Actions
Forum Thread Details
  • State Not Answered
  • Replies 11 replies
  • Subscribers 240 subscribers
  • Views 1767 views
  • Users 0 members are here
Related
See a helpful answer?

Be sure to click 'more' and select 'suggest as answer'!

If you're the thread creator, be sure to click 'more' then 'Verify as Answer'!

How does the oscillating frequency of a charge pump make a difference?

BigG

I recently purchased some Microchip TC7660S charge pumps as I plan to use them to generate a -12V DC output from a +12V DC input.

Based on the datasheet, the circuit seems simple enough.

However, I am rather puzzled by Pin 1, which the datasheet refers to as the Switching Frequency boost pin. Now, why would you need to boost the oscillating frequency - I'm guessing this is for when you want to generate a positive voltage multiplier.

Then on the bottom of page 10 (section 5.4 Changing the TC7660S Oscillator Frequency) it then talks about slowing the oscillating frequency using the following method:

For example, the addition of a 100 pF capacitor between pin 7 (OSC) and pin 8 (V +) will lower the oscillator frequency to 1 kHz from its nominal frequency of 10 kHz (a multiple of 10), and necessitate a corresponding increase in the values of C 1 and C 2 (from 10 μF to 100 μF).

So, as a learning exercise I'm also trying to fathom when does one decrease the frequency. I believe a clue is given in the datasheet, but not still not sure why one would want to, that you should only slow down the frequency for low currents. I'm given very little else.

I'm sure there's good reason for all of the above.

I look forward to receiving some insight.

Thanks.

Colin.

  • 0

    The higher frequency will allow you use small capacitors, while low frequency will require larger capacitors, but also reduce EMI.

  • 0
    BigG said:
    Now, why would you need to boost the oscillating frequency

    5.4 Changing the TC7660S Oscillator Frequency
    It may be desirable in some applications (due to noise or other considerations) to increase the oscillator frequency. 

    45 KHz is outside the audio range. For audio devices it may be good to have that frequency out of the operational bandwidth.

  • 0 in reply to Jan Cumps

    Ah, that's good to know. Yes, I can imagine that's important for audio (and noise sensitive analog) applications.

  • 0 in reply to genebren

    Ah yes, I suspect that should you have an EMI frequency spike that's above acceptance limits, this offers you an option to adjust the switching frequency to remove or reduce this EMI noise.

  • 0

    Just speculating .... Charge pumps can transfer "fixed" charge on each cycle of the clock. At higher frequencies the amount of charge and thus average current is increased (more charge transfers per second) which allows a heavier load and lower ripple. There may be higher switching losses at higher frequencies and some components may get hotter so you would only boost the frequency when you need the extra current.

  • 0

    One of the videos in my capacitor project shows the effect, on voltage, ripple and current, of increasing the charge pump oscillator frequency....

     Polymer Cap Charge Pump 

  • 0 in reply to dougw

    Thanks for sharing. Great project, btw, and the video was very helpful.

  • 0

    Hi Colin,

    It will affect the output. Surprisingly the TC7660S PDF datasheet doesn't mention the output waveform in any detail, but the TC7662B PDF datasheet does, and the same info will apply. There's a formula there too.

    image(Image source: TC7662B datasheet)

    Incidentally, if the 12V input is unregulated and might be a bit higher, then the TC7662B may be a better choice, since it supports up to 15V (whereas 12V is the limit for the TC7660S). However the TC7660S seems to be a lot lower-cost, so would be the better option if the input is stable.

    I'm presuming this is to drive an op-amp or similar? If so, it could be worth leaving some space on the PCB for optional filtering, in case you find you need it. The filtering could be RC (either a ferrite bead creating resistance at noise frequencies, but it could also just be a small value resistor, say 1 to 10 ohm, and a capacitor such as 10 uF), or it could involve a negative voltage regulator, or both. The 7908 (e.g. 79L08 is available in a small TO-92 package) might be OK, but might not be, since the TC7660 output will drop a bit under load, and the 7908 might come out of regulation. An LDO would be a better choice. You'll get excellent results by doing this. Or, you could just leave pads for patching in a regulator if the need arises. Anyway this is assuming that a full -12V is not needed. If you do need very close to -12V then that limits options a bit.

    A nice way to move some problems to later, is to use zero-ohm resistors in the path between circuit modules, so that things can be easily isolated for testing, or for inserting additional circuitry. I frequently do that just to make life easier when testing.

    image

  • 0 in reply to shabaz

    Thanks for this. I hadn't spotted the TC7662. Gosh yes, I see that it's double the price. One wonders if that's a fair price to pay just to squeeze out that little bit extra on Vmax. As you say, you fly very close to wind, if the power supply is slightly more than 12V as max voltage on TC7660 is 13V. I'm certainly going to study the TC7662 datasheet as it looks to provide more technical background info.

    One application I was aiming to use the charge pump in was to generate a ±12V PWM signal using a PWM signal from a MCU.  Here I used a differential comparator rather than an op amp: www.ti.com/.../lm211.pdf

    image

    Mind you, I have wondered what sort of signal response I'd get if using mosfet switches etc. instead. Here's the output I get with the chargepump + comparators - I'm using my new toy (SCOPY)

    image

  • 0 in reply to BigG

    Hi Colin,

    Regarding the MOSFET idea you mention, an option could be to use a 'opto-isolated gate driver'.

    For example the part code below is UCC23513 which is £0.66 in single quantities (available at some distributors, unfortunately not Farnell, but there should be other similar-ish options). It is an all-in-one solution, containing the MOSFETs. 

    image

    The left side is an LED so could directly connect to the PWM from the microcontroller (with say a 100 ohm resistor to protect the LED).

    The right side VCC and VEE would connect to the +12V and -12V supply rails. The PWM would be directly usable on the Vout pin. 

     Anyway, maybe this is overkill, perhaps the PWM doesn't need to be so sharp-edged! Still, it's a cheap-ish solution (and does protect the microcontroller a lot), although the comparator method is low-cost too.

    Also, if you find the 7660 doesn't have enough output power (I don't know what you're driving), then a ready-built DC-DC converter module could be an option too (they cost about £2.50, i.e. a bit more than the TC7660S, but a lot more output current capability.