Parts: 133MHz-16.2kHz programmable oscillator (DS1077)

cover

The DS1077 is a 5volt, 133MHz to 16kHz programmable clock source. The internal frequency divider is configured over a simple I2C interface, and the chip requires no external parts. Not bad for under $2. We used the Bus Pirate to test this chip before using it in a project. Grab the datasheet (PDF) and follow along.

DS1077, $1.69 direct from Maxim + $10 shipping.

This chip isn’t available at any major distributors yet, but Maxim has them for under $2/each with a flat $10 shipping charge. This is an 8pin SOIC surface mount chip, so we made a small breakout board for testing.

Test circuit

schematic-450

Pin connections

Bus Pirate DS1077 (pin)
SCL SCL (8)
SDA SDA(7)
AUX OUT1 (1)
+5volts Vcc (3)
GND GND, CTRL (4,5,6)

We powered the DS1077 from the Bus Pirate’s 5volt power supply. Two resistors, R1 and R2, pull-up the I2C bus to 5volts when it’s not in use. Capacitor C1 is 0.01uF and C2 is 0.1uF, as recommended by the datasheet. Control pins provide some additional functions, but we bypassed them to ground during our test. Output1 is the primary clock signal pin.

Interfacing

Address Purpose
0b10110000 Default base address (0xB0)
0xB0 Write address
0xB1 Read address

We put the Bus Pirate into I2C mode (M, options: I2C, 100kHz). The external pull-up resistors hold the bus at 5volts, so it’s important to leave the on-board 3.3volt pull-up resistors off (default).

I2C>{0b10110000} <– DS1077 write address
210 I2C START CONDITION
220 I2C WRITE: 0xB0 GOT ACK: YES <– got ACK
240 I2C STOP CONDITION
I2C>

First, we broadcast the DS1077’s address and see if it acknowledges. The address of the DS1077 is 1011, plus three programmable bits (000 by default), and the read (1) or write (0) bit. We got an ACK, so we know that the circuit is working and our connections are good.

Address Bytes Register
0x01 2 10 bit clock divider, n+2 (DIV)
0x02 2 Prescaler, CTRL pin functions. (MUX)
0x0D 1 Address select, EEPROM write control. (BUS)
0x3F 0 Save settings to EEPROM (E2)

The DS1077 is controlled by writing values to the locations shown in the table.

I2C>{0xb0 0x0d 0b00001000} , <–write to BUS register
210 I2C START CONDITION
220 I2C WRITE: 0xB0 GOT ACK: YES <–DS1077 write address
220 I2C WRITE: 0x0D GOT ACK: YES <– BUS register
220 I2C WRITE: 0x08 GOT ACK: YES <– BUS register setting
240 I2C STOP CONDITION
I2C>

By default, the DS1077 saves all changes to the EEPROM. We don’t need this during testing, so we disable it by setting bit 3 (0b1000) of the BUS register (0x0d). The first four bits must be left as 0, the last three bits select the address to accommodate multiple DS1077s on the same I2C bus. See datasheet page 7.

I2C>{0xb0 0x02 0b00011000 0b00000000} <–set the 16bit MUX value
210 I2C START CONDITION
220 I2C WRITE: 0xB0 GOT ACK: YES <–DS1077 write address
220 I2C WRITE: 0x02 GOT ACK: YES <–MUX register
220 I2C WRITE: 0x18 GOT ACK: YES <–data byte 1
220 I2C WRITE: 0x00 GOT ACK: YES <–data byte 2
240 I2C STOP CONDITION
I2C>

The MUX register controls the prescalers, CTRL pin functions, and frequency divider.  We disable the prescaler and CTRL pins, and enable the 10bit frequency divider.  The MUX register is explained on page 5 of the datasheet.

Specific frequencies are generated by dividing the 133MHz reference frequency through the prescalers and a 10bit (1025 level) programmable divider.  The clock is divided by the amount specified in the DIV register, plus two. When DIV=0, the output is 133MHz/2=66MHz.

graph

This scheme gives the best frequency resolution in low ranges, and no steps between 133MHz and 66MHz.

I2C>{0xb0 1 0b11111111 0b11000000} <–DIV=1025
210 I2C START CONDITION
220 I2C WRITE: 0xB0 GOT ACK: YES <–DS1077 write address
220 I2C WRITE: 0x01 GOT ACK: YES <– DIV register
220 I2C WRITE: 0xFF GOT ACK: YES <– bits 9:2
220 I2C WRITE: 0xC0 GOT ACK: YES <– bits 1:0
240 I2C STOP CONDITION
I2C>f  <–do a frequency count
9xx FREQ COUNT ON AUX: 16128Hz (16kHz) <– DS1077 frequency
I2C>

We set all the bits in the DIV register to 1 for maximum frequency division. ‘F’ measures the frequency on the AUX pin, which is connected to the DS1077 clock output. With DIV=1025, the frequency is about 16kHz.

I2C>{0xb0 1 0 0} <– DIV=0, 133MHz divide by 2

9xx FREQ COUNT ON AUX: 0Hz <–66MHz, too fast to count

————-

I2C>{0xb0 1 0 0b10000000} <– DIV=2

9xx FREQ COUNT ON AUX: 3339696Hz (33MHz) <–133MHz/4

————-

I2C>{0xb0 1 0b00000001 0b00000000} <–DIV=4

9xx FREQ COUNT ON AUX: 22192384Hz (22MHz) <–133MHz/6

We can play with the divider and generate a range of frequencies. The output is always equal to the reference frequency (133MHz) divided by DIV+2. The Bus Pirate’s input pin is only capable of measuring about 50MHz, so the highest speeds don’t register. A future version of the Bus Pirate should include a gigahertz prescaler for high frequency measurement.

I2C>{0xb0 0x3f} <–write E2 register

Finally, we can write the E2 register (0x3f) to save these setting in the EEPROM. The DS1077 will now return to these settings at power-on.

Conclusion

The DS1077 simplifies complex clock sources by moving a programmable oscillator and frequency divider into a single chip. It isn’t available from distributors, but you can buy it directly from Maxim. If you need better control of high frequencies, check out the DS1085 with 10kHz steps from 133MHz to 8kHz. The DS1085L is a 3.3volt, 66MHz version available at Digikey.

bread-450

Comments

  1. andy says:

    Thank you for these new posts. I’ve been visiting hack a day almost since it started! This is defiantly a new reason not to miss a post. Every one of these posts has been bookmarked!

  2. A. says:

    Thanks for the article… Maybe I will learn how to solder SMD parts :-)

  3. Ian Lesnet says:

    @a- SOIC stuff is a great size to solder and etch at home. It’s the biggest SMD size, and there’s really nothing tricky about it. I think it’s easier to solder than through-hole parts, but don’t forget the flux.

  4. mmalluck says:

    I like Maxim parts. They’re very kind with the free samples too :-)

  5. ngth says:

    It’s also worth noting that there is a +3.3V version available.

  6. sverre says:

    Stupid question maybe?

    What is the point of using 2 decoupling cap? one 0.1 and one 0.01?

  7. Jancans says:

    Stupid (incorect) answer maybe.
    Smallest cap charge and discharge faster, so it smooth vcc

  8. matt says:

    There are two different caps because higher value ones aren’t good at blocking very high frequencies (though theoretically their high value would suggest they do) so you put a lower value one in parallel which, again theoretically, seems useless because contributes almost nothing to the total capacitance, but it can effectively block high frequencies.
    It’s not uncommon to see in well designed circuits power supply lines even 3 caps in parallel, for example one 10 uF, one 100nF and one 100pF, when there are high frequencies or high self oscillation risks involved (switching power supplies, audio amplifiers, rf circuit, etc).

  9. therian says:

    that strange why large caps cant block high frequency, maybe because of internal inductance ?

  10. Andrew says:

    To the guy who loves Maxim, they’re great if you want one or two, but beware if you want to buy anything between that and 50000. They’ve been bad enough that we have a policy at work that nothing sole source from Maxim is allowed unless you can justify a reason to the VP.

    P.S. 10^6 Hertz is abbreviated MHz not mhz, and 10^3 Hertz is kHz, not khz.

  11. Andrew says:

    grr. someone needs to lose the thing that squashes caps to lower case…

  12. jimmys says:

    Ok, so what might it be used for? a sweep generator for testing, wide band frequency hopping data, adjusting for unstable carriers.. what else?

  13. timour says:

    bad news for battery projects,
    this chip is really power hungry.
    it needs about 20ma to operate at 5V.

  14. Odin84gk says:

    “Stupid question maybe?

    What is the point of using 2 decoupling cap? one 0.1 and one 0.01?”

    This is a question I asked for a long time, and I never got a proper answer until I read the following app note:

    http://www.intersil.com/data/an/an1325.pdf

    to make a long story short, the inductance within a capacitor create resonant frequencies (well, frequencies that they are less likely to attenuate). You need two different capacitors along with two different package sizes to attenuate the most amount of noise. If you don’t use different package sizes, you might as well not even bother. (See app note before replying please)

  15. nuerobyte says:

    maxim? they have a realy good sampling policy, $2 may be good but FREE is better.

  16. Dove says:

    We offer programmables from multiple vendors. Request a free sample today. Not too sure if they will let me post a link so just go to google and type in dove programmable oscillators…good luck guys!

  17. tmac says:

    Maybe it is elementary, but, can someone tell me what the resolution is on this chip. I saw where another similar chip had a 10kc resolution, but did not see resolution for this one. Im going to try to get mine working with basic stamp first and then switch to pic chip using C. If anyone already has this worked out for either platform would appreciate shared code.
    Tom -WA4FYN

  18. ben nguyen says:

    In case you can’t get the 3.3v “(L)” type..

    The 5v ds1077 *might* work with a 3.3v i2c bus… the datasheet says the minimum for logic high is .7 of vcc, and minimum for vcc is 4.75, so the minimum for logic high would be 3.325 if you were to supply vcc at 4.75.

    Of course you’d want to make sure whatever you’re connecting to it can tolerate 5v inputs.

    Maxim has the DS1070K / DS3900K kits that could be useful in evaluating these chips:

    http://datasheets.maxim-ic.com/en/ds/DS1070K.pdf

    And according to this nice comparison sheet, the freq step size, at a minimum is 5khz. Which is kinda high if you’re looking for fine tuning low frequencies:

    http://pdfserv.maxim-ic.com/en/an/AN230.pdf

  19. rperkins says:

    I was following along this tutorial. The Bus pirate now uses ‘[' and ']‘ to enclose i2c comands where in this tutorial it is listed as ‘{‘ and ‘}’, which I believe was the case at one time. I am new with the bus pirate.

    http://dangerousprototypes.com/forum/viewtopic.php?f=28&t=1987#p19384

  20. Piratinio says:

    >”The external pull-up resistors hold the bus at 5volts, so it’s important to leave the on-board 3.3volt pull-up resistors off (default).”<

    Hi there,
    why didn't you use Bus Pirates' pull-up resistors?
    Is it because 3900ohm is better than 10kohm there?
    Power supply come by Bus Pirates, so you could use Vpull-up tying the first with the latter and turning on Bus Pirates' inner pull-up resistors.
    That should be more easy and cheap than add two external 3k9ohm resistors.
    Sorry but I don't understand.
    Regards,
    Piratinio

    • Piratinio says:

      Hi there,
      me answer myself.
      It’s because the article is about an old, very old version, of the Bus Pirate which it was much different compared to the current today’s production.
      Regards,
      Piratinio

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Follow

Get every new post delivered to your Inbox.

Join 94,423 other followers