Need a bigger brain! Question about LT1054

[garcho]

While 15-20k is still technically audible...

Depends how long you've been gigging with drummers

[GGBB]

Depends how long you've been gigging with drummers

Hey Gord!

So, are you suggesting that trying the small-ish cap between Pins 2 and 7 of the LT1054 to lower the pump's oscillator frequency might be worth a shot?

I do have some MAX1044 and TC1044SCPA pumps here. I may try those as well just to see if I get any noise changes.

I think anything is worth a shot - you'll never know until you try it.  According to the datasheet, a cap between 2 and 7 will raise the frequency whereas a cap from 7 to ground will lower it.  Both are pretty easy to try.  I was suggesting that raising might not work, whereas lowering it might.  But it all comes down to what the noise is that you are hearing.  If the osc freq is 25k, then you aren't hearing that, so maybe it's the difference frequency that Morocotopo mentioned.  If it is the difference between the mistress range freq and the osc freq, then the noise should change as you adjust the range pot.  Does this happen?  It should be audible only at the low end of the control because once the range pot gets you up to 45k and higher, the difference freq goes above audible.  At min, the range control is 35k (according to you), which produces 10k difference freq with the 25k osc freq.  10k is probably audible.  To move difference frequencies outside the audible range, you need to separate the two frequencies by ideally more than 20k without having either frequency below 20k.  Lowering the LT1054's 25k osc freq to 15-20k gets you close to this, probably close enough for a pedal.  But I have to stress that all this is purely mathematical speculation on my part - the only way to find out is to try it.

[Govmnt_Lacky]

If it is the difference between the mistress range freq and the osc freq, then the noise should change as you adjust the range pot.  Does this happen?

ABSOLUTELY! When running on either the charge pump OR with 12V regulated I get the "Tune in Tokyo" effect in flange mode. The noise changes with the rotation of the Range pot. The Color and Speed pots have their indicated effects as well (Color = Louder. Speed = Faster/Slower). Of course, when in Filter mode, the noise is "static" and constant depending on where the Range pot is set. There are a few "sweet spots" where I can get a good clean flange/filter signal but most of those appear to be lumped towards the middle of the Range pot rotation (from 10 o'clock to 2 o'clock)

What is killing me is that the noise I am hearing is WAY WORSE when using the charge pump circuit like I described above. When running on the 12V regulated, the noise subsides considerably!

It should be audible only at the low end of the control because once the range pot gets you up to 45k and higher, the difference freq goes above audible.  At min, the range control is 35k (according to you), which produces 10k difference freq with the 25k osc freq.  10k is probably audible.  To move difference frequencies outside the audible range, you need to separate the two frequencies by ideally more than 20k without having either frequency below 20k.  Lowering the LT1054's 25k osc freq to 15-20k gets you close to this, probably close enough for a pedal.  But I have to stress that all this is purely mathematical speculation on my part - the only way to find out is to try it.

I'll try some smallish caps (5pF to 22pF) between the pump's pin 7 to ground and see where that leads me. I also still want to try a TC1044SCPA in there and see what happens (with and without the boost)

Thanks for the replies everyone. More info to follow later 

[Govmnt_Lacky]

Here is a link to another thread that I started that deals with the situation I am in right now. There is a short video in one of the posts that shows exactly what type of noise I am getting.

Posting a link to the other thread for more info and continuity.

http://www.diystompboxes.com/smfforum/index.php?topic=98903.0

[GGBB]

I'll try some smallish caps (5pF to 22pF) between the pump's pin 7 to ground and see where that leads me. I also still want to try a TC1044SCPA in there and see what happens (with and without the boost)

If the small caps don't seem to do anything, try bigger (100pF - 470pF).  I *think* the idea is to raise the built-in 150pF capacitance so that the cycle time increases (osc freq drops), but there's no guideline as to how much capacitance causes how much freq shift.  I'm not even sure it's clear whether the pin 7 to ground cap is parallel or in series with the internal cap - my guess would be parallel.

Also - thinking way outside my level of knowledge now - could lowering the osc freq increase voltage ripple and therefore require higher filtering caps?  Something to consider perhaps.

[Govmnt_Lacky]

UPDATE--

I tried the TC1044SCPA in the circuit and it did not go very well. Constant hissing and static in the audio even with the Boost feature in use. I am guessing that the current limit of the TC1044 became a factor. Switched back to the LT1054.

Tried different values of caps (5pF to 27pF) between Pin 7 of the LT1054 to GND. This seemed to help quite a bit although not entirely. With the 27pF cap, the noise on the CCW end of the Range pot (highest frequencies) seemed to have dissipated however, there was still noise on the CW end of the pot travel which is where the lower clock frequencies reside. I was playing it safe and going with the recommended values in the datasheet so I did not have anything bigger to try.

Inserting caps between Pins 2 and 7 did not seem to help with the noise and in some cases made it worse.

The next step is to slowly move up, starting at about 33pF, and see if the noise on the CW end of the pot rotation will disappear 

More to follow.....

[armdnrdy]

I took a listen to the noise issue that you are encountering in the link that you provided.

I have used LT1054s in quite a few builds with nothing even close to that occurring. I am well versed in the art of "not leaving any stone unturned" while troubleshooting a problem, but it is my humble opinion that you might be "barking up the wrong tree".

The "classic" charge pump whine that I've encountered is a constant frequency produced by the charge pump and dumped throughout the circuit it is feeding. I have never encountered being able to manipulate the frequency with any of the effects controls. The fact that you state that the problem persists even with a regulated 12 volt supply (albeit a lessoned problem) solidifies my reasoning.

If you haven't already, I would audio probe the circuit to see where the sound in question is entering the signal path.

[Govmnt_Lacky]

I took a listen to the noise issue that you are encountering in the link that you provided.

I have used LT1054s in quite a few builds with nothing even close to that occurring. I am well versed in the art of "not leaving any stone unturned" while troubleshooting a problem, but it is my humble opinion that you might be "barking up the wrong tree".

The "classic" charge pump whine that I've encountered is a constant frequency produced by the charge pump and dumped throughout the circuit it is feeding. I have never encountered being able to manipulate the frequency with any of the effects controls. The fact that you state that the problem persists even with a regulated 12 volt supply (albeit a lessoned problem) solidifies my reasoning.

If you haven't already, I would audio probe the circuit to see where the sound in question is entering the signal path.

The sound in the video is exactly what I hear when I am using the charge pump circuit AND I rotate the Range knob. When I use the rectified 12V, the noise is still there however, it is much, much less pronounced. I don't doubt that the noise is possibly originating from a possible noisy BBD chip.... I am just saying that something with the charge pump circuit is causing it to be WORSE. I was able to filter out about 1/2 of the noise with the 27pF cap from LT1054 Pin 7 to GND. I think that the above statements about heterdyning may be correct in some fashion. I am going to try larger caps this evening to tell.

In the mean time, I do have some new BBDs (all be it from a Hong Kong source) on the way... hoping for an arrival today. I think between squashing the heterodyning issues, some better power filtering, and possibly better and quieter BBDs, this should be a fairly quiet and awesome circuit.

Fingers crossed and more to come!

[Govmnt_Lacky]

PROGRESS REPORT:

I mocked up the original power supply section of the DEM. Schematic below:

http://www.metzgerralf.de/elekt/stomp/mistress/images/deluxe-electric-mistress-v4-schematic.gif

When I bypassed the onboard charge pump circuit and injected the power from the mocked up version I got ABSOLUTLEY NO NOISE AT ALL!! 

So, this tells me that the circuit is good. It also tells me that the Charge Pump circuit is what is creating all of the unwanted noise. 

Next, I am going to breadboard the pump circuit AWAY FROM THE PCB and inject the power from there to see if it is just a matter of keeping the pump away from the circuit pcb.

More to come later (when I order a new breadboard  )

[armdnrdy]

Do you have an image of the layout that you're using for the DEM?

What values are you using for filter caps for the charge pump circuit? Do you have a layout image for that circuit?

[Govmnt_Lacky]

Do you have an image of the layout that you're using for the DEM?

What values are you using for filter caps for the charge pump circuit? Do you have a layout image for that circuit?

I have a PCB layout but it is not mine. Basically, the intention is to work out the bugs and then share with the forum 

As for filter caps, I am using the generic values according to the voltage doubler circuit on the datasheet for the LT1054 Sheet 10.

Something just hit me though... after looking at the PT80 circuit on GGG, I noticed that they DO NOT ground Pin 5 of the MAX1044. Like the LT1054 positive doubler circuit in the datasheet, the MAX1044 says to ground this pin for positive voltage doubling however, the PT80 does not and it works quietly and just fine.

I wonder if removing Pin 5 of the LT1054 from this circuit will help 

[midwayfair]

I wonder if removing Pin 5 of the LT1054 from this circuit will help 

Huh. I've NEVER grounded pin 5. I didn't even notice until you said this that the datasheet has it grounded.

[Govmnt_Lacky]

Huh. I've NEVER grounded pin 5. I didn't even notice until you said this that the datasheet has it grounded.

Yeah... it kinda hit me to look at some other circuits that use charge pumps. I looked at the PT80 and noticed this. I also looked at Brian's Road Rage and he uses it for the negative voltage output but, if you do not hook it up it is essentially left floating.

Fingers crossed!! 

[PRR]

> a 9V to 12V charge pump

Charge-pumps are really, IMHO, best used "no"-load. Substrate bias, static CMOS, etc.

Yes they can power small chips.

But there are more serious approaches.

http://www.pololu.com/catalog/product/799

True, it's $12, but you've spent that much time already.

And it can approach 1 AMP out for 9V in 12V out, far beyond any $1 pump.

[Govmnt_Lacky]

Tried the circuit with the LT1054 Pin 5 removed. STILL NOISY!! 

Next up.... breadboard the charge pump circuit and run it away from the flanger circuit and put the charge pump ground path in parallel to the flanger circuit ground path.

Im now thinking this is either a series shared ground issue OR it is a proximity issue with the flanger circuit being too close to the charge pump's oscillator.

More to follow....

[chromesphere]

Hey Gov, did you end up working out the solution?  I've got a FL301 that has the same sound as in that video you mentioned.  My FL301 did benefit from the low noise TC1044 but unfortunately, its still noisey.  Might have to try some of the suggestions in this thread.
Paul

[Govmnt_Lacky]

Hey Gov, did you end up working out the solution?  I've got a FL301 that has the same sound as in that video you mentioned.  My FL301 did benefit from the low noise TC1044 but unfortunately, its still noisey.  Might have to try some of the suggestions in this thread.
Paul

Sorry Paul but I was unable to get the noise to go away with any kind of pump circuit. My only solution was to incorporate a transformer or run the circuit from a wall wart and incorporate regulation within the circuit.

[chromesphere]

Thanks for the reply gov.  Shame theres no way of getting the charge pump working

All the best
Paul

[Govmnt_Lacky]

Thanks for the reply gov.  Shame theres no way of getting the charge pump working

All the best
Paul

If you figure it out without using about 10 or 15 filtering caps.... 

I'll be the first one to give you mad props! 

[chromesphere]

haha, i think it might be a bit beyond me, but i will definitely post here if i do work it out.

Paul