Author Topic: Hot 74LS04, replace or look elsewhere?  (Read 11588 times)

0 Members and 1 Guest are viewing this topic.

Offline StonentTopic starter

  • Super Contributor
  • ***
  • Posts: 3824
  • Country: us
Hot 74LS04, replace or look elsewhere?
« on: July 09, 2015, 08:09:35 pm »
I've got a 74LS04 on my IBM PCjr motherboard that I'm trying to fix that is getting too hot to touch.  I probed the voltages and got this:



Would you think that the LS04 is failing and getting hot, or would you think maybe there's a problem somewhere else (maybe a short?). Pins one and two look like they may be ok. Pins 12 and 13 concern me a little bit. 11 and 10 look almost shorted, as well as 5 and 6.  9 and 8  I'm not sure if the voltage is too low to invert.  And 3 and 4 I'm not sure either.

This is the only part on the board that is too hot to touch after about 30 seconds of power.  Using a 50K thermistor I could get it as high as 44C but I know it was higher than that because it was painful to hold the back of my finger to it.  If I wiped a damp finger across it, the water would evaporate in about 3 seconds.

There's an 74LS374 very close to it that's warmer than all the other LS chips on the board but not hot.

Unfortunately I can't tell which LS04 in the schematic this one is since the part label on the board layout sheet is too blurry to read, but it is very close to the 8088 processor.

Schematics: https://drive.google.com/folderview?id=0B3nYGMOtVmrVfkoyU0VNVnBqd0doVkUwWWhTQ0hlUS1zZkxRaXRUMDlxZzIwX3FEMFVDcDg&usp=sharing

The LS04 chip is the one directly next to the left end of the 8088 CPU if you have the board rotated with the cartridge slots in front.
« Last Edit: July 09, 2015, 08:30:32 pm by Stonent »
The larger the government, the smaller the citizen.
 

Offline grumpydoc

  • Super Contributor
  • ***
  • Posts: 2905
  • Country: gb
Re: Hot 74LS04, replace or look elsewhere?
« Reply #1 on: July 09, 2015, 09:01:25 pm »
Hard to say, but easy enough to replace and see.

It's hard to say because we don't know what signals are on the gates, you would be better off looking with a 'scope if you have one

TTL should drive a logic 1 output >= 2.8V and a logic 0 output should be <= 0.4V so, on the face of it a gate with 0.13V in and 1.7V out is struggling - if those are DC values.

BUT if it is a train of pulses swinging between 1 and 0 where the average value works out as 0.13V at the input then 1.7V out might be reasonable.

However we might guess that an input of 0.05V has to be low most of the time so the output should be high most of the time - making the figures for the 3rd gate a bit suspicious. Similarly the 6th gate (pins 12&13) looks a bit odd because for the average at the input to be 5v it is probably pulled up to VCC and even if there are some "0" pulses they are going to be short and infrequent - hence the output looks too high.

What test equipment do you have apart from the multimeter, and which multimeter do you have?
 

Offline rolycat

  • Super Contributor
  • ***
  • Posts: 1101
  • Country: gb
Re: Hot 74LS04, replace or look elsewhere?
« Reply #2 on: July 09, 2015, 09:44:39 pm »

What test equipment do you have apart from the multimeter, and which multimeter do you have?

Ooh! ooh!  I know this one:

Tenma 72-720 20Mhz Scope, Radio Shack 22-168A Multimeter

 

Offline grumpydoc

  • Super Contributor
  • ***
  • Posts: 2905
  • Country: gb
Re: Hot 74LS04, replace or look elsewhere?
« Reply #3 on: July 09, 2015, 10:08:43 pm »
Ooh! ooh!  I know this one:

Tenma 72-720 20Mhz Scope, Radio Shack 22-168A Multimeter
:)

OK, Stonent actually filled that in. Well, assuming it's accurate even a 20Mhz 'scope should be fine for a PCjr

This -  http://read.pudn.com/downloads156/sourcecode/asm/696308/IBM%20PCjr%20Technical%20Reference%20manual.pdf if you haven't already found it might be useful as it has the circuit diagrams (many copies kicking around do not contain Appendix B).



 

Offline StonentTopic starter

  • Super Contributor
  • ***
  • Posts: 3824
  • Country: us
Re: Hot 74LS04, replace or look elsewhere?
« Reply #4 on: July 09, 2015, 10:58:42 pm »
 :-DD

What test equipment do you have apart from the multimeter, and which multimeter do you have?

Ooh! ooh!  I know this one:

Tenma 72-720 20Mhz Scope, Radio Shack 22-168A Multimeter
The larger the government, the smaller the citizen.
 

Offline StonentTopic starter

  • Super Contributor
  • ***
  • Posts: 3824
  • Country: us
Re: Hot 74LS04, replace or look elsewhere?
« Reply #5 on: July 09, 2015, 11:06:26 pm »
I probed earlier today and didn't seem to be seeing any thing on the scope unless I turned it to the most sensitive setting which I think is .5mV/Div.  At that level and a sweep that was one click faster than a sweep that you could follow with your eyes, some of the pins had digital oscillations on them, but at levels of a few mV.   They looked a lot like what you'd see if you were probing I2C. Some of the pins seemed to have white noise on them in the few mV. range.

When I tried higher voltage levels, I just got solid voltage with no visible oscillation.  DC coupling just got me the voltages that I was expecting.  AC coupling was just  flatlined.

I started recording a video to show you guys but ran out of room on my phone and it stopped about 5 seconds before I finished.  I'll try to upload in a few minutes.
The larger the government, the smaller the citizen.
 

Offline flynwill

  • Regular Contributor
  • *
  • Posts: 143
Re: Hot 74LS04, replace or look elsewhere?
« Reply #6 on: July 09, 2015, 11:13:39 pm »
Indeed almost any 'scope is better than going at this with a multimeter....

As others have said it's difficult to tell for sure, but I would hazard a guess that whatever pin 12 is driving has died.  This is based on the multimeter indicating pin 13 at a solid high (nearly 5 volts) and pin 12 which should be low being at 2.5 V.  And I remember that a common failure mode in TTL was an input that sourced way too much current.

Digging through the schematic I only find one pin 12/13 section of a '04 being used:  ZM16 -- Sheet 7 C4 generates signal -ZA9 (inverted version of buffered address bus bit 9).   Following the schematic that appears to only drive ZM7 pin 2 (an LS32) Sheet 10 A8.  So I'd recommend doing some circuit tracing and if indeed you can find conductivity between pin 12 of your hot chip and pin 2 of an LS32 then I would start by replacing the LS32.

Best of luck!   8)
 

Offline StonentTopic starter

  • Super Contributor
  • ***
  • Posts: 3824
  • Country: us
Re: Hot 74LS04, replace or look elsewhere?
« Reply #7 on: July 09, 2015, 11:54:43 pm »
It looks like the one of the left says ZM44.  That's the one. ZM40 is the 8088 cpu.

I see 4 references to ZM44 on the page that has the CPU. 




« Last Edit: July 10, 2015, 12:50:10 am by Stonent »
The larger the government, the smaller the citizen.
 

Offline flynwill

  • Regular Contributor
  • *
  • Posts: 143
Re: Hot 74LS04, replace or look elsewhere?
« Reply #8 on: July 10, 2015, 12:37:16 am »
Well if it's ZM44 then ignore what I said in the last post.   ZM44 is indeed a '04, but I can't find the 12/13 element anywhere on the schematic.  I'd suggest a close look at the PCB to see if there are any traces to pins 12/13.

Do you have any experience at replacing a dip package on a PCB?
 

Offline StonentTopic starter

  • Super Contributor
  • ***
  • Posts: 3824
  • Country: us
Re: Hot 74LS04, replace or look elsewhere?
« Reply #9 on: July 10, 2015, 01:07:10 am »
Well if it's ZM44 then ignore what I said in the last post.   ZM44 is indeed a '04, but I can't find the 12/13 element anywhere on the schematic.  I'd suggest a close look at the PCB to see if there are any traces to pins 12/13.

Do you have any experience at replacing a dip package on a PCB?

Not a whole lot, just installing them in some kits I've built.  I'm a bit blobby with the solder unfortunately. 

I did get the video uploaded.  It ends suddenly because I filled up the memory. All the interesting stuff is when I turn down the voltage.
I say pin 8 at the beginning but mean pin 4.  I will probably go back and leave it at 1V/div and poke around again to be sure there weren't any higher voltage signals I was missing.

« Last Edit: July 10, 2015, 01:12:56 am by Stonent »
The larger the government, the smaller the citizen.
 

Offline Paul Moir

  • Frequent Contributor
  • **
  • Posts: 926
  • Country: ca
Re: Hot 74LS04, replace or look elsewhere?
« Reply #10 on: July 10, 2015, 02:43:37 am »
Whoa, you say pin 8 but it looks like you're probing 14?  Is there really that much noise on your 5v rail?

(Can't imagine any normal situation where a 74LS04 would be hot though.  Warm even running full speed.)

(Pin 6 doesn't make sense considering it's the output of pin 5.  Short to ground?)
« Last Edit: July 10, 2015, 02:51:21 am by Paul Moir »
 

Offline poot36

  • Frequent Contributor
  • **
  • Posts: 678
  • Country: ca
Re: Hot 74LS04, replace or look elsewhere?
« Reply #11 on: July 10, 2015, 03:02:38 am »
If you want to clean up that scope trace try connecting the ground clip to something closer to the problem chip as well as make the ground clip as short as possible to reduce picked up noise.  I am not sure using what looks like the composite video output jack as a ground is a good idea because it may not be directly referenced to system ground as well as having high frequency video signals potentially interfering with your scope measurement.  You should see a squarish looking signal but it should not have that much spikes on it.  I would check the noise on the 5V rail from the power supply by setting your scope to AC coupling and the scope voltage to 0.2V per division.  If you are getting more then 200mv of ripple then I think it would be a good idea to replace the power supply caps just to avoid sending bad power to the chips and potentially damaging more of them.
 

Offline StonentTopic starter

  • Super Contributor
  • ***
  • Posts: 3824
  • Country: us
Re: Hot 74LS04, replace or look elsewhere?
« Reply #12 on: July 10, 2015, 03:57:47 am »
If you want to clean up that scope trace try connecting the ground clip to something closer to the problem chip as well as make the ground clip as short as possible to reduce picked up noise.  I am not sure using what looks like the composite video output jack as a ground is a good idea because it may not be directly referenced to system ground as well as having high frequency video signals potentially interfering with your scope measurement.  You should see a squarish looking signal but it should not have that much spikes on it.  I would check the noise on the 5V rail from the power supply by setting your scope to AC coupling and the scope voltage to 0.2V per division.  If you are getting more then 200mv of ripple then I think it would be a good idea to replace the power supply caps just to avoid sending bad power to the chips and potentially damaging more of them.

Yeah I just couldn't find a good spot to clip to. Though the scope and motherboard's ground plane are both earth referenced so if I'm just checking DC logic high/low, I can leave the clip "flapping in the breeze".

I think I checked the power rail and it was pretty smooth even by modern power supply standards.  I can't do much with the schematic for the power board since mine appears to be a different revision that is a completely different design. Mine spans the whole board, but the one in the manual is very short.  The board layout in the book makes it look like there's a huge TO5 packaged device in the center with a big heatsink.  Mine has 2 SIPP devices side by side with a heatsink on each.

I will double check the power rail just to confirm.
The larger the government, the smaller the citizen.
 

Offline flynwill

  • Regular Contributor
  • *
  • Posts: 143
Re: Hot 74LS04, replace or look elsewhere?
« Reply #13 on: July 10, 2015, 03:59:36 am »
Not a whole lot, just installing them in some kits I've built.  I'm a bit blobby with the solder unfortunately. 

The best way, particularly if you have limited tooling, is to clip all of the part's legs with diagonal cutters, and remove the plastic body first.  Then you can grasp the legs with your needle nose pliers while heating the solder connection with your iron and pull them out one at a time.  Finally use solder wick or a solder sucker to clear the holes.

Watching your video, I agree with the others, take a close look at the power supply as well.  Your dodgy scope ground might explain some of the bad waveforms I saw, but not all of them. 
 

Offline grumpydoc

  • Super Contributor
  • ***
  • Posts: 2905
  • Country: gb
Re: Hot 74LS04, replace or look elsewhere?
« Reply #14 on: July 10, 2015, 06:36:16 am »
Well if it's ZM44 then ignore what I said in the last post.   ZM44 is indeed a '04, but I can't find the 12/13 element anywhere on the schematic.  I'd suggest a close look at the PCB to see if there are any traces to pins 12/13.
Unless anyone spots it on the schematic it might be an unused gate with the input tied to +5V (via a resistor, usually 1-10k).

Do you have any experience at replacing a dip package on a PCB?

Not a whole lot, just installing them in some kits I've built.  I'm a bit blobby with the solder unfortunately. 
Get a small pair of snips or flat face cutter - snip through the legs one by one and lift the body off the board, then carefully remove each leg with your soldering iron. Finish off the hole with a solder sucker. On boards this old, fresh flux or a bit of fresh solder applied to the joint will aid the removal process. If you have any scrap boards practice first - it's quite easy to lift tracks and damage through hole plating but if you are careful you can avoid this.

EDIT: Oops - I see flynwill already covered this; must do more than glance at posts :)

As clock frequencies in these boards are not high put the new chip in a decent quality turned pin socket.
« Last Edit: July 10, 2015, 07:34:10 am by grumpydoc »
 

Offline flynwill

  • Regular Contributor
  • *
  • Posts: 143
Re: Hot 74LS04, replace or look elsewhere?
« Reply #15 on: July 10, 2015, 01:19:21 pm »
Since you appear to have an analog 'scope I'll pass on another tip from the ancient days of microprocessor debugging...

Somehow hookup your signal generator set at 1-10 kHz to the reset line of the system.  (You will need to review the schematics to find the best place to do this.  TTL is pretty tolerant of "wire or'ing" another signal into an active-low line because the pullups are pretty weak).  Then trigger your 'scope off the rising edge of reset and you can methodically scope out exactly what the address and data lines are doing for the first few cycles.  Of course it helps is you know what the first few instructions in the ROM are, but even if you don't this will often help to find faults in the address decoding or address/data busses.
 

Offline grumpydoc

  • Super Contributor
  • ***
  • Posts: 2905
  • Country: gb
Re: Hot 74LS04, replace or look elsewhere?
« Reply #16 on: July 10, 2015, 01:33:28 pm »
Also LS TTL is pennies per chip - when you buy replacements get 5 or 10 and don't solder a new chip directly back into the board but put a socket in. That way even if you do fry the first replacement, or even the 2nd it's no big deal while you debug the rest of the board.

I'd get that 04 out before going much further though - if the gate between pins 12 & 13 really is unused then thet IC is probably fried anyway.

 

Offline StonentTopic starter

  • Super Contributor
  • ***
  • Posts: 3824
  • Country: us
Re: Hot 74LS04, replace or look elsewhere?
« Reply #17 on: July 10, 2015, 03:17:55 pm »
Also LS TTL is pennies per chip - when you buy replacements get 5 or 10 and don't solder a new chip directly back into the board but put a socket in. That way even if you do fry the first replacement, or even the 2nd it's no big deal while you debug the rest of the board.

I'd get that 04 out before going much further though - if the gate between pins 12 & 13 really is unused then thet IC is probably fried anyway.

I've got 5 ls04's on order.  $2.50 shipped from "ThaiShine" which is one of Tayda's many Ebay stores.  They list their location as Louisville, Colorado, which is where all of my Tayda orders come from.

If I had wanted to spend 2 dollars more, I could have gotten some new-old-stock Fairchild LS04 chips with a 1983 date stamp and kept it looking original. (most of the 2 digit 74 series logic are Fairchild on this board) The LS04's I ordered are Renesas which will kind of stick out.
The larger the government, the smaller the citizen.
 

Offline SeanB

  • Super Contributor
  • ***
  • Posts: 16272
  • Country: za
Re: Hot 74LS04, replace or look elsewhere?
« Reply #18 on: July 11, 2015, 08:37:22 am »
Cut it out and probe again on the stubs ( the board will still not work, but you will see if the TTL is doing anything on the board. Measure both from Vcc and GND, to detect weak pull ups and pull downs with the 10M meter impedance, so you can see floating lines.

Put in a socket for the replacement if you can, though with the one board I worked on the official IC replacement measure ( aside from changing the entire vapour phase reflowed board at a price of $FARK) was to cut the legs off at the body ( or destroy the CERDIP package and remove the ceramic on those, we were FUBAR with the side brazed ceramics, but luckily they were not our trouble children) then clean the inner remains with a needle file and shape the new IC to fit on those stubs, often by gently filing the plastic package down slightly to narrow it. If you say just desolder it then you try to desolder a DIP package where they essentially used both sides of the board filled edge to edge with TTL packages and used solder paste in the holes as soldering. Done to get the board small enough to fit the case. No LSI in 1970, only SSI and MSI, and custon was not common on Mil spec at the time, they did not want to make a custom LSI for a sub 200 production run. Only the USA where the buyer was the government, and they ordered 10 000 planes as an order, was it viable.
 

Offline helius

  • Super Contributor
  • ***
  • Posts: 3632
  • Country: us
Re: Hot 74LS04, replace or look elsewhere?
« Reply #19 on: July 11, 2015, 09:18:26 am »
SeanB: Assemblies that can only be soldered by VPR, can be desoldered with VPR:

But from the description it's not clear why hot air wouldn't work too.
 

Offline SeanB

  • Super Contributor
  • ***
  • Posts: 16272
  • Country: za
Re: Hot 74LS04, replace or look elsewhere?
« Reply #20 on: July 11, 2015, 09:41:27 am »
One DIP package in an assembly of 400, and without damaging the rest of the fragile 20 layer board or the brittle connectors? Plus we did not have the equipment, only standard through hole irons, though I did buy myself a flat pack tip and kept it for myself. made changfing the flat pack chips easy, no issue, no damage to the fragile traces. These boards were made in the 1970's, and would delaminate with only slight heat, so I used a lot of superglue to reattach lifted traces after rework.
 

Offline helius

  • Super Contributor
  • ***
  • Posts: 3632
  • Country: us
Re: Hot 74LS04, replace or look elsewhere?
« Reply #21 on: July 11, 2015, 09:52:40 am »
OK, I understand better the kind of boards you are talking about. Challenging for sure.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf