How can I filter and convert a differential current output from an amplifier?

[Jdo300]

Hello Guys,

I really need some help with this one. I'm still working on the circuit design for this AD9959 programmable pulse generator and I found (from reading the datasheet) that the output from the DAC uses differential current output and not voltage output (something like 0.5V @ 20 mA max).

I've never played around with this type of output before but I ultimately want to filter the sine wave (which is DDS so it has squareness to it), and then use it to trigger a Schmitt trigger buffer IC to convert the sinusoidal output to a square wave. From there, the square wave signal will trigger a programmable pulse chip (3D7622) to produce programmable output pulses.

While looking around for a suitable Schmitt Trigger IC, I came across the SN74LVC1G17. In the datasheet for the input threshold, it mentions it in turns of threshold current rather than threshold voltage. I'm not sure exactly what that means or if it will work for my application. If not, would anyone have any suggestions for how to deal with this? I will be running the DDS chip in a frequency range of 0.1Hz to 20MHz.

Thanks,
Jason O

 

[berkeman]

I've never used the SN74LVC1G17, so I'm no help on that one. Given your brief description (and 20MHz requirement), I'd be inclined to terminate the differential current output in a resistor (maybe 100 Ohms), and pin one side of the resistor to a 2.5V supply. Then use a high-speed comparator with 20mV of external positive feedback hysteresis about 2.5V to convert the DAC output signal to the digital signal that you want. Just an idea...

 

[Jdo300]

Hi Berkeman,

Thanks for the suggestion. I was also thinking about this too. I looked up high speed comparators and came across this one here called the LT1721IGN. I'm still reading the datasheet but just thought i would run it past you for comments. I'm not sure what features to look for other than the speed and adjust ability of the threshold. Also, could you sketch a circuit for the setup you described? I think I know what you are getting at but I want to be sure.

Thanks,
Jason O

 

[Jdo300]

Hello again,

Just a quick question. Is there a type of comparator out there that is specifically made to convert a sine wave input to a square wave output (like sensing when the sine wave goes through 0V? Maybe that way I wouldn't have to put a bias voltage on it?? I don't know much about comparators.

Thanks,
Jason O

 

[berkeman]

That comparator looks like a good candidate. Be sure to read over any application notes for it, and follow any recommended PCB layout guidelines carefully. When you are using fast comparators, it is difficult to keep them from buzzing at transitions, especially for slowly-varying input signals. Positive feedback hysteresis is fundamental, and you often need to add a speedup capacitor around the feedback resistor to make it work.

As for a zero-crossing detector circuit, they are common in 50/60Hz applications, like for lighting dimmer circuits with triacs, but those are low frequency. You could make your comparator inputs 0V referenced, but that would complicate your hysteresis feedback scheme. It would appear to be best to stick with the 0V to 5V rails, and bias your current DAC output with respect to 2.5V like I mentioned. That will give you the best symmetry on your output 0V to 5V square wave from the comparator.

Also, could you sketch a circuit for the setup you described? I think I know what you are getting at but I want to be sure.

Nope. That's an exercise for the reader...


EDIT -- Oh, and if you end up using a dual or quad comparator package, be sure to tie off the unused comparators in the package to some benign state.

 

[Jdo300]

Hi Berkeman,

Thank you for the suggestions for the circuit. Taking your hint, I went ahead and drew up what I *think* will work. I also attached the page of the datasheet which talks about the outputs from the AD9959. It seems like they are saying that the outputs oscillate 0.5V above and below the power supply voltage (1.8V). If I am reading this correctly, then all I need to do is bias the comparator at 1.8V to have it trigger correctly. Could you look at the attached sheet and let me know if I'm doing this right? I'm wondering if the output behaves like this in all cases or if this is true only when you use an output transformer like what they show in the diagram.

Thank you,
Jason O

 

[berkeman]

First off, I'm confused by the DAC output biasing that is partially described on the datasheet page that you attached. Do you know why the output needs to be terminated into AVdd? What rails does it normally run between? (I guess I could pull down the full datasheet from Analog Devices, but I'm short on time this morning.)

And on the comparator circuit, you will need positive feedback hysteresis and a speedup cap around the feedback resistor, I would think.

 

[Jdo300]

Hi Berkeman,

I too was confused about the explanation about te DAC on that page of the datasheet. I finally called Analog Devices' tech support people to ask them what is going on with it. They explained to me that both of the outputs of the DAC are supposed to terminate to AVDD through load resistors if you don't use a transformer like they show in the datasheet.

I then told them what I want to do and the comparator that I plan on using, and I drew a circuit from their explanation. I attached a couple of diagrams for you to check out. The first one is the one I made from the tech's description of how it should be connected. I'm not quite sure how to implement the hysteresis feedback on this but I also drew a second diagram of how I *think* it goes. please let me know what you think.

Thanks,
Jason O

 

[berkeman]

Yeah, your 2nd diagram looks like a good first cut. I'm still not clear on the AVdd termination thing out of the AD part -- have you run across a detailed internal diagram of the DAC that might explain what they are using it for? I guess they are making the output swing symmetrically differentially about AVdd?

 

[Jdo300]

Hi Berkeman,

Yes, I think that they are using AVDD as the reference voltage for the DAC to swing positive/negative on. As for an internal circuit diagram, I have yet to come across one. I don't remember the datasheet going into any details about the internal workings of the DACs either but I could be wrong. Now my next question is how do I determine what value to use for the resistor and capacitor on the comparator for the hysteresis loop? Also, what exactly does the speedup cap do?

Thanks,
Jason O

 

[berkeman]

As a first cut, set the feedback resistance to about 10x the input resistance, and calculate what hysteresis that gives you, and what it does to the trip points. You may want to mess around some with the values and the AVdd bias point, if that helps you achieve your symmetry goals.

For the speedup cap, it is used to null out the slowdown effect of the input capacitance of the + input of the comparator. Its value is related to that input capacitance, and the ratio of the feedback and input resistors. You need to work through the math to see which way you should ratio the speedup capacitor...

BTW, I was looking for an explanation of the compensation capacitor in oscilloscope probes (it is used for almost the same thing that you are using the feedback speedup cap for here), and I came across a couple good resources from Tektronix. I thought I'd post them here for folks who might be interested:


Tek Probes Primer: http://physics.wustl.edu/ClassInfo/321/Pdf/abcProbes.pdf

Tek Oscilloscope Primer page: http://www.tek.com/Measurement/cgi-...rement/App_Notes/XYZs/&FrameSet=oscilloscopes


And here is a page that discusses the role of the compensation capacitor in oscilloscope probes... Scope probes vary the input capacitor and have a fixed feedback capacitor, which is kind of the opposite of what you have in your circuit (but it's the same idea about ratio-ing the capacitors versus the two resistances):

http://www.eng.yale.edu/ee-labs/morse/probe/probe2.htm

 

[Jdo300]

Hi Berkeman,

Thanks again for the information and all the great links. I haven't yet had the time to pour through them but I will as time permits.

One more question. I am trying to determine how to calculate the value for the feedback resistor but I have having some trouble finding any reference material on it. Most of the equations and calculations I found all assume that one of the pins has the voltage bias on it while th other one has the signal. How do I calculate it in the particular setup that I have since both pins are biased and have the differential signal on them?? Do you know of any good resources I could check out to figure this out?

Thank you,
Jason O

 

[berkeman]

The positive feedback resistor will add a voltage offset to the + input of the comparator when the output is high at 5V, and it will subtract an offset voltage at the - input when the output is low at 0V. The amount of this offset is determined by the ratio of the input (AVdd tie) and feedback resistors. More hysteresis is better for preventing output buzzing, but too much hysteresis gives you assymmetries and distortions of the "zero crossing" waveform information. Does that help?

 

[Jdo300]

Perfect, thanks .