AD9059, Many questions

Discussion in 'General Electronics Chat' started by CVMichael, May 13, 2010.

  1. CVMichael

    Thread Starter Senior Member

    Aug 3, 2007
    416
    17
    Continuing to my question here: High Speed ADC ?

    I purchased a few of the AD9059 (it was $30 each :eek: ), but I must say: this chip is AWESOME !!!!!
    I connected it to my Intronix LogicPort Analyzer, and I basically I turned my digital analyzer into an analog oscilloscope also.

    So, the questions:
    1) I used 2 * LM7805 power regulators, one for VD power, and another for VDD power. The thing that I'm worried about is that the AD9059 is getting a little worm to the touch, and also the LM7805 power regulator that goes to the VD input is a little more worm than the ADC chip, but not too hot so you can't touch it.
    I was not expecting that one chip that small (especially an ADC) to draw so much power to get a LM7805 power regulator worm at all.
    The data sheet says "Low Power: 400 mW Typical", is that enough to make the LM7805 worm ?

    2) I tried to change the VREF to 5V, so I connected the VREF (Pin 2) to VD (5V) (I also tried VDD which is 5V also), and the thing went DEAD ! as soon as I disconnected the VREF from the VD, then it worked fine after that...
    So my question is, how can I change the VREF then ?

    3) Is there a chip that can output a preset number of frequencies ?
    What I mean is, I want to give 60Mhz, and to give out, 50MHz, 40MHz, 30MHz, and so on ?
    I want to be able to change the frequency that goes in the ADC.

    4) Is there an Opto-isolator that can go at 60MHz (or higher) so that I can isolate the ADC output ?

    5) Why do they use the 74ACQ574 chip in the Evaluation Board Schematic ? does that act as an Opto-isolator ? (It does not look like it...)
     
    Last edited: May 13, 2010
  2. CVMichael

    Thread Starter Senior Member

    Aug 3, 2007
    416
    17
    6) I measured a 17KHz signal (from a stepper motor driver L297, connected to pin 16 OSC)

    Why is there so much noise ? (see picture attached)
     
  3. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Yes. Didn't you take that into account when designing your circuit board? The traces should be wide to help dissipate the heat. Orienting the PCB vertically will help to take advantage of the air convection currents.

    It's supposed to be 2.5v. The VREF pin is provided so that you can connect an 0.1uF cap between it and ground. You can also use a more accurate external 2.5v reference if you so desire. But don't go up to 5v with it.

    If you want to change your scaling, use a zero-drift very fast opamp on the input.

    Well, you might find multi-frequency TCXO's or OCXO's that have built-in frequency dividers. Or, you can buy crystal oscillators in a variety of frequencies, and wire up the dividers yourself.

    Have a look at Roman Black's page; he has plans to build a really cheap and simple OCXO.

    That's mighty zippy.
    If you can find one that will go to 5MHz, let us know?
    No, but it acts as a data latch.
     
  4. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    You have a noise problem somewhere.

    Are your data lines all the same length?

    Are you using the eval board, or a board of your own design?
     
  5. CVMichael

    Thread Starter Senior Member

    Aug 3, 2007
    416
    17
  6. someonesdad

    Senior Member

    Jul 7, 2009
    1,585
    141
    Michael, a few years ago I was thinking about getting a logic analyzer and I looked at 10 or 12 different USB ones. By reading web pages, the Intronix looked to me like the best analyzer to buy.

    I'm not intending to hijack your thread, but if you'd sometime take the time to write up your opinions of it and what you like and don't like, I'd sure appreciate it...
     
  7. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Very interesting - thanks. I didn't know they had them quite that zippy.

    Fiber optic transceivers can have extremely high data rates, but frankly I'm not familiar with what might be available.

    7805 regulators are very handy for "quick and dirty" linear regulator projects, but when you look at the tolerances in the datasheet, they really have more of a tolerance range than you'd want for something that you are using as an instrument.

    Errors (like enemies) accumulate.

    What is the source voltage that you are feeding the regulators from?

    Keep in mind that if you are using 12v for an input to a 400mW load, power dissipation in the regulator will be (12v-5v)*(400mW/5)=560mW. Thermal coupling to ambient for the 7805 (no heat sink) is 65°C per Watt, so your temp rise will be roughly 37°C over ambient.

    If your room temp is 25°C/77°F, the reg temp will be 25+37=62°C or 143°F; what most people set their hot water heater to. But, you're powering other stuff besides the ADC, so yours will be even hotter.

    Are you feeding both the analog and digital side from the same supply? If so, keep in mind that the grounds appear to be common to both sides. Your wiring from the regulators to the board need to be very short, as they have inductance that accumulates at a rate of ~15uH per 10mm @ 10MHz.

    More on your Vref: looking at the schematic of your eval board, Vref also serves as the noninverting input for U3 and U4; if you raise Vref to 5v, the outputs of the opamps will be against the rails. If you remove the jumpers from E10 and E13 (they ARE jumpers, aren't they?) the opamps will be removed from the loop, and you will be reading the unbuffered input.

    Are you using O-scope probes for your signal input? They will help to reduce the noise, as O-scope probes are shielded. They also have selectable 1:1 and 10:1 (some have 100:1) which lightens the load on the signal source.
     
  8. CVMichael

    Thread Starter Senior Member

    Aug 3, 2007
    416
    17
    I love this thing !! I've been able to debug my applications with it very well. You just have to learn how to set up triggers, and your good from there on.
    I also love that it can convert the digital input to analog, as you can see in the attached picture in the second post in this thread.

    The negative things:
    - I wish it also had an analog input (that's what I'm trying to do now, to add analog)
    - I wish it had more memory, to store more samples in one shot...

    But for this price, compared to the competitors, this one is much better.
    And I also love that it has 34 digital inputs, so even after I add this AD9059, having 2 * 8 bit analog inputs, I will still have 18 digital inputs left ! which is still more than what I need on daily basis...
    And if I don't need to use the analog, then I can simply unplug it, and use all the 34 digital channels.
     
  9. CVMichael

    Thread Starter Senior Member

    Aug 3, 2007
    416
    17
    Hi SgtWookie,

    I want to be able to change the frequency that goes into the AD9059, so I made a schematic using DipTrace (first time to do something like this).

    Can you please take a look at the schematic ?

    I want to be able to change the frequency using a microcontroller, and to also show on a display the frequency selected, and probably 2 buttons for "Up/Down" to increment & decrement the frequency. (the programming for the microcontroller is the easiest part for me)

    I did not focus on the AD part, just the part that generates the frequency for the "ENCODE" input.

    Is there a way to simplify the schematic ?
    For example is there an 74LS154 that has the outputs reversed ?
    Or are there any chips that have more OR inputs ? because right now I am using 5 * OR4 gates to collect all the frequencies.

    I attached the PDF, and also the DCH file zipped.
     
  10. CVMichael

    Thread Starter Senior Member

    Aug 3, 2007
    416
    17
    OK, maybe I was wrong to direct my previous post only to SgtWookie...

    Can ANYONE please take a look at the attached schematic (in the previous post), and let me know if it looks OK, and any improvements ?

    Thank You :)
     
Loading...