Need a bigger brain! Question about LT1054

[Govmnt_Lacky]

I am using an LT1054 as the heart of a 9V to 12V charge pump in a circuit and I suspect that it is adding some unwanted frequencies and noise into the output. So, my question is....

Can anyone see a way to change the frequency of the oscillator like you can with the MAX1044 chips (Freq Boost)?

Link to datasheet:

http://cds.linear.com/docs/en/datasheet/1054lfg.pdf

Thanks for the help 

[midwayfair]

Huh. I wish I knew how to help, but the clock frequency (25KHz) in the LT1054 should be waaaay above human hearing. This is worrying because I really like using it ever since the forum taught me it was safer.

[tomfoy]

Take a look at the Texas Instruments datasheet for this chip. I think I remember seeing something about setting the oscillator frequency in it.

[pinkjimiphoton]

i had a whine on the first klon(e) i built... if i remember correctly, grounding i think it was pin 7 directly took care of the whine.

sorry greg, i don't remember which was either vref (pin 6) or osc (pin 7) that did the trick. i think it was pin 7...

i don't have that klon(e) anymore, so i can't check.. but according to this on the forbidden planet, yah, pin 6 or 7...

http://.......org/viewtopic.php?f=13&t=1256&start=440#p180878

yah, pin 7. if it whines, short that to ground, that should take care of it bro.

[armdnrdy]

Put a 5-25pf cap between pins 2 and 7.

You can socket the cap until you find the "magic" value, then install permanently.

[chromesphere]

Havent tried the cap idea suggested by armdndry (have to give that a shot!) but i can tell you first hand the LT1054 in certain circuits will produce excessive whining.  FSH1 could be a candidate havent tried swapping that one yet, but fairly certain the LT isnt filtering well enough.  FL301 is DEFINITELY a candidate.  I swapped that one over from an LT1054 to a TC1044(scpa -> the low noise type).  Reduced noise significantly.  But alas, still some noise...

The LT in fuzzes and even some modulations can be fine though.  If i could find the TC1044 at a reasonable price i would just get them.  Havent had any reliaiblity issues with the TC, like the LT. 

I wouldnt recommend a MAX to my accountant.  They are terribly unreliability imo.

Paul

[pinkjimiphoton]

Put a 5-25pf cap between pins 2 and 7.

You can socket the cap until you find the "magic" value, then install permanently.

larry, why a cap between 2 and 7?

[blueduck577]

yah, pin 7. if it whines, short that to ground, that should take care of it bro.

Careful!

larry, why a cap between 2 and 7?

From page 6 of the datasheet:

The frequency can be lowered by adding an external capacitor (C1, Figure 2) from Pin 7 to ground. This will increase the charge and discharge times which lowers the oscillator frequency. The frequency can be increased by adding an external capacitor (C2, Figure 2, in the range of 5pF to 20pF) from Pin 2 to Pin 7.

Don't ground it as that will lower the oscillator frequency into the audio range. 

[pinkjimiphoton]

weird!! worked good on the klon, only way i could get the whine to stop.

[Govmnt_Lacky]

Can anyone comment on using the TC1044SCPA in a circuit? Couldn't find the current output limits on the datasheet. I "think" the circuit I am using is pulling somewhere between 15-20mA.

[Scruffie]

Can anyone comment on using the TC1044SCPA in a circuit? Couldn't find the current output limits on the datasheet. I "think" the circuit I am using is pulling somewhere between 15-20mA.

What are you trying to power? And what proximity is the pump to the circuit and have you used shielded wire.

[Govmnt_Lacky]

What are you trying to power? And what proximity is the pump to the circuit and have you used shielded wire.

Trying to power a Deluxe Electric Mistress clone at 12V. I have verified that the circuit has "some" noise but it is tolerable however, when I run an LT1054-based pump circuit, the noise gets much worse. The wiring is not affecting the noise as it is barely audible when I directly inject 12V regulated power into the circuit and bypass the pump portion of the circuit.

I experimented with the power input electro cap and discovered that it needs to be larger however, I am still wondering if the LT1054 is still injecting noise. This is still at the experimental stages and I would just like to rule out the pump as a source of noise.

[Morocotopo]

Gov, you may be having heterodyning problems. Two oscillators (the LFO one and the one in the charge pump) running at different speeds can produce sum and difference "beating" frequencies through the power or ground. I know because years ago I had that problem with a 301 flanger that I added a charge pump to, to be able to use it with a 9V power supply. I solved it by adding lots and lots of caps to the power (something like ten 100 uF´s. Really funny looking).

[Scruffie]

What are you trying to power? And what proximity is the pump to the circuit and have you used shielded wire.

Trying to power a Deluxe Electric Mistress clone at 12V. I have verified that the circuit has "some" noise but it is tolerable however, when I run an LT1054-based pump circuit, the noise gets much worse. The wiring is not affecting the noise as it is barely audible when I directly inject 12V regulated power into the circuit and bypass the pump portion of the circuit.

I experimented with the power input electro cap and discovered that it needs to be larger however, I am still wondering if the LT1054 is still injecting noise. This is still at the experimental stages and I would just like to rule out the pump as a source of noise.

The LT1054 has a clock, the mistress has one, you'll have to be careful where it is located to prevent heterodyning which may be the noise you're experiencing.

The LT1054 chip itself is not noisy, i've used them loads as have others, it's layout, cross talk and power conditioning that are causing it to work noisily with the circuit most likely. So don't give up on it yet, add more filterng and try and move the charge pump out of the way of anything, shield the wires from it, keep them short etc. a regulated 12V source doesn't need the same treatment as a circuit doubling the voltage.

Edit: Morocotopo got there first with the heterodyning suggestion. It can certainly add some strange noise when you experience it.

[Govmnt_Lacky]

Thanks for the suggestions about heterodyning.

Ill try to parallel a 1000uF electro and see what it does  

Should the larger cap go between power and ground BEFORE or AFTER the charge pump? When I did some experimenting, I noticed the greatest change in noise when I paralleled an electro before the pump input.

If I remember correctly, the Mistress clock starts at ~40KHz. If I were to replace the pump with a TC1044SCPA and added the boost, this would cause the pump oscillator to run at 45KHz. Would this cause BIGGER problems since they would be soo close? 

[Morocotopo]

Thanks for the suggestions about heterodyning.

Ill try to parallel a 1000uF electro and see what it does  

Should the larger cap go between power and ground BEFORE or AFTER the charge pump? When I did some experimenting, I noticed the greatest change in noise when I paralleled an electro before the pump input.

If I remember correctly, the Mistress clock starts at ~40KHz. If I were to replace the pump with a TC1044SCPA and added the boost, this would cause the pump oscillator to run at 45KHz. Would this cause BIGGER problems since they would be soo close? 

I guess the big cap should go after, but try it in both locations, can´t rememeber where I had it since it was so long ago... and that pedal is no longer in one piece.

Regarding the frequencies, my (limited) knowledge is that the produced frequencies are sum and difference of the two original ones, so for example if you have a 45 kHz and a 40 kHz, you could have heterodyning at 45-40=5 kHz (ouch! right in the middle of the audible spectrum), 40+45= 85 kHz (only problematic if you are a bat), and so on (all the harmonics).

[Govmnt_Lacky]

Thanks Ariel! 

So, with what has been said so far.... If the charge pump is going to be mounted on the same PCB as the circuit...

Is there any way to get rid of, or at least minimize the oscillator interference that would be coming from the pump circuit?

Also, the 35KHz (found the correct frequency elsewhere on the forum) is the MINIMUM clock frequency used my the DEM. When the Range pot is adjusted, the frequency gets much higher (I believe it can go up to and even above 500MHz) Not sure of the exact numbers right now but, I believe it would be a range just like any other flanger-type circuit (MXR M117, etc.)

[armdnrdy]

Sorry for the delay in my reply. Sick with the FLU!!!!

Take a look at the better detailed Linear Technology data sheet.

http://cds.linear.com/docs/en/datasheet/1054lfg.pdf

[GGBB]

Just speculating, but if difference frequencies are the issue here, then raising the oscillator frequency might not solve the problem.  According to the spec, the nominal oscillator frequency is 25kHz which would produce difference frequencies of 10k and upward as the the range pot is adjusted (assuming the 35kHz is right).  If you raised the osc freq to 40k, then you'd have differences from 5k down to zero then back upward as the pot is adjusted.  If you raise it to anything above 35k, there will be some sections of the range pot where the difference frequencies will be audible, since the range pot will sweep its frequency from 35k up to the osc freq and beyond.  On the other hand, if you lowered the osc freq to 15k or 20k, you'd have differences of 20k or 15k on upward.  While 15-20k is still technically audible, it might not be in this circuit and could probably be filtered out without too much negative impact. The spec isn't clear exactly how high or low you can adjust the frequency however.  Again - just playing around with the math and speculating here - so maybe none of this applies in the real world?

[Govmnt_Lacky]

Just speculating, but if difference frequencies are the issue here, then raising the oscillator frequency might not solve the problem.  According to the spec, the nominal oscillator frequency is 25kHz which would produce difference frequencies of 10k and upward as the the range pot is adjusted (assuming the 35kHz is right).  If you raised the osc freq to 40k, then you'd have differences from 5k down to zero then back upward as the pot is adjusted.  If you raise it to anything above 35k, there will be some sections of the range pot where the difference frequencies will be audible, since the range pot will sweep its frequency from 35k up to the osc freq and beyond.  On the other hand, if you lowered the osc freq to 15k or 20k, you'd have differences of 20k or 15k on upward.  While 15-20k is still technically audible, it might not be in this circuit and could probably be filtered out without too much negative impact. The spec isn't clear exactly how high or low you can adjust the frequency however.  Again - just playing around with the math and speculating here - so maybe none of this applies in the real world?

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.