of the PCB will have a very slightly revised layout but won't require another round of prototyping.


Here's a link to the build doc:
https://docs.google.com/document/d/1P7uOYFRH6ZW5eGoquhAiMQMyUU76bod_gNKkiSlPNJM/edit

The build doc contains the perboard layout, a primitive etch image, the bill of materials, notes on mods and substitutions, and an extensive "How It Works" section.

The schematic:
(https://dl.dropbox.com/u/9878279/Jon%20Patton%27s%20layouts/Hamlet/Hamlet%20Delay%20Schematic.png)

Here's the PCB:
(https://dl.dropbox.com/u/9878279/Jon%20Patton%27s%20layouts/Hamlet/Hamlet%20Delay%20PCB.png)

Here's the demo again so that everything's in the same place:

wow, jon,
that's freakin' sweet~

finally a pt2399 that don't suck!! ;)

well done!!!!!

Cool!
Something very loosely based on this will probably make it into the amp I'm building ATM (by ATM i mean the amp build that has been going on for about a year now :icon_biggrin:)
Still not entirely sure what the tone control does... does it change the amount of treble cut on each successive repeat?

(http://www.mediafire.com/conv/9dcdb4ee4f83cfe4d156d9f39b5b8eebdbbe9c833e5d142c16839b94388063516g.jpg)
As far as I know, you shouldn't use junctions in the kind of points I coloured red. Junctions are for connecting more than 2 wires meeting. Components are connected to wires just like that - no need for junctions there. Btw. If you don't, you should use Net function rather, not Line/Wire.
You should use junction in the point I coloured blue, if the wires are meant to be connected.
Anyway, the idea is great, and from what I see in the video, works just as well. Shame that only half of the pots are outside. ;p At least the tone pot should also be on the outside, I think.

Very nice job Jon. Certainly a few ideas that will probably end up in my next PT2399 project.
The tone control is a nice addition. Did you consider bringing it out of the box?
I assume it will work OK if we run VA at +9V, but I guess you are using the charge pump to give the preamp more headroom? Seeing as the PT2399 can only cope with 5V signal, I can see why the clipping diode became real important.

Edit: Shouldn't pin 5 of the charge pump be grounded, or is that only the MAX1044?

Quote from: meffcio on February 04, 2013, 02:36:46 AM
As far as I know, you shouldn't use junctions in the kind of points I coloured red. Junctions are for connecting more than 2 wires meeting. Components are connected to wires just like that - no need for junctions there. Btw. If you don't, you should use Net function rather, not Line/Wire.
You should use junction in the point I coloured blue, if the wires are meant to be connected.
Anyway, the idea is great, and from what I see in the video, works just as well. Shame that only half of the pots are outside. ;p At least the tone pot should also be on the outside, I think.

Eagle can be funny about connections when using the line tool sometimes. I threw in junctions to be safe because I got tired of going back to the schematic to add them. :/ I had a few components that just wouldn't stick together, though weirdly the one you circled in blue worked without a junction (both of those pins are in fact grounded in the PCB).

Anyone's free to mount the tone pot outside. I just prefer an uncluttered panel and I wouldn't change it once I had it set. And it's easier to turn a trimpot into an external pot than vice versa. ;)

samhay: pin 5 apparently does not need to be grounded in either chip for a basic +18v operation. And according to a recent post it might cause the charge pump to whine. I just did it the way it's done in tons and tons of madbean projects and the like.

As far as running it at 9v, the +18v also helps ensure that the single transistor has enough output for the split signal. (You can't use a high-gain transistor like a Darlington because it will freak out the PT2399 -- oscillation, not distortion. I'm not sure why.) But yes it's also a headroom thing. I was able to overdrive a 9v transistor setup biased to 4.5v with a booster. And since using the charge pump gave both headroom and enough gain, whereas using a second clean gain stage would have only given more output, it seemed the right call here. I thought pretty hard about using a charge pump in the design, since they're a fairly expensive part and I usually avoid those in designs I release. :)

The limiter LED is just for protecting against boosted signals -- the output on the emitter is below unity with the guitar, so even a high-output humbucker by itself wouldn't overdrive it even without the diode.

Quote from: midwayfair on February 04, 2013, 08:35:50 AM
Eagle can be funny about connections when using the line tool sometimes. I threw in junctions to be safe because I got tired of going back to the schematic to add them. :/

There is an error check function (Erc) in schematic editor, which will point out suspicious junctions (besides the other errors)..

Quote from: Thomeeque on February 04, 2013, 09:36:22 AM

Quote from: midwayfair on February 04, 2013, 08:35:50 AM
Eagle can be funny about connections when using the line tool sometimes. I threw in junctions to be safe because I got tired of going back to the schematic to add them. :/

There is an error check function (Erc) in schematic editor, which will point out suspicious junctions (besides the other errors)..

Yes, I've used it, and sometimes the net isn't created when using the line too. Is there a compelling reason to change it if it's functioning, or is this just telling me for next time I create a schematic? Or is it actually confusing for other people reading the schematic? p.s. not trying to be hostile, I guess I just don't understand why it's even come up.

Quote from: midwayfair on February 04, 2013, 09:42:10 AM
Yes, I've used it, and sometimes the net isn't created when using the line too.

These situations are pointed out by Erc too (unfortunately I don't have example now). It does not happen to me when drawing lines but usually when I connect segments by moving parts (or segments) around - sometimes it attaches them correctly, sometimes it does not. I use Erc after these actions (or just drag part and move it around to see if all connections follow) and clean it immediately, no big deal IMO.

Quote from: midwayfair on February 04, 2013, 09:42:10 AMIs there a compelling reason to change it if it's functioning, or is this just telling me for next time I create a schematic? Or is it actually confusing for other people reading the schematic?

It makes schematic just little bit harder to read to me but it's not that bad (circuitry is still readable correctly), I would go with "next time" :)

Quote from: Thomeeque on February 04, 2013, 09:55:44 AM

Quote from: midwayfair on February 04, 2013, 09:42:10 AM
Yes, I've used it, and sometimes the net isn't created when using the line too.

These situations are pointed out by Erc too (unfortunately I don't have example now). It does not happen to me when drawing lines but usually when I connect segments by moving parts (or segments) around - sometimes it attaches them correctly, sometimes it does not. I use Erc after these actions (or just drag part and move it around to see if all connections follow) and clean it immediately, no big deal IMO.

Quote from: midwayfair on February 04, 2013, 09:42:10 AMIs there a compelling reason to change it if it's functioning, or is this just telling me for next time I create a schematic? Or is it actually confusing for other people reading the schematic?

It makes schematic just little bit harder to read to me but it's not that bad (circuitry is still readable correctly), I would go with "next time" :)

okay, thanks! When I get home I will see if I can just delete the junctions without changing anything on the board.

Great work!!

Sounds great Jon, nice work.
Not to derail the thread but I always enjoy your demos if only for your guitar playing, do you play in a band or anything be interesting to hear it.

Quote from: slacker on February 04, 2013, 12:39:29 PM
do you play in a band or anything be interesting to hear it.

This is the sort of derail I'm perfectly fine with!

The band linked in my sig (Midway Fair) is on a more or less permanent hiatus until some stuff gets sorted out with a couple band members, but our records are free downloads. I'm hoping we'll be able to practice some more in the Spring and put out an EP early next year, because the last four songs we worked on are really, really good and definitely need the full band treatment since our keyboard player did a lot of the arranging. I also play mandolin with a bunch of people in the Baltimore area, most recently an Americana group Whale Show. I've hada solo EP being in the works for what seems like an eternity. I recorded it over the summer; it's just not done mixing. And I'm doing FAWM this year, so ironically I'll have a solo "record" done before the EP ... :P

Quote from: midwayfair on February 04, 2013, 08:35:50 AM
samhay: pin 5 apparently does not need to be grounded in either chip for a basic +18v operation. And according to a recent post it might cause the charge pump to whine. I just did it the way it's done in tons and tons of madbean projects and the like.

Cool. I have only used charge pumps for bipolar supplies, so I have learned something today

Quote from: midwayfair on February 04, 2013, 08:35:50 AM
As far as running it at 9v, the +18v also helps ensure that the single transistor has enough output for the split signal. (You can't use a high-gain transistor like a Darlington because it will freak out the PT2399 -- oscillation, not distortion. I'm not sure why.) But yes it's also a headroom thing. I was able to overdrive a 9v transistor setup biased to 4.5v with a booster. And since using the charge pump gave both headroom and enough gain, whereas using a second clean gain stage would have only given more output, it seemed the right call here. I thought pretty hard about using a charge pump in the design, since they're a fairly expensive part and I usually avoid those in designs I release. :)

The limiter LED is just for protecting against boosted signals -- the output on the emitter is below unity with the guitar, so even a high-output humbucker by itself wouldn't overdrive it even without the diode.

Fair enough. Ultimately, it sounds good, so I guess it is/was worth it.

Oh, and like Slacker, I always enjoy your playing. I too play sans plectrum - just not as well as you unfortunately - and enjoy trying to rip off your riffs.

Quote from: midwayfair on February 04, 2013, 12:54:16 PM
The band linked in my sig

Doh, right there in front of my eyes, or it would be if I didn't have sigs turned off.

Quote from: slacker on February 05, 2013, 12:33:04 PM

Quote from: midwayfair on February 04, 2013, 12:54:16 PM
The band linked in my sig

Doh, right there in front of my eyes, or it would be if I didn't have sigs turned off.

Huh! I didn't even know people could do that. It's www.midwayfair.org. :)

errr.. Back to the delay discussion now.  :icon_mrgreen:

Can I use 7660 SCPA for charge pump?

Quote from: soundclone on February 06, 2013, 09:40:16 PM
Can I use 7660 SCPA for charge pump?

As long as you connect pin 1 to pin 8. As noted in the build doc, if you do this substitution, you'll also want to use a 10v Zener.

I tend to use the LT1054 because it's less likely to blow up. It's just worth the extra $1 to know I won't have to replace it if my power supply spikes.

OK TANKS. But i can't find LT1054 in local store. ???

Quote from: soundclone on February 08, 2013, 11:01:03 AM
OK TANKS. But i can't find LT1054 in local store. ???

I usually pick a couple up if I have to order stuff from Mouser. I don't even try to buy most stuff locally ... too difficult to find most parts.

Im relatively new to delays, so excuse my ignorance... I see you mentioned this is compatible with the taptation controller. Taptation's documentation states a delay time up to ~1100 ms. Does that mean it can essentially double the delay time on the PT2399? Or would this require using two 2399s?

Quote from: Brossman on April 27, 2013, 05:39:25 PM
Im relatively new to delays, so excuse my ignorance... I see you mentioned this is compatible with the taptation controller. Taptation's documentation states a delay time up to ~1100 ms. Does that mean it can essentially double the delay time on the PT2399? Or would this require using two 2399s?

The taptation uses a 100K digital pot; since the PT2399 works by more resistance = more delay, using a larger value pot (with no limiting resistor like I use here) will mean more delay.

In reality, I've never heard a PT2399 that can get above about 800 milliseconds.

I'm not the right person to ask about the taptation, though. I suggest finding the thread in the digital subforum.

I was honestly after Taptation for the modulations it can provide for the delay repeats.  However, it seems that the singing/ringing quality obtained by your extensive filter removal can be mildly close to that, albeit lacking "swirl", yes? Conversely, I suppose I could experiment with running a phaser of some sort AFTER the Hamlet, thus modulating the repeats, no?

Quote from: Brossman on April 28, 2013, 03:37:13 PM
I was honestly after Taptation for the modulations it can provide for the delay repeats.  However, it seems that the singing/ringing quality obtained by your extensive filter removal can be mildly close to that, albeit lacking "swirl", yes? Conversely, I suppose I could experiment with running a phaser of some sort AFTER the Hamlet, thus modulating the repeats, no?

You couple splice in an effects loop on the repeats (taken after the decoupling cap on Q2) if you want some strange form of modulation on the repeats like tremolo or phasing. If you want the normal vibrato effect of modulated delay, you would just add an LFO modulation circuit on the delay pin. There are many, many examples of delay modulation for the PT2399 if you search for them.

The filtering of the repeats here does not create modulation.

Sounds great! Nice guitar playing too!  ;D

Okay, gonna try to add an LFO to my build on perf. 

I'd like to know if you think this build would be seriously hindered by NOT using film caps and metal oxide resistors? I've been wondering because the electronics store in town has Mylar caps, but I can get WIMA/CDE metal films from Mouser (cheaper), and I have Carbon film resistors, so buying metal oxides would be another throw down. I mean, this is a really cheap build (maybe $50), factoring in buying MULTIPLES of all components, enclosure, etc. I could cut my cost, etc etc by buying from Mouser...

Bottom line, how would the fidelity of this be affected based on the choice of cap and resistor materials?

Thanks

Quote from: Brossman on April 30, 2013, 04:43:37 PM
Okay, gonna try to add an LFO to my build on perf. 

I'd like to know if you think this build would be seriously hindered by NOT using film caps and metal oxide resistors? I've been wondering because the electronics store in town has Mylar caps, but I can get WIMA/CDE metal films from Mouser (cheaper), and I have Carbon film resistors, so buying metal oxides would be another throw down. I mean, this is a really cheap build (maybe $50), factoring in buying MULTIPLES of all components, enclosure, etc. I could cut my cost, etc etc by buying from Mouser...

Bottom line, how would the fidelity of this be affected based on the choice of cap and resistor materials?

Thanks

You should be able to build this for less than $35, probably closer to $25 including the PCB if you paint your own enclosure and get parts from Tayda.

I would use metal film for R3 at least. You probably won't notice a difference for most of the other resistors in this build. The other one that matters, the drain resistor on Q2, is a trim pot, though you could replace it with a fixed metal film and maybe reduce a very tiny amount of noise. But this isn't a high-gain effect, so unless you've just got horrible parts, the dry path will be fine, and nothing to do with resistors and caps will make a PT2399 less a less noisy chip except more filtering (i.e., the tone control).

Caps are caps, with a couple exceptions: As noted in the build document, the 10uF filtering cap on the PT2399 should ideally be tantalum for best performance (this is explained in the build document), and the 100nF from +9v to ground should be film because film is best at filtering very high frequencies (that's why it's there in the first place). Many of the caps in this build are not part of the audio path. It's nice to use box film caps because they're 5% tolerance.

Have you looked at Tayda?

Thanks for all of this... Im looking at about $30 for the build, but Im buying pt2399 in bulk, extra charge pumps and JFETS, and extras of most of the resistors and caps while im getting a few... I like the idea of tight tolerances and being able to hand pick parts if I so desire. The other things are also considerably harder to source locally, not to mention more expensive. I should not have to pay $2 for an NTE equivalent JFET single... thats absurd.

Not sure if you know, but Mouser has a cheaper residential shipping op thats like... $5-6, so for small orders, Im still paying about 1/3 of what I could locally. Mind you, I get dealer-break prices at the shop cuz they know Im hard up and Im a frequent customer so...

I also planned on using WIMA films for the tolerance and size.  :-\

Thanks for the tips!

EDIT::  I had a look at Tayda lastnight... wow. They've got almost everything!! And for really good prices too.  Can't beat 50x 2n3904s for $1 ;D

Hey Jon,

Im trying to squeeze this build into my (pre-drilled) enclosure and ai noticed im having trouble with larger pots... Ive looked at Tayda for the Mix and Tone controls (both external is the goal). Tayda only has 20ka pots - will using these adversely affect my range of control..?  Thanks.

Quote from: Brossman on July 01, 2013, 09:41:43 PM
Hey Jon,

Im trying to squeeze this build into my (pre-drilled) enclosure and ai noticed im having trouble with larger pots... Ive looked at Tayda for the Mix and Tone controls (both external is the goal). Tayda only has 20ka pots - will using these adversely affect my range of control..?  Thanks.

20k should be fine for the mix. Use one of the 30k trims for the tone (I do), or go with a 50k if you're making it external. Bigger value = more available rolloff.

Did a small redesign on this after learning a few things from the Cardinal V2.

(https://dl.dropboxusercontent.com/u/9878279/Jon%20Patton%27s%20layouts/Hamlet/Hamlet%20Delay%201.3.png)

(10K for the bias trim might be overkill.)

Kinda neat that you can share the collector of a BJT with the drain of a FET. The biasing is more complicated but it can be dialed in by ear by moving the trims a little at a time, and you get a little more overall control over the maximum volumes of each. I also added an optional second base resistor, which would let the same layout use any type of transistor.

This also gets rid of the slight signal loss that came from mixing the signals after the decoupling caps in the older version.

I haven't fully tested it and I don't really have time to do so at the moment, but a preliminary test sez it's sound.

I just finished building version 1.2 on vero.

I'm getting strange noises when in the tails bypass mode, the delay and tone knobs affect the noise.

Seems like the PT2399 doesn't like FS1 hanging open.

I was able to eliminate the noise while in tails bypass by grounding FS1. I tried grounding FS1 with the stomp switch but had a pop echoing whenever the switch was opened or closed. 

I eventually gave up the tails and went with true bypass and so far everything seems fine.

Any suggestion on how to get the tails bypass quiet?

Dunno. I've built a dozen of them on PCBs are perf board without it and this is the first I've heard anyone having the problem.

The source and PT are usuallypretty close in voltage but ground is way different, hence the pop. There are other ways to do tails, like swithing in a very large resistor instead of fully disconnecting, or switching a gigantic cap to ground at pin16 so you just don't get sound but don't DC ground it.

If I think of something else I'll let you know but the oscillation mint indicate some other problem. Is the veto layout you used verified? Which layout is it?

> share the collector of a BJT with the drain of a FET.

"Plate mixer". RDH 3rd.

It's a pretty sad thing with vacuum triodes. Better with pentodes, but that spoils the cost and simplicity. Ought to work great with BJT or FET. Using both (and especially with the variable bias on BJT) may be more clever than elegant or mass-repeatable.

Quote from: PRR on August 12, 2016, 03:04:31 PM
> share the collector of a BJT with the drain of a FET.

"Plate mixer". RDH 3rd.

It's a pretty sad thing with vacuum triodes. Better with pentodes, but that spoils the cost and simplicity. Ought to work great with BJT or FET. Using both (and especially with the variable bias on BJT) may be more clever than elegant or mass-repeatable.

Ah. I neglected to update this thread. I remembered over at Madbean: That experiment was a bust. It sounded fine when clean but when pushed even a little bit, it was just terrible.

It does work great with only FETs (which is what I did in the Cardinal), and this design can be built with two of them if one is willing to sacrifice some input headroom and redirect the 4M7 to ground. You get far more accurate high frequency response in exchange for the headroom. Ironically it ends up sounding less like the (FET-based) EP3 that inspired it than the BJT version. I didn't feel the need to change anything in the end.

Two BJTs with their collectors joined is equally awful. I also tried MOSFETs but that didn't work out at all.

Quote from: midwayfair on August 12, 2016, 03:11:52 PM

Quote from: PRR on August 12, 2016, 03:04:31 PM
> share the collector of a BJT with the drain of a FET.

"Plate mixer". RDH 3rd.

It's a pretty sad thing with vacuum triodes. Better with pentodes, but that spoils the cost and simplicity. Ought to work great with BJT or FET. Using both (and especially with the variable bias on BJT) may be more clever than elegant or mass-repeatable.

Ah. I neglected to update this thread. I remembered over at Madbean: That experiment was a bust. It sounded fine when clean but when pushed even a little bit, it was just terrible.

It does work great with only FETs (which is what I did in the Cardinal), and this design can be built with two of them if one is willing to sacrifice some input headroom and redirect the 4M7 to ground. You get far more accurate high frequency response in exchange for the headroom. Ironically it ends up sounding less like the (FET-based) EP3 that inspired it than the BJT version. I didn't feel the need to change anything in the end.

Two BJTs with their collectors joined is equally awful. I also tried MOSFETs but that didn't work out at all.

I never noticed this project (but started a few years ago)...Kudos Jon, really nice work!

I have played around with this mixing scheme in simulation and implemented a variant of it in a stereo-to-mono mixing box for an old tube audio receiver.

What worked out acceptably well for the stereo-to-mono mixer was to use the Sziklai pair to compose a "super transistor", and of course a healthy-sized (4.7k) emitter resistor on each node.  From a 15V rail I was able to achieve about 6Vpp clean headroom which is about double what I needed to drive my mono receiver at maximum volume.

Using the Sziklai configuration the "collector" voltage has much less effect on the overall gain of the stage so the distortion is very low until the point that it actually starts clipping.  Because of this the inter-modulation between the two separate signals is much lower up until the point at which it starts clipping.

It wasn't phenominal but not bad in simulation.  I was simulating something like 70 dB signal/distortion ratio on the 6Vpp sum, which amounts to 0.03%. I avoid using the term "THD" because I was measuring 2 different signals mixing at a common node so the distortion figure includes both harmonic and inter-modulation distortion.

To my ears it was perfectly acceptable, and probably a lot lower distortion at the actual levels I was using to listen to music.  I would guess for more practical levels the distortion was below the specs on my mp3 player output.

Anyway if you get bored and want to take this a step further you might try cooking something up with some Sziklai pairs.  You can also couple N-ch JFET with PNP BJT if you want the convenience of ground-reference biasing -- which actually works really well with a JFET because it's operating much closer to pinch-off, which drives the source voltage much higher so you can handle larger input signals.

Thanks Jon, I designed the vero myself.

http://guitar-fx-layouts.42897.x6.nabble.com/Hamlet-Delay-by-Jon-Patton-td32472.html

I had 1st tried the perf layout (my 1st attempt at perf) didn't go well. Low output and the tone control created smoke. Didn't attempt to troubleshoot as I found the perf process frustrating.  (I'm a long time vero builder)

Anyway, I'll go over the vero layout again to see if something is amiss. Somehow it sounds fine with true bypass.

And I'm sure you've heard this many times, thanks again for all the great videos and designs.     

Turns out, the tone control (and playing) was masking the noise.

I am getting the oscillation with the effect engaged, without playing, even without the charge pump installed.

Most noticeable with the tone control and mix on full. It sounds like a clock or LFO type noise. The delay control effects the noise, almost sounds like dialing in a old radio or CB.

Won't be able to troubleshoot until tomorrow...

Check that Q1 isn't backwards. The pedal will oscillate if it is.

I was hoping that was it, but no. Q1 oriented correctly.

Went through the layout and version 1.2 build doc and they seem to match. I couldn't find any errors with my populated board either.

Audio probe at PT2399 pin 15 sounds fine, pin 14 has the noise.

It seems the noise appears when not playing and with delay resistance above 40k. Delay fully CW the noise sounds like motorboating. 7 of 8 PT2399s have the same noise (2 or 3 suppliers). Some have noise with the delay knob turned all the way down too.

Is that a clue or does this indicate 7 bad PT2399s? 

1 of the 8 PTs has much less noise, only a slight motorboating at full CW. I think if I go with a 50k pot and no parallel resistor, it may be suitable. 

I don't really know. I do know that PT2399s can vary pretty wildly especially from seconds-sellers like Ebay folks, but usually even at worst you can get get to 25K before the noise becomes noticeable -- it's certainly extremely abnormal for the motorboating to appear with the delay at minimum even with the tone control at 0 here.

Unfortunately we're sort of at the mercy of the chip itself for a lot of things and the chip variability can complicate troubleshooting, but I'm sure you already know that. :(

Sorry, the noise with delay at minimum wasn't quite motorboating, more of a white noise, hiss kind of thing.

The motorboating only happened with the delay full on, or close to full on. As I backed off the Delay control, the motorboating would stop but there was a white noise and sometimes other strange noises until I got to about 40k. Then they would quiet down, but a couple of them would be noisy again at minimum.
 
One of them seems to be working well with a 50k delay pot. I double checked the tails bypass with this "quieter" pt, but there is still noise when not playing. Frustrating. Makes me think I have two issues going on, poor PTs and a build error. And I think I already mentioned this, but if I ground the FS1 connection, the noise in tails bypass goes away.

I think at this point I'm going to order a bunch of PTs from Tayda and in the meantime build another vero to see what I get. 

     

Quote from: maoriente on August 15, 2016, 01:20:39 AMmore of a white noise, hiss kind of thing.

Again, hiss is pretty normal when you start getting above 200mS of delay, but having it at minimum with the tone resistance pretty much anywhere on the dial would be weird. The only thing I can suggest there is double check that you've used 56nF for the tone cap and not 5.6nF. But considering you still have some unusual behavior when the bypass is disconnected, I'm more inclined to suspect the chip.

One thing that might help is if you build a comparison on a known working layout. Unfortunately, while I know of at least two etch layouts (other than the etchable version of my perf layout), I don't remember if anyone else did a build on vero. When I get home I'll try to remember to look through my circuitboard box and see if I still have a PCB (which you can have). If I don't, I'll see if I can get a comparison build to you some other way. Are you in the U.S.?

Wow, too kind Jon. I'm in Utah.

Recently, I have mostly been using monolithic caps. Some are a bit out of spec so I measure each one and usually end up doubling them up trying to get the exact schematic values. For the tone cap in this build, I paralleled a 473 and a 103 to get the 56n.

I re-cut and re-flowed all the joints yesterday just because. Played it again tonight using headphones and there is definitely some strange noises in the background. No change using a battery supply.   

Even with the best sounding PT, and with the guitar volume turned all the way down, there is a shift in the noise while adjusting the delay pot. It quiets down close to full CCW, and then there is a little bump in noise at full CCW, not nearly as loud as full CW. Adjusting the tone control just masks the noise.         

I just finished up a madbean Dirtbaby (pcb) tonight, I increased the filtering quite a bit per the build doc and also increased C13 from 4n7 to 47n. I already had a seperate PT2399 installed, I'm not getting any of those strange noises on this one. I swapped in a couple of the suspect PTs and they were acceptable. The Dirtbaby is much darker though, but even with the tone control full CCW on the Hamlet, the noise levels are night and day between the two.   

Conclusion: there must be something wrong with my build and not necessarily the PTs.

I'm traveling the rest of this week but will try to knock out another vero the following week to compare.

I'll see if I can record and post the noise before I take off tomorrow. If not, I'll do so when I get back.



 

Haven't had a chance to build another one just yet, hopefully over the weekend, but here is a sound sample.

I recorded direct to pc from headphone jack of amp, also recorded a Dirtbaby for comparison. (I've increased the filtering on the Dirtbaby making it darker on the repeats that with the stock values)

Hamlet: Max Delay, Max Mix, Max Tone, Min Repeats
https://soundcloud.com/maopedals/hamlet?in=maopedals/sets/delays (https://soundcloud.com/maopedals/hamlet?in=maopedals/sets/delays)
Bypass 0-3 sec
Effect On 3-10sec
Guitar Volume turned down 10sec
Rotate Delay knob slowly CCW 15-24sec
Rotate Delay knob slowly CW 24-33sec
Rotate Delay back and forth a little quicker 33-45sec

Dirtbaby: Max Delay, Max Blend, Min Repeats, Rate & Depth Min
https://soundcloud.com/maopedals/dirtbaby?in=maopedals/sets/delays (https://soundcloud.com/maopedals/dirtbaby?in=maopedals/sets/delays)
Bypass 0-3 sec
Effect On 3-10sec
Guitar Volume turned down 10sec
Rotate Delay knob back and forth 10-27sec

As usual, Jon was right on.

I replaced the 56n tone cap and the noise is gone! 

I had originally paralleled 2 multilayer ceramics in the same holes to get the 56n, I replaced with a single box cap and the noise is gone. My guess is I had a bad solder joint with the 2 caps-per-hole.

I still have some noise with the tails bypass though, I assume I may have the same issue with a couple of the other caps I had paralleled. It's a sputtery kind of white noise with the tails bypass.

Going to dive back in with an audio probe after lunch. 

So while audio probing in tails bypass, I found noise at the collector of the 2n3904. The base and emitter are clean. The same noise makes it through the 1u to lug 3 of the volume trimmer.

When I remove the charge pump, the noise goes away. I tried a couple of 7660s and a LT1054 with pin 1 lifted. No change.

Same noise when powered with 9v battery too.

The noise is also on PT2399 pin 14, and the Drain, Source and Gate of the 2n5457

Not sure what to make of it. Maybe the charge pump is to close to the PT2399?

Here is the layout I used.
(http://guitar-fx-layouts.42897.x6.nabble.com/file/n32472/Hamlet.png;cid=1469929273565-451)

this picture before I swapped out the 56n tone cap(s)
(http://guitar-fx-layouts.42897.x6.nabble.com/file/n32472/Hamlet_pcb.png;cid=1469929273565-451)

I would expect charge pump noise to be a high pitched whine, and I think that's what I'd also expect if proximity were a problem. That's really not the noise I expected from the PT either. If it's coming from the PT it makes sense that it's at the collector of Q1 -- it would just go through the cap back to the collector. You could try lifting the 1uF decoupling cap from Q1 if you want to be able to probe the outputs of Q1 and Q2 separately.

If it is the charge pump doing it, I am not sure what else could be done (especially because I haven't encountered it before). Maybe increase the 10uF filtering capacitor.

I increased both 10u, neither affected the noise.

So here is the latest:
With FS1 and FS2 closed, the noise is barely detectable. I think there may be some noise way in the background, but it's hard to hear even with headphones.

When I disconnect FS1 and FS2, there is a noticeable increase in hiss. I get these weird noises while adjusting the delay knob too. If I ground FS1, the hiss is somewhat reduced, but still noticeable compared to shorting FS1 and FS2.

Not sure what to make of that.

When I remove the charge pump, the noise seems to be gone altogether. Or it's just so faint I can't tell its there.

The Delay without the charge pump still sounds pretty good, maybe just not as sharp. 

At this point, I think I'm going to run this particular vero at 9v, re-bias Q2 and call it a day. 

I'm going to redesign the vero and try again.

Thanks for all your help Jon.   

     

Hello,
I dare to ask a quite dumb question, but i have started studying the (very interesting !) Hamlet Delay Build Doc (https://docs.google.com/document/d/1P7uOYFRH6ZW5eGoquhAiMQMyUU76bod_gNKkiSlPNJM/edit), but, as english is not my native language, i am not positive about the meaning of one sentence :

QuoteC9 provides some final high-pass filtering. Increasing it will retain some additional bass and subharmonic frequencies, but no fundamental guitar information is lost even with a comparatively small capacitor in this position.

Does "retain" actually mean that the capacitor will allow the bass to pass through (like "the bass will be kept in the source signal, remain intact"), or does it mean that the capacitor will block the bass, and doesn't allow it to pass through ?  :icon_redface:

I am trying to understand how to further hipass filter a delayed signal, so i am trying to figure out which capacitor value I should have for C3 (input cap) and C9.
- Should I focus more on C3, or on C9 ?
- Why value should i choose in order to filter out the frequencies below 100-150hZ ?

Thank you !

Quote from: snk on January 27, 2019, 04:30:28 PM
Hello,
I dare to ask a quite dumb question, but i have started studying the (very interesting !) Hamlet Delay Build Doc (https://docs.google.com/document/d/1P7uOYFRH6ZW5eGoquhAiMQMyUU76bod_gNKkiSlPNJM/edit), but, as english is not my native language, i am not positive about the meaning of one sentence :

QuoteC9 provides some final high-pass filtering. Increasing it will retain some additional bass and subharmonic frequencies, but no fundamental guitar information is lost even with a comparatively small capacitor in this position.

Does "retain" actually mean that the capacitor will allow the bass to pass through (like "the bass will be kept in the source signal, remain intact"), or does it mean that the capacitor will block the bass, and doesn't allow it to pass through ?  :icon_redface:

I am trying to understand how to further hipass filter a delayed signal, so i am trying to figure out which capacitor value I should have for C3 (input cap) and C9.
- Should I focus more on C3, or on C9 ?
- Why value should i choose in order to filter out the frequencies below 100-150hZ ?

Thank you !

Increasing that capacitor will cut less bass. However, that bass is not in a guitar signal.

If you want to cut more bass (high pass), then you could make that capacitor smaller.

"- Why value should i choose in order to filter out the frequencies below 100-150hZ ?"

Use a calculator like this one:
http://sim.okawa-denshi.jp/en/CRlowkeisan.htm

Or, better, experiment by ear.

Thank you very much, Midwayfair (as well as for the Hamlet design !), and sorry for the average understanding of your sentence  :icon_redface: :icon_wink:

QuoteIf you want to cut more bass (high pass), then you could make that capacitor smaller.

Would you advise to decrease C3 first, and then experiment with lower values for C9, or the opposite ?

QuoteUse a calculator like this one:
http://sim.okawa-denshi.jp/en/CRlowkeisan.htm

Using the Hipass calculator (http://sim.okawa-denshi.jp/en/CRtool.php) (it's a hipass filter, right ?), it gives me that the values used in Hamlet set a cutoff point around 72Hz. Is that correct ?
If i set the resistor value to 110nF, it should have the cutoff point set around 144Hz, which seems fine (unless i'm mistaken) ;)

QuoteIncreasing that capacitor will cut less bass. However, that bass is not in a guitar signal.

I will use it with synths and drum machines :)

Quote from: snk on January 28, 2019, 01:16:07 AM
Would you advise to decrease C3 first, and then experiment with lower values for C9, or the opposite ?

C3 will affect the bass for the whole delay signal, INCLUDING the first repeat.

C9 will affect every signal AFTER the first repeat.

If you ALWAYS want less bass in the delay, make C3 smaller. Try 47nF.

Thank you very much, Midwayfair.
That is very interesting to know.

Will the change have a big influence on the feedback behaviour (like getting faster into self oscillation, or not being able to reach self oscillation) ?

Am I right assuming that a 47nF capacitor with a 10K resistor will make a 330Hz HP cutoff point ?

Hey Jon, hope all is well.

What's the recommended way to get 10db gain for the dry signal?

Increase the 4k7 collector resistor or decrease the 1k emitter resistor?

Or something else?