Harbinger tone

[HailToTheBlues]

I guys,
So i'm finishing an Harbinger One project, and all it takes to close the enclosure at last is to adjust the offset and the gain trimmers to get my sound. One of the tones i would like to replicate would be the "any colour you like" univibe tone gilmour as, just for fun, but the harbinger doesn't seem to be fast enough, and it's not intense enough, the effect on full speed sounds like a subtle vibrato at the high frets, and gilmour's is very present and "dramatic". I'm also comparing it to a mini dejavibe demo from youtube, since they look very similar. About the details, the cells are faced to the lamp, but i'm using a shield as well, they are all matched. So guys, is this normal, or is it some issue? It could be from my cheap amp, or the cells.

Best regards

[madbean]

It sounds like it is not working quite right. It should not be subtle by any means (although one interesting thing I have noticed about Univibes is that if you play one for a while, your ear gets pretty used to it so it seems less intense than when you first turn it on).

Anyway, one thing you can do is decrease the value of R46. This was originally designated as 47k, and I have seen more than one schematic list it as such. But there does seem to be more of a consensus that it should be 4k7, which is the value listed in the BOM of the current release. That may increase the overall intensity for you. If it still is not working correctly then you might need to dig into checking voltages, trying different photocells, etc.

To increase the max speed, simply lower the values of R44 and R41. There are a couple of modern repo's that go down to 2k2 or 1k8. Just don't set it too low or the bulb may not be able to respond fast enough.

[HailToTheBlues]

Thanks for the suggestions brian, really helpfull. Just another question i forgot to ask: how could i know if one of the cells got damaged during the soldering and other stuff, or even arrived broken, would it even work with just 3 cells, or would it have sound but not the sound a univibe should have? And how could i spot this?

[RobA]

Thanks for the suggestions brian, really helpfull. Just another question i forgot to ask: how could i know if one of the cells got damaged during the soldering and other stuff, or even arrived broken, would it even work with just 3 cells, or would it have sound but not the sound a univibe should have? And how could i spot this?

Unless I'm looking at this all wrong, you should get some modulation with just two of the LDR's working. You should be able to see if the LDR's are all working by measuring the resistance across the LDR's pins with the pedal powered up. If you put it on a slow enough speed then you should be able to track theresistence change to some degree with a DMM.

...
Anyway, one thing you can do is decrease the value of R46. This was originally designated as 47k, and I have seen more than one schematic list it as such. But there does seem to be more of a consensus that it should be 4k7, which is the value listed in the BOM of the current release. That may increase the overall intensity for you. If it still is not working correctly then you might need to dig into checking voltages, trying different photocells, etc.
...

Mine has the 47k in it now. It does function quite nicely with that value in there. But, I'm going to be modding it to take a CV control for the Speed anyway, so I'll try changing the value of that resistor and report back what it does. I should be able to get it to 4.7k by clipping a 5.2k across the 47k as a simple test.

[RobA]

I finally got to do the mods to my Harbinger. So these are the results.

Moving to 4.7k for R46 does seem to help to me. I think it helped in being able to dial the gain into a better setting.

I put R41 and R44 at 1k. I did this because I can use the CV control section to put the minimum resistance where I want it to be. But, I did try it with the normal 100kC pot in there first. 1K is too small. It gets too fast for the lamp to even respond and then going further CW just locks up the LFO. I'd guess that a 2.7k resistor for these two would be about as fast as you'd want it to go. With the CV controlled board in there, I have the smallest resistance at about 3.3k and that goes plenty fast for what I want.

The depth and strength of the vibe or chorus stayed good and strong up to the point where the lamp really couldn't respond fast enough, so it should be fine to reduce those two resistors to get a faster LFO going.

[HailToTheBlues]

I finally got to do the mods to my Harbinger. So these are the results.

Moving to 4.7k for R46 does seem to help to me. I think it helped in being able to dial the gain into a better setting.

I put R41 and R44 at 1k. I did this because I can use the CV control section to put the minimum resistance where I want it to be. But, I did try it with the normal 100kC pot in there first. 1K is too small. It gets too fast for the lamp to even respond and then going further CW just locks up the LFO. I'd guess that a 2.7k resistor for these two would be about as fast as you'd want it to go. With the CV controlled board in there, I have the smallest resistance at about 3.3k and that goes plenty fast for what I want.

The depth and strength of the vibe or chorus stayed good and strong up to the point where the lamp really couldn't respond fast enough, so it should be fine to reduce those two resistors to get a faster LFO going.

I still haven't tried to reduce those resistors to get it to go faster, but i'll try it too someday. I'm still not satisfied with it, i tried to reduce the 47k resistor to 4,7k and it started to work better, the lamp was brighter, and sounded more intense, but as i roll the speed control all the way, it also looses intensity, it seems like a vibrato, not close to the strong effect of gilmour in "Any colour you like", and i think if i get it to go faster, it'll be even more subtle. So my question is, is this what is suppose to happen, is it suppose to be a little more subtle at high speeds, or should i try to swap my photocells? Or could this be just a matter of trying to tweek the gain and offset a little further?

[RobA]

One thing that definitely happens is that as the speed of the LFO gets really high, the lamp can't respond as well. Certainly, by the time the rate is at 50Hz, no lamp is going to have any variation at all. You can see that just in the way a normal light works with an AC power supply. Where the drop off in response is going to hit and how much probably depends on the specific lamp used.

I modded mine to have a pot in place of R11 and R12. Actually, I replaced R11 and R12 with two 47k resistors and a 100kΩ pot. This lets you dial in the chorus notches to various depths. I think that the chorus is stronger with the preamp out of the circuit (modern mode). I still always use it in modern mode because I think it sounds better overall. I'd also note that my thinking that the chorus is stronger in vintage setting could be purely my imagination influenced by other tone differences. I also did the Geofex mod that subs in 10µF caps for all the 1µF caps in the phase shifter section. It's subtle, but I do think it sounds stronger. The mod I'd recommend most is the depth pot thing.

I have no idea if this means anything, but Gilmourish suggest that the EMS Synthi Hi-Fli might have been used to do the Leslie like sounds on Any Colour You Like.

[TNblueshawk]

I'm always leary of trying to replicate something I hear on a record. The one thing you can't account for is who knows what they did in the studio and if there is one band that did a crap ton of stuff in the studio after recording a song it was Pink Floyd. So whether it got tweaked after Gilmour recorded Any Colour who knows for sure.

Just a thought FWIW.

[HailToTheBlues]

I'm always leary of trying to replicate something I hear on a record. The one thing you can't account for is who knows what they did in the studio and if there is one band that did a crap ton of stuff in the studio after recording a song it was Pink Floyd. So whether it got tweaked after Gilmour recorded Any Colour who knows for sure.

Just a thought FWIW.

Yeah, exactly the way i think, the only thing i was saying is that my univibe doesn't feel strong at the highest speed, and maybe that's the way they work, i don't know

[TNblueshawk]

I'm always leary of trying to replicate something I hear on a record. The one thing you can't account for is who knows what they did in the studio and if there is one band that did a crap ton of stuff in the studio after recording a song it was Pink Floyd. So whether it got tweaked after Gilmour recorded Any Colour who knows for sure.

Just a thought FWIW.

Yeah, exactly the way i think, the only thing i was saying is that my univibe doesn't feel strong at the highest speed, and maybe that's the way they work, i don't know

I've built 3 vibes, a Warhead being one of them from Bean, which were great and had plenty of warble like it should. I've not yet built a Harbinger but intend to since I sold my Warhead.

You might want to post some pics if you haven't already in the troubleshoot section and see if anyone sees anything funky that has built one. Could be worth a shot.

Oh and +1 on All Hail to the Blues ! 

[midwayfair]

I'm always leary of trying to replicate something I hear on a record. The one thing you can't account for is who knows what they did in the studio and if there is one band that did a crap ton of stuff in the studio after recording a song it was Pink Floyd. So whether it got tweaked after Gilmour recorded Any Colour who knows for sure.

Just a thought FWIW.

Yeah, exactly the way i think, the only thing i was saying is that my univibe doesn't feel strong at the highest speed, and maybe that's the way they work, i don't know

There are multiple things going on here.

One is the lamp. If you've used the specified lamp, you'll be fine for the entire speed range of the pedal for normal operation.

The other is the photocells, and this requires really careful parts selection, because they have several different properties, none of which are particularly easy to measure. But this is one place you might want to experiment if the rest of the circuit is definitely correct and you have already been very careful about dialing in the trim pots.

1. Switching range. This is the total range, in the circuit, between fully on and fully off for the photocells.

-Minimum on resistance: You want a low on resistance -- 10K is okay, 5K is great, 1K is spectacular. Lower resistance means that more out of phase signal gets mixed in, which creates a more intense effect. There is actually something you can do in-circuit to help if your photocells have too high a dark resistance, though! See those 47K resistors in series with the photocells? Make them smaller, e.g. 22K. You can even jumper them.

-Maximum OFF resistance: This is the opposite -- you also need a point where the out-of-phase signal is pretty much completely taken out of the circuit. Since the photocells are beating against an impedance of about .. 38K in each stage, you need a photocell that can reach multiple megaohms to achieve that. While it's not too hard to find a cell with 1M maximum dark resistance (like Tayda's), those cells won't get anywhere near the maximum in the actual circuit, especially when the bulb is going fast and not going completely dark. So you have to go higher ... The Silonex 7532 was pretty popular, but as Smallbear noted, at some point the construction technique changed and they went from being a cell with 20M, 50M, or even higher to rarely reaching 10M. So Smallbear had a version made called the "hi-dark" that hits 200M when fully off. You won't HIT fully off, but you will get way higher resistance on the dark cycle if you use those. There's not really anything you can do in the circuit that will fix things if your dark resistance isn't high enough.

2. And the rise and fall times.

-The Rise time is how long it takes for a photocell to reach some percentage of its maximum dark resistance. Just like how your eye might have an afterimage when exposed suddenly to a bright light, a photocell will retain some light memory. A faster rise time means that the photocell will "forget" the lamp's brightness -- this is absolutely critical to retaining a deep effect at faster speeds. For reference, one of the fastest devices out there, the VTL5C1, still only has a Rise time of 30mS (and I haven't found anything like that in a discrete photocell, though I really wish I could!). The hi-dark that Smallbear carries is more like 70mS, I think. That's only about 14hZ before the light blinking on and off has almost no effect. While the Univibe's maximum speed is probably only about 10Hz (and it might even be more like 8Hz stock), that's still getting pretty close even in a photocell that's basically made to drop into a Univibe.

-Fall time is how quickly the photocell drops to its lowest resistance. This depends on whether the light is bright enough to actually drop it there. Remember how I said that you want an extremely low resistance? Well, now you have to contend with the fact that the light needs to be bright enough to drop it there and on long enough to actually get it there. So you want to add this time to the Rise time ... if you get anything more than 100mS, you're at the point where the effect will be very shallow at faster speeds. You're fighting physics in that case and no changes in the circuit will help you.

I haven't even touched on proximity, which can vary according to photocell.

And this is just a tiny portion of why this can be such a complicated build ...

I don't know how to measure rise and fall times, because it's hard to set up good testing conditions. I think you could make a small box with some tiny holes in it and a light inside, put the leads through with the cell on the inside, close the box, and measure the leads, but your multimeter will have to be very accurate.

[HailToTheBlues]

I'm always leary of trying to replicate something I hear on a record. The one thing you can't account for is who knows what they did in the studio and if there is one band that did a crap ton of stuff in the studio after recording a song it was Pink Floyd. So whether it got tweaked after Gilmour recorded Any Colour who knows for sure.

Just a thought FWIW.

Yeah, exactly the way i think, the only thing i was saying is that my univibe doesn't feel strong at the highest speed, and maybe that's the way they work, i don't know

There are multiple things going on here.

One is the lamp. If you've used the specified lamp, you'll be fine for the entire speed range of the pedal for normal operation.

The other is the photocells, and this requires really careful parts selection, because they have several different properties, none of which are particularly easy to measure. But this is one place you might want to experiment if the rest of the circuit is definitely correct and you have already been very careful about dialing in the trim pots.

1. Switching range. This is the total range, in the circuit, between fully on and fully off for the photocells.

-Minimum on resistance: You want a low on resistance -- 10K is okay, 5K is great, 1K is spectacular. Lower resistance means that more out of phase signal gets mixed in, which creates a more intense effect. There is actually something you can do in-circuit to help if your photocells have too high a dark resistance, though! See those 47K resistors in series with the photocells? Make them smaller, e.g. 22K. You can even jumper them.

-Maximum OFF resistance: This is the opposite -- you also need a point where the out-of-phase signal is pretty much completely taken out of the circuit. Since the photocells are beating against an impedance of about .. 38K in each stage, you need a photocell that can reach multiple megaohms to achieve that. While it's not too hard to find a cell with 1M maximum dark resistance (like Tayda's), those cells won't get anywhere near the maximum in the actual circuit, especially when the bulb is going fast and not going completely dark. So you have to go higher ... The Silonex 7532 was pretty popular, but as Smallbear noted, at some point the construction technique changed and they went from being a cell with 20M, 50M, or even higher to rarely reaching 10M. So Smallbear had a version made called the "hi-dark" that hits 200M when fully off. You won't HIT fully off, but you will get way higher resistance on the dark cycle if you use those. There's not really anything you can do in the circuit that will fix things if your dark resistance isn't high enough.

2. And the rise and fall times.

-The Rise time is how long it takes for a photocell to reach some percentage of its maximum dark resistance. Just like how your eye might have an afterimage when exposed suddenly to a bright light, a photocell will retain some light memory. A faster rise time means that the photocell will "forget" the lamp's brightness -- this is absolutely critical to retaining a deep effect at faster speeds. For reference, one of the fastest devices out there, the VTL5C1, still only has a Rise time of 30mS (and I haven't found anything like that in a discrete photocell, though I really wish I could!). The hi-dark that Smallbear carries is more like 70mS, I think. That's only about 14hZ before the light blinking on and off has almost no effect. While the Univibe's maximum speed is probably only about 10Hz (and it might even be more like 8Hz stock), that's still getting pretty close even in a photocell that's basically made to drop into a Univibe.

-Fall time is how quickly the photocell drops to its lowest resistance. This depends on whether the light is bright enough to actually drop it there. Remember how I said that you want an extremely low resistance? Well, now you have to contend with the fact that the light needs to be bright enough to drop it there and on long enough to actually get it there. So you want to add this time to the Rise time ... if you get anything more than 100mS, you're at the point where the effect will be very shallow at faster speeds. You're fighting physics in that case and no changes in the circuit will help you.

I haven't even touched on proximity, which can vary according to photocell.

And this is just a tiny portion of why this can be such a complicated build ...

I don't know how to measure rise and fall times, because it's hard to set up good testing conditions. I think you could make a small box with some tiny holes in it and a light inside, put the leads through with the cell on the inside, close the box, and measure the leads, but your multimeter will have to be very accurate.

Thanks for the time you took for that explanation, really helpful!

So if i intended to measure the minimum and maximum resistance of the photocells i have installed, which by the way, are from a univibe kit from Banzai, could i do it with them on circuit? And how would i do it?

About the rise and fall time, i really doubt i could measure them, my multimeter is really basic, and i don't think it is fast enough to measure that.

In case i decide to go for those photocells from smallbear you referred, which ones are you talking about:                         http://smallbear-electronics.mybigcommerce.com/photocell-silonex-advanced-photonix-hi-dark/
or http://smallbear-electronics.mybigcommerce.com/photocell-silonex-advanced-photonix-hi-dark-to-5/ ?
And also, could i ask for a kit with 4 of those matched to those values you referred? Or would i need to get a couple and match them myself, which would be much more expensive, i think.

[midwayfair]

Measuring the photocells would be almost impossible in-circuit. To do that, you'd need the cover off, which means light is getting on the photocells, so there's no way you can measure the dark resistance or the actual in-use light resistance.

Umm... Silly question, but you ARE testing the effect with the circuit in total darkness, right? I usually put a cardboard box over my test rig and turn out the lights.

In case i decide to go for those photocells from smallbear you referred, which ones are you talking about:                         http://smallbear-electronics.mybigcommerce.com/photocell-silonex-advanced-photonix-hi-dark/
or http://smallbear-electronics.mybigcommerce.com/photocell-silonex-advanced-photonix-hi-dark-to-5/ ?
And also, could i ask for a kit with 4 of those matched to those values you referred? Or would i need to get a couple and match them myself, which would be much more expensive, i think.

Those are the same photocell. If you're planning on dunking your Univibe in slightly acidic water fairly often, you might want the hermetically sealed version. (But I get those because they look cool.)

Re the Banzai cells, I'm not even sure what matching would entail for this effect. I suppose one could match response times, but I sincerely doubt that's what Banzai's doing because it would be very difficult. You could match dark resistances, but that doesn't strike me as meaningful. I guess you could match them to some range for on/off resistances, but they don't tell you what they've matched them for.

[HailToTheBlues]

Measuring the photocells would be almost impossible in-circuit. To do that, you'd need the cover off, which means light is getting on the photocells, so there's no way you can measure the dark resistance or the actual in-use light resistance.

Umm... Silly question, but you ARE testing the effect with the circuit in total darkness, right? I usually put a cardboard box over my test rig and turn out the lights.

In case i decide to go for those photocells from smallbear you referred, which ones are you talking about:                         http://smallbear-electronics.mybigcommerce.com/photocell-silonex-advanced-photonix-hi-dark/
or http://smallbear-electronics.mybigcommerce.com/photocell-silonex-advanced-photonix-hi-dark-to-5/ ?
And also, could i ask for a kit with 4 of those matched to those values you referred? Or would i need to get a couple and match them myself, which would be much more expensive, i think.

Those are the same photocell. If you're planning on dunking your Univibe in slightly acidic water fairly often, you might want the hermetically sealed version. (But I get those because they look cool.)

Re the Banzai cells, I'm not even sure what matching would entail for this effect. I suppose one could match response times, but I sincerely doubt that's what Banzai's doing because it would be very difficult. You could match dark resistances, but that doesn't strike me as meaningful. I guess you could match them to some range for on/off resistances, but they don't tell you what they've matched them for.

Couldn't i put some tape on the photocells and measure them? Would that work? Or is better to desolder them? Also, to measure the minimum on dark resistance, i measure them completely in the dark, but what about the maximum off resistance?

About the matching, you were saying earlier that getting photocells with 1k was better for a good sound, so are those smallbear photocells close to that, or do i need to search for those values in various photocells of those? Or could i even ask someone from smallbear to get me 4 of those with 1k minimum resistance (if i can even do that)?

EDIT: And yes, when i change anything of the trimmers or measure something, it's completely in the dark, but from what i've seen around the web, it's not that silly of a question, believe me!

[midwayfair]

I've tried a ton of ways of actually covering a photocell to cover it, with kind of poor results. Electrical tape seems to still bleed through some light. This might be more trouble than it's worth, but some heat shrink over electrical tape might work. In the end, it could be simpler to desolder them.

Those photocells will be more like 5-10K for light resistance under most circumstances. It's low enough and in line with what I've heard of people measuring in vintage units. Like I said, if you want to fake a lower light resistance, you could always lower those 47K resistors.