triangle

Here's my maybe-final triangle wave generator.

formatting link

This will be used for production test of a 12-bit digitizer. We'll run the triangle into the ADC take a bunch of random-time samples and analyze them, looking for stuck bits and bad bit statistics. I haven't decided if it would be best to rail the ADC a bit at both ends, or stop just short. May as well rail, and check for 0x000 and 0xFFF, just to have one more thing to test.

It has a lot of parts, but it's parts that we have, and it's just a test set. I hope I got all the polarities right.

Bad photo. Pencil on d-size vellum doesn't photograph well.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin
Loading thread data ...

Huh.. (I'm not saying it doesn't work) But I'm missing something. Why doesn't that first opamp slam towards the rail (at it's slew rate) when the filp flop changes state. (oh, I see charging the cap..)

Still it's not at all obvious to my how that works. I don't see what's controlling that first opamp.. (Is it OK that I often play "the fool" here on SED.)

George H.

Reply to
George Herold

It looks like a clean design.

Reply to
bloggs.fredbloggs.fred

The decadent method would be to whip out the credit card and buy something like this:

formatting link

Also comes much smaller sans display.

The photo looks alright. Sometimes it helps to load it into IrfanView, transfer to monochrome and click "Sharpen" 2-3 times. That's how I make web-derived street maps readable via a B/W laser printer.

--
Regards, Joerg 

http://www.analogconsultants.com/
Reply to
Joerg

It's an integrator. When it hits the upper or lower limit, we reverse the input current and ramp it in the opposite direction.

The current into the integrator is + or - 270 uA, depending on the state of the flipflop. The exact frequency, and the ramp slopes, aren't critical here: I just want to scan all the ADC codes.

I should add another 1K resistor, from the opamp into the upper comparator, so the amp can't ever stuff a zillion mA into the comparator ESD diodes. If I use the nice little open-drain MCP6567, I'll need pullups, too.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

Except for a couple of trivial blunders, like protecting the upper comparator and adding pullups to the comparator open drains.

We have Preliminary and Critical design reviews to catch silly mistakes. SED is review zero.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

Oh snap. Yeah.. you know that cap was confusing me. I didn't see it was in the feedback loop... I want my integrating caps drawn over the top of the opamps.

George H.

Reply to
George Herold

Yeah, the old all-in-one triangle generators (ICL8038, XR206) seem to have left the building. Looks like a winner, though; only improvement would be if you made the thresholds less dependent on power rails, with a V reference.

I presume the 1k resistor at lower comparator input is for power-supply sequencing? And, that the little capacitor at the upper comparator is for ?

Reply to
whit3rd

You can still find function generators on Ebay. Well at least the kits to make a ICL8838 one. See ebay item 321937147204 $3.99 including shipping.

Dan

Reply to
dcaster

"Should be noted that the amplitude and DC offset can not be programmed"

And I want the triangle slopes to be smooth, unquantized, to test the statistics of my 12-bit ADC.

And besides, I'm a circuit designer.

I did that!

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

So that opamp can never force a ton of current into the comparator front-end. I need to do that for both comparators. AD8033 looks like a good opamp, rail-to-rail output. But it can source/sink 40 mA. I'm a little gun-shy about jamming current into ESD diodes.

To remove any supply noise from the comparator threshold; probably not needed. This is a statistical test, so a little noise wouldn't matter.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

AFAIK it's a potmeter just like in your circuit.

You be hard-pressed to find the quantization steps in a DDS output with a 12-bit converter.

Yes, same here, and that justifies rolling our own circuits. It's the manly thing to do.

The topper was a guy who sat for an exam at my university, one of the few where electronic calculators were allowed as long as they were not programmable. He came with a really dilapidated model, held together with tape, a bunch of wires, a loosely connected bridge rectifier with an electrolytic dangling off two solder joints, and a doorbell transformer with a toaster mains cable. Plus a substantial extension cord. "Got no batteries in that?" ... "They went flat a long time ago".

I think readability is quite ok. Much better than the prescriptions from my previous dentist.

--
Regards, Joerg 

http://www.analogconsultants.com/
Reply to
Joerg

See "Function_Generator_OmniComp-GenRad.pdf" (September, 1986) on the S.E.D/Schematics Page of my website... instrumentation grade... very low distortion Square, Sine and Triangle. ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
 Click to see the full signature
Reply to
Jim Thompson

The equally nice little MCP6562 comparators have push-pull outputs :)

piglet

Reply to
piglet

See "MicrochipComparators.zip" on the Device Models & Subcircuits Page of my website... I've re-written the Spice models so they don't choke on LTspice. ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
 Click to see the full signature
Reply to
Jim Thompson

We don't stock it, so it's easier to just add pullups.

They are very nice comparators, and only 60 cents. The data sheet specifically says that ESD diode current is tolerable, unlike the 339 types, where it's intolerable.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

Depends on the number of bits in the generator's DAC. And its lowpass filter behavior. At low frequencies, below the filter rolloff, a DDS will have output steps.

We weren't allowed to use calculators in tests. They gave an unfair advantage to the few who could afford them. My HP35 cost $400, the same as my first motorcycle.

--
John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  
 Click to see the full signature
Reply to
John Larkin

Den torsdag den 28. april 2016 kl. 23.56.31 UTC+2 skrev John Larkin:

when I was at uni we could any calculator, the test were usually about how you arrive at the result not the actual result, if you don't know what equations to use a fancy calculator doesn't help

-Lasse

Reply to
Lasse Langwadt Christensen

True, but miniscule. If I set my SFG-2120 to 7kHz 3Vpp triangle the steps are about 5mV and 100nsec. An audio range ADC wouldn't see those. They can easily be scrubbed if needed.

That's what I thought as well. Although, if one was serious enough about becoming an engineer in the early 80's one would invest in that tool. It was way more valuable than three text books costing the same.

I can't understand some youngsters these days. Many times a young guy told me "No, I don't have an oscilloscope". A PC scope can be had for under $100. But in front of the apartment was a shiny lowered sports car with mag wheels and a massive stereo in there. I had an oscilloscope before I turned 15. But for a long time no car or any other such luxury.

[...]
--
Regards, Joerg 

http://www.analogconsultants.com/
Reply to
Joerg

My HP45 was $400, too, but they were acceptable in all but a couple of the engineering departments and they were soon overruled by the dean's office. At the time, calculators weren't an absolute necessity, since most exams were set up to have nice numbers (to make it easier for slip-stick users)[*]. Engineering isn't in the touchy-feely humanities. Suck it up. College is expensive.

[*] In my junior-senior summer, our linear circuits class (and lab) had six people with calculators. Two with HP35s and four with SR50s. The two owners of the HPs got As, the other four failed the course and out of school. ...just proves that HPs were better calculators. ;-)
Reply to
krw

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.