POR and 74LS input current

As someone else pointed out it's on the web as an 84MB pdf, the search feature on Agilents website could be better.

Assuming the resistor is still 100K that means 34uA is flowing.

A 74LS132 datasheet says the high-state input current is typically

20uA.

The low-state input current is negative which suggests to me that most of that 34uA is going through the capacitor. You might like to take that cap out and measure the leakage current with 4V on it to see how bad it is.

The capacitor has almost certainly become more leaky over time. The manual says it's a 15V tantalum.

I don't think you can complain about somthing that has worked for twenty years.

I doubt the input current of the 132 will drift much over time so I suspect the cap is very leaky. Suppose they had used a 47K resistor and a 47uF cap. A cap of twice the size in the same range will typically have around twice the leakage current so there might not have been any benefit. Even using 47K and keeping the 22uF cap might only have made it last another year.

Adam

Joerg, I think your solution is correct (smaller R, bigger C). I didn't check the data sheet, but if 20uA is correct for the high input current, and you have 2 inputs, then: .

40uA X 0.1Meg = 4V drop across the resistor, worst case. Regards, Jon

"Joerg" schreef in bericht news:jPV3g.17030$ snipped-for-privacy@newssvr11.news.prodigy.com...

Joerg,

I'd bet a beer on a leaky capacitor. The 74LS132 may be defective or the

100k resistor may be interrupted, but I consider that almost only theory. Unless really mistreated these components last much, much longer than an elco. A guy named Kirchhof ever said that current has to flow somewhere and a leak is the only thing I can think about :)

petrus bitbyter

Hello Adam,

I have replaced it with a new good quality cap. Same thing. With lower resistance and a good 100uF cap it works fine.

Oh, I don't. I was just surprised. The logo on the circuit board reads "YHP" so maybe this section was a Yokogawa design.

That would reduce the reset time to half. The best solution would be to build a real reset circuit, using a transistor or a reset chip. Although, with reset chips I had found some bad apples over the years so I tend to roll my own.

One item such instruments would have really needed would be a precision NiCd monitor. Then an error code would have popped up before the thing leaked all over the place. The nickels for a TL431 and a CD4060 wouldn't have made much of a dent in the profit margin on a >30k instrument. Oh well, other than that the 4191 has been good to me. What I really like is that it doesn't create the jet engine fan noise of equipment such as the 3577. Meaning less allergies as well.

Regards, Joerg

Hello Jon,

Yes, but as Adam pointed out it would flow into the cap. Nevertheless, about 3.5V across the resistor is what I am seeing. Before I leave a bigger cap in there I'd have to meaure how fast the 5V die out after turn-off and see whether the voltage divider clamp bleeds that cap fast enough so it won't fry the emitters on the 74LS132 inputs. If I could just swap it out for a 74HCT132 I'd be home but the via plating of those thick circuit board is not very good, too risky.

Regards, Joerg

Hello Petrus,

'tis what I also thought. New cap, no change. Measured the old cap, no leakage to write home about. Somewhere

"Joerg" a écrit dans le message de news:ti64g.64546$F snipped-for-privacy@newssvr29.news.prodigy.net...

theory.

and

Being LS, and from the date, this must be DIP package. Why not cutting the pins with a plier and remove the rest, one pin after another. No damage to the PCB.

If the IC has developped some pb that you cure with a lower value resistor, it might well worsen in the future and you'll be back to square one. If it were me I'd change it now the beast is opened. But I'm not you :-)

--
Thanks,
Fred.

Hello Fred,

Yes, I might just do that. It's all through-hole parts.

The reset must have always been marginal. I remember when I bought it that I wondered about the long time for the display to turn on. I always thought this was some kind of controlled heat-up phase since the manual says that the unit won't be ready before a 10 minute warming phase. The engineer at the company I bought it from said "It's always done that, even when new". It must have been considered normal since it had gone through many calibration labs. Guess now I got to the ground of it. The display comes on right after the reset is done.

There are also two transistors in Darlington configuration that additionally hold the reset until 5V is fully deployed. Just took out the transistors, looks like some kind of special order type from Motorola with a long number on them. Their beta appears to be a bit marginal so now they'll be replaced with something better.

Regards, Joerg

Joerg, You are correct in that initially, the current would flow into the cap. However, as the cap voltage gets into the "undefined" region between a 0 and a 1, the current direction would be into the gate. Regards, Jon

Hello Jon,

It's now fixed, reaches 2.5V which is well above the worst case threshold. Does a clean reset, passes all tests. However, the analyzer displays grossly wrong readings between calibration points. That's where the service manual isn't too helpful anymore because the code for the controller isn't in there.

Regards, Joerg