Crutschow said:
..............
I'm concerned that in the future it will be difficult to find discrete parts that the hobbyist can even readily use (they'll all be tiny surface mount devices).

I think that our addiction to cheap, throwaway things is going to contribute to that. Nobody wants anything repaired anymore. They want to throw away the old broken one and buy a shiny new one. And that means continuing sales so the manufacturers are happy with the concept. It also means that there's little reason to produce and stock repair parts. If nobody repairs their things anymore, the manufacturers don't have to design the things to be repairable (by the consumer or the manufacturer).

Harvey

 

Crutschow said:
..............
EVERYTHING for the electronic hobbyist was better 40 years ago.

Radio Shack was cool when the stores started popping up in the 1960s! If you were a budding mad scientist, Radio Shack was where you could be found after school. No matter what you needed for a project, Radio Shack would have it somewhere on a huge wall of pegboard filled with transistors, resistors, capacitors, and more!

But then, Radio Shack figured out that there was a greater profit margin in selling consumer electronics. So the huge wall of components shrunk a little to make space for clock radios, flashlights, electronic games, and the like.

Then Radio Shack started producing and selling its own line of stereo equipment and the display of little bubble packs of components became an aisle. Only us mad scientists mourned when we discovered that the tube tester had been hauled off.

Then Radio Shack started specializing in cheap electronic children's toys, especially around Christmas (April to January). To make room, all of those magical electronic components that drew in Radio Shack's first customers were relegated to a cabinet of drawers in the back of the store near the fire exit. After that came computers (really awful ones) then pagers, then Radio Shack was restructured into a cell phone store and that was the beginning of the end for Good 'Ol RS. And all of Radio Shack's managers and stockholders wondered when it all started going downhill.

Maybe it was when the bicycle racks in front of the stores were no longer filled after school.

Harvey

 

Well, I assembled the circuit on the new breadboard again today. It didn't work again. Not having a scope is really inconvenient! What are the "node" voltages mentioned in post #90 that you wanted me to check?

Harvey

 

Well, I assembled the circuit on the new breadboard again today. It didn't work again. Not having a scope is really inconvenient! What are the "node" voltages mentioned in post #90 that you wanted me to check?

All the nodes carrying the signal, all inputs and outputs of the op amps.
If you put them on the schematic that would be extra helpful.

 

Please see the attached image.

Note that all voltage readings were taken with an analog meter and in reference to the circuit ground (15VDC power supply -) and not the Vgnd. All readings were positive. Also note that all readings were steady (i.e. no modulation detected).

Thanks for your help!

Harvey

Mod edit: converted .doc to .pdf for you

 

As no node (apart from the supply output) is above 2.5V I'd guess pin 4 of the LM324 isn't properly connected to the 15V supply.

 

Yes, check the voltage at the LM324 pin 4 and check that pin 11 goes to ground.

You don't have a digital multimeter??
I suspect the analog meter is loading some of the high impedance nodes.

 

Yes, check the voltage at the LM324 pin 4 and check that pin 11 goes to ground.
Sorry. I had the readings on my notes but failed to put them on my attachment. Yes, pin 4 of the dip is getting +15v and pin 11 is connected to 0V on the power supply. (And I did ohm it out to make sure it was, in fact, a good connection all the way to the p/s.)

You don't have a digital multimeter??
Yes, I do have one but I elected to take the initial measurements with an analog meter to see if there was any hint that the circuit was trying to operate (modulate).

I suspect the analog meter is loading some of the high impedance nodes.
But the circuit doesn't work whether the analog meter is connected or disconnected.

What a bad time to be without a scope!

Harvey

 

The output voltages of U1 and U2 are way off what they should be, given the inputs. If the IC is plugged in, make doubly sure it is the right way round.

 

If the IC is plugged in, make doubly sure it is the right way round.

I've got the IC plugged into a 14-pin DIP socket oriented horizontally with the notch to the right. That makes pin #1 in the upper right corner, #2 is to its left, and so on counterclockwise around the socket and ending up with #14 below the notch from #1.

Correct?

Harvey

 

I've got the IC plugged into a 14-pin DIP socket oriented horizontally with the notch to the right. That makes pin #1 in the upper right corner, #2 is to its left, and so on counterclockwise around the socket and ending up with #14 below the notch from #1.

Are you talking top or bottom view?
Here's top view.
Are you measuring all the voltages directly on the device pins, not the socket pins?

 

If you rotate your drawing 180 deg, it'll match the orientation (and pin ID) of mine on the breadboard. (Yes, top view.) I've been taking my measurements on the DIP pins themselves.

Harvey

 

Then you appear to have a bad op amp because those voltage readings don't follow from a properly operating amp.

The two inputs of an op amp will be essentially at the same voltage (within a few mV) when it is operating in the linear region.

If not, then the output will be saturated either at the rail or ground depending upon whether the plus (+) input is higher or lower than the minus (-) input.

All of your measurements appear to violate those rules.

But do any further measurements with the digital meter, not the analog.

 

Today's goal is to strip down the breadboard and build it back up. If the problem persists, I have an extra LM324, TL431, and two 4138s. (I bought duplicates of the sensitive components in case I zapped anything while assembling the circuit.) I'll then take readings with a digital VOM and let you know how everything turns out.

Harvey

 

IT WORKS!!!

I got the circuit torn down today and then put it back on the breadboard. It didn't work at first but after moving some + and - power supply points around, my test bulb started showing some signs of life. After doing some twiddling on the three trimpots, I got it working! Apparently, I don't fully understand how the pins of breadboards are connected together under its plastic surface!

Coupla questions...
1. How should I set the trim pots?
2. The ramp up and down of the sine wave isn't smooth but has a slight "jerky" effect on my test bulb. It's a bit like the sine wave is more of a stepped output. (My test bulb is a 60w incandescent. I haven't tried the LED bulb yet.) Any idea what I could do to smooth it out? If nothing, I can easily live with what I've got.

Guys, I can't begin to tell you how much I appreciate your help through these past 115 posts!

Harvey

PS
I didn't have to change out the op amp after all.

 

Glad you got it to work.

You should always check a breadboard wiring with an ohmmeter after it's built, to verify the connections.
That can save a lot of grief.
It's seldom that I've built a circuit breadboard, (or had one built) that didn't have at least one wiring error.

Adjusting the trimpots will vary the output waveshape but it's kind of hit and miss to know what to adjust without an oscilloscope.
Sounds like the gain may be too high.
First try adjusting the output gain (pot U5).
Then just try varying the other two pots, one at a time, to see if you can get the light to behave as you want.

 

Once you've got it working, check your parts against that last PCB I posted -- if we made some modifications since I last posted that let me know (PM please, to hopefully get my attention, heh) and I will update the PCB to accommodate the latest revisions.

 

DJ,

It looks like your board in post #58 incorporates all of our changes and is the one to build.

How do I order the boards?

Harvey

Note: PM ("conversation") also sent.

 

The board can now be ordered @ https://oshpark.com/shared_projects/G1iJ6ymm ... I thought I'd set it as "shared" in OSHPark before, but apparently I was wrong. Anyways, it's shared now

Keep in mind it's designed with the Mouser parts list in mind; the most notable caveat would be the trimpots and capacitors, as they come in many flavors and footprints. The trimpots here are the top-adjust inline-pin variety (TTE designates these with a "W" in their part number). The capacitors here are MLCC, which theoretically should have a longer life and be less susceptible to temperature/humidity variations than electrolytics.

I've updated the parts list slightly, since the 10K resistors originally specified are currently shown as backordered. Thankfully 10K is very common so it was easy to find a substitution! http://www.mouser.com/ProjectManager/ProjectDetail.aspx?AccessID=7bc3b14352

 

Thanks DJ! I just ordered and paid for the three boards. I also saw your PM regarding your advice to build the boards one at a time. Since I already bought (from Mouser) enough parts to populate one circuit for breadboarding, I'll use those to build the first board. If all goes well, I'll then order the parts (2/3rds of your Mouser cart) for the other two boards.

I DO understand the possible fitment headaches from subbing parts for a PCB. Therefore, I have full intention of ordering the Mouser parts that you've specified.

THANKS FOR YOUR HELP IN ALL THIS!!!

Harvey