Logic Inverter "noise"

[staze]

All,

I'm working on diagnosing a troubled Multimeter (here: https://www.eevblog.com/forum/chat/bench-multimeter-(another)/) and one thing that's come up is the hex inverter seems to be generating runt/crap pulses now and again (see image). The input to the inverter is "perfect".

So, I'm curious since I can't find anything anywhere... does it make sense that an inverter would only randomly produce crap (rather than always)? What would cause a failure like this?

Thanks!

[c4757p]

Check the rails powering the inverter. Could be power supply noise due to a faulty cap. Failing that, probe both input and output at the same time; see what the input does precisely during generation of these crap pulses. If all the inputs are fine (including power) and the output is not, the only one to blame is the inverter itself; replace it.

[SeanB]

If it is in a socket it could have an intermittent ground pin, or the bond wire internally is broken and barely making contact.

[PA0PBZ]

Is this the output of U49E? Did you replace the chip?

[staze]

Wow, stupid non-standard notify.

I haven't done both input and output at the same time (lack hands), but the input never wavers a bit. The output is the troublesome part. I'm just confused why only occasionally it would be bad.

No, it's not in a socket, but I've ordered a replacement chip (NTE74HCT04) and plan to install it in a socket (since there's room). So no, haven't replaced the chip yet. And yeah, it's the output from U49E.

The rails look perfect. No noise at all.

I guess my only concern was some loading on the output that was causing the noise, but it doesn't sound like that's likely. So I'll just wait for the replacement to arrive and get it swapped. Either way, my understanding is TTL level signals really need to be within about 5% in order to work, so those noisy/runty ones probably aren't doing their job?

[AndyC_772]

That output looks like it's struggling to pull down to GND. Check the ground connection to the IC, and if that's OK, change the device. I'd speculate that the lower (N channel) output FET in the inverter has died.

[staze]

That output looks like it's struggling to pull down to GND. Check the ground connection to the IC, and if that's OK, change the device. I'd speculate that the lower (N channel) output FET in the inverter has died.

Okay, that makes sense... except, why would it produce SOME good TTL pulses, and some bad ones?

I'll check the ground (resolder), and see what I get. But it sounds like it's an issue with the IC, and not something further down the line. So I'll replace if it's still showing as bad.

Any idea what would cause this? Just, random, or static discharge, or...?

[AndyC_772]

Speculation again, of course... maybe an internal connection has failed, but a transistor gate is still getting a weak signal by capacitive coupling. It looks as though each pulse is slightly stronger than the one before it, which suggests that something is happening which depends on time. It's the sort of behaviour you see when a high impedance node is floating, and picks up signals by capacitive coupling from nearby.

[staze]

Speculation again, of course... maybe an internal connection has failed, but a transistor gate is still getting a weak signal by capacitive coupling. It looks as though each pulse is slightly stronger than the one before it, which suggests that something is happening which depends on time. It's the sort of behaviour you see when a high impedance node is floating, and picks up signals by capacitive coupling from nearby.

Oh, that makes perfect sense. okay. Anyway, I'll post back once I've replaced it (once the part gets here). I just don't have a pile of logic chips around (sadly). I'm guessing, and maybe I'm wrong, this modern version will be a bit more robust as well... I know it's got protection from ESD, where I bet this original one didn't... but I don't recognize the brand mark on there, so I have no idea.

[staze]

As another note, since I'm still waiting for the replacement part, the output of the inverter on another "channel" is extremely noisy (the "strobe" channel on the Keithley 199. The input is clean, but the output is crap. Rather than being a 5V "0" and a 0V "1" or whatever, that 5V "0" is just high speed oscillation. So yeah, clearly something is wrong.

I'm going to unsolder and put in a 14pin socket while I wait for the part.

[PedroDaGr8]

This same IC failed on my Keithley 199. Except unlike his it failed with a constant 5V output. Since I wasn't sure if the 5V came from this IC or was injected later, I removed the IC, probed the same line and it's back fluctuating at around 1mV (which I am assuming is just noise I have picked up). The input is nice and clean with no real obvious issues.

[staze]

This same IC failed on my Keithley 199. Except unlike his it failed with a constant 5V output. I removed the IC, probed the same line and it's back fluctuating at around 1mV (which I am assuming is just noise I have picked up).

So, when you tested out of circuit, you provide a 5V rail, ground the other (well, negative rail on the power supply), and then feed it some TTL level pulses from an FG?

Thanks!

[PedroDaGr8]

This same IC failed on my Keithley 199. Except unlike his it failed with a constant 5V output. I removed the IC, probed the same line and it's back fluctuating at around 1mV (which I am assuming is just noise I have picked up).

So, when you tested out of circuit, you provide a 5V rail, ground the other (well, negative rail on the power supply), and then feed it some TTL level pulses from an FG?

Thanks!

No I didn't go that far into it. I just checked to see the pin to see if 5V was still present (which it would be if the 5V was injected from farther up). I actually planned on testing that tonight; I found an exact match of this IC on a scrap board at work, out of an old fraction collector.  I planned on testing the two side by side. Though in hind sight I should have tested it earlier because it is easy to do.

[staze]

ah, okay then. =)

I'm actually ordering some basic logic chips to have around (74HC stuff) for cases like this... waiting for stuff from China/Thailand, or in this case, New Jersey, is a bit annoying. The Aussies on this forum are spoiled with having a Jaycar/Farnell. Best we have is Frys, if you're close enough to one (closest for me is about 100 miles). =/ I miss the good ol' days of Radioshack.

[PedroDaGr8]

ah, okay then. =)

I'm actually ordering some basic logic chips to have around (74HC stuff) for cases like this... waiting for stuff from China/Thailand, or in this case, New Jersey, is a bit annoying. The Aussies on this forum are spoiled with having a Jaycar/Farnell. Best we have is Frys, if you're close enough to one (closest for me is about 100 miles). =/ I miss the good ol' days of Radioshack.

I'm about 20min away from a Fry's. Oddly enough I didn't even think of them for getting the part. Despite the fact when Im there I often pick up an IC or two just on principle.

[staze]

Do they have much in the way of Logic IC's? If I was heading to Portland anytime soon, I would have just done that.

[PedroDaGr8]

Do they have much in the way of Logic IC's? If I was heading to Portland anytime soon, I would have just done that.

I just checked and the have the NTE74HCT04 and it is in stock near me for $1.

http://www.frys.com/product/1003903

[staze]

Do they have much in the way of Logic IC's? If I was heading to Portland anytime soon, I would have just done that.

I just checked and the have the NTE74HCT04 and it is in stock near me for $1.

http://www.frys.com/product/1003903

boo....

[MacAttak]

Frys out here hasn't been so great. Almost all of their stock is wildly overpriced (if they have what you want at all) and ancient. And frequently with crushed pins and/or suspicious packaging.

Easier and cheaper to just buy from digikey.

[staze]

Frys out here hasn't been so great. Almost all of their stock is wildly overpriced (if they have what you want at all) and ancient. And frequently with crushed pins and/or suspicious packaging.

Easier and cheaper to just buy from digikey.

Yeah, I actually just ordered from RS over eBay. Free shipping, and reasonable prices. they do free FedEx Ground, which kind of sucks since it's slow. But, it is what it is.

There is a small shop here in town, but I'm guessing they'd want $5 for a 74HTC04. =/ I'm willing to pay double for something small like that locally... but sometimes it's just crazy.

[staze]

So, being a glutton for learning (or at least, confirming), I pulled U49 (the hex inverter), and breadboarded it with a +5V rail, and ground (from my bench PSU), then fed it a 5Khz 5Vpp square wave to the 4 inputs I could see were used on it (guess I could check the two unused ones, but why?). For interest, I set my FG to HiZ "destination" (50R source), put a Tee on the output, and put one output directly into channel 2 on the scope (blue trace), and the other half of the T goes to some minigrabbers that I hooked to the input/gnd.

Something is DEFINITELY wrong. Two of the inverters (1A and 5A) give basically no output (although it still looks capacitive), one (2A) gives something that definitely looks capacitive, and the other output (3A)... god, hell if I know what it's doing. I do know on a couple of the inverters, when I'd feed it a signal, the IC would draw about 30mA of power. On the other 2, it would draw about 10mA. I do know the 3A output was basically what I was seeing in circuit (drives the strobe channel).

Also of interest, all of them seem to either offset the input voltage, drag it down, or completely destroy it. This could be my fault for not understanding something required... maybe the FG can't drive enough... but I would think the inverter inputs would take VERY little current to drive.

Definitely failed.

[staze]

I just realized I did something dumb. I was feeding in a square wave... with a 0V offset... these are CMOS/TTL chips... so the signal shouldn't go negative. doh.

either way, chip definitely seems borked. will retest tomorrow with a proper signal. gah.

[ElectroIrradiator]

(Snip.)
... maybe the FG can't drive enough... but I would think the inverter inputs would take VERY little current to drive.

Yep. 74HC CMOS inputs are very high impedance ... as long as the input voltage stays between the power rails.

[staze]

(Snip.)
... maybe the FG can't drive enough... but I would think the inverter inputs would take VERY little current to drive.

Yep. 74HC CMOS inputs are very high impedance ... as long as the input voltage stays between the power rails.

.......*sigh*........

[staze]

not being able to sleep, I fired the scope, fg, and psu back up, and tested again with a +pulse this time... and largely similar results. Could be I toasted it with my square wave testing previously... ah well. Also note from looking online, CMOS logic chips should have all their inputs and outputs connected, even if not in use... so maybe that's part of the issue too.

Either way, looks like the chip is still bad. Wish the replacement would hurry up and get here. I found RS has 74LS series, but I understand you generally don't want to downgrade your logic. Maybe I'll get impatient and go shopping locally tomorrow. =P