(Topic ID: 252890)

IC’s operating temperature

By oldschoolbob

4 years ago


Topic Heartbeat

Topic Stats

  • 44 posts
  • 9 Pinsiders participating
  • Latest reply 4 years ago by mbwalker
  • No one calls this topic a favorite

You

Linked Games

No games have been linked to this topic.

    Topic Gallery

    View topic image gallery

    pic_71_4 (resized).jpg
    pic_71_3 (resized).jpg
    pic_71_2 (resized).jpg
    pic_71_1 (resized).jpg
    IMG_3868 (resized).JPG
    meters 04 (resized).jpg
    meters 03 (resized).jpg
    meters 02 (resized).jpg
    pasted_image (resized).png
    12 and 5 volt power supply r6a (resized).jpg
    Final layout (resized).jpg
    pcb layout 15 (resized).jpg
    IMG_3073 (resized).JPG
    pasted_image (resized).png
    pasted_image (resized).png
    IMG_3846 (resized).JPG

    #1 4 years ago

    Tonight after I finished replacing the headers on a MPU 100 I decided to bench test the new work. Using my oscilloscope I was testing for activity on the new headers. The board was in attract mode (seven flashes). When I laid my arm on the board I noticed U7 and U8 felt quite warm. (About 90 degrees). After I finished my tests I checked the temperature. I needed something to compare so I turned on my Meteor. The ROMs were all about the same - 90 degrees. However, U7, U8, U9, U10, and U11 were all about 8 to 10 degrees warmer on the MPU 100 on the bench.

    Is this normal?
    Should I be concerned?
    What is the normal operating temperature of these IC’s?

    Thanks
    Bob

    #2 4 years ago

    Normal temperature range for plastic commercial grade devices is 0-70 C or 32-158 F.

    #3 4 years ago

    °C or °F?

    #4 4 years ago

    To answer the question regarding is it normal -- It is normal for many ROMs to run hot. Depends on the technology used to create the ROM. Parts with CMOS (low power cores) tend to run cooler. But based on the date of that machine - it is probably a bipolar ROM... these normally run hot.
    And, the ROMs that have more activity on them will run hotter than those with less activity.

    #5 4 years ago

    Sorry, I forget we're worldwide on Pinside. I was using fahrenheit.

    The ROMs in both games were about the same temperature. But why would the other IC's be hotter on the bench test. I'd think they'ed be hotter in the game. The rest of the boards (both in the game and on the bench) were about the same 78 to 80 degrees F.

    Bob

    #6 4 years ago

    The hotter chips aren't AMI brand are they? I notice they run hotter.
    What voltage are you reading at the 5V rail on the MPU board?

    #7 4 years ago

    Hi Quench

    Most of the chips are Motorola. No AMI chips.

    The bench power supply reads 5.28 VDC. TP5 reads 4.6 VDC.

    The Meteor game reads 5.02 VDC at TP5.

    I thought you were onto something but it looks like the game's voltage is higher.

    Bob

    #8 4 years ago
    Quoted from oldschoolbob:

    TP5 reads 4.6 VDC

    TP5 is only 4.6 volts? Where are you losing the 0.7 volts from the power-supply to the board? What current reading are you getting on the 5V rail on your bench power-supply?

    #9 4 years ago

    I didn't know the 0.7 voltage loss was a problem - I guess it is.

    Today I did some checking. I checked both the Stern board and the Bally. I got similar results.

    To refresh your memory - the power supply has meters on the 12 volt line and the 5 volt line. But only the 12 V has amps (5 V only reads volts.) The 12 V reads the amps for both 12 and 5 volts.

    The 5 V meter reads 5.32 - When I connected my DMM to the test probe I read 5.31 V. No loss there.

    Then I connected the MPU and booted it up. I checked the TP5 with the DMM and got 4.59 V. The amp meter on the 12 V meter read 0.44 amps. (remember that's both 12 and 5 volts). Then I disconnected the 5 V probe and the amp meter read 0.06 amps. I guess that's telling me the 5 V is pulling 0.38 amps.

    I hope I haven't lost you in my explanation.

    Bob

    IMG_3822 (resized).JPGIMG_3822 (resized).JPGIMG_3825 (resized).JPGIMG_3825 (resized).JPGIMG_3826 (resized).JPGIMG_3826 (resized).JPG
    #10 4 years ago

    Check your test points and measurement points. Make sure they are clean and corrosion free. Most likely this is poor contact point since you are not pulling much current.

    #11 4 years ago

    Thanks Dave

    You brought up a good point. I checked the resistance from the alligator clip to R23 (first component) and got 0.1 to 0.0 ohms. Then I checked the ground connection. From clip to ground plain I got the same low resistance. I'm sure the board is clean.

    But a good suggestion.

    Thanks

    Bob

    #12 4 years ago
    Quoted from oldschoolbob:

    I hope I haven't lost you in my explanation.

    Yeah I remember most of it. But there was something about the current meter on the ground wire.

    Where do you measure the 0.7 voltage loss though?
    With a board powered on, measure the voltage across the 5V source on the power-supply banana connectors, then between the black banana connector and ground on the MPU board, then between the 5V banana connector to 5V rail on the MPU board.

    Quoted from oldschoolbob:

    Today I did some checking. I checked both the Stern board and the Bally. I got similar results.

    Back to the original question, how about the chip temperatures between the Bally and Stern boards on the bench? Similar or different.

    BTW, can I make a suggestion, every PC power-supply has the red wires as 5 volts and yellow wires as 12 volts. It's almost defacto. Feel like changing them?

    #13 4 years ago
    Quoted from Quench:

    But there was something about the current meter on the ground wire.

    The current meter can only read through the ground and I have a common ground. So the current meter reads both 5 and 12 volts.

    The voltage loss seems to be on the board. With the 5 volt connected to TP5 (and the board booted) I read 4.71 V at the alligator clip. When I remove the clip I read 5.21 Volts.

    Then I connected TP5 only (no 12 volts) and turned it on (so it wasn't booted) and I still read 4.67 Volts.

    Quoted from Quench:

    red wires as 5 volts and yellow wires as 12 volts

    Leave it to me to get it backwards. You'll notice I marked the clips TP2 and TP5. I think I can clean off the markings and switch them then re-mark. Thanks.

    I'll check the temperatures and be right back.

    Bob

    #14 4 years ago

    What about when the board is booted, what voltage do you measure across the power-supplies banana connectors (the connector outputs for your 5 volt supply)?

    #15 4 years ago

    Temperatures Stern v Bally v Meteor (in Fahrenheit) :

    Note: Stern and Bally were on the bench. Meteor was in the game.

    U2 – S = 88, B = 80, M = 90
    U6 – S = 89, B = 88, M = 90
    U7 – S = 90, B = 89, M = 82
    U8 – S = 82, B = 90, M = 77
    U9 – S = 89, B = 83, M = 80
    U10 – S = 85, B = 78, M = 78
    U11 – S = 86, B = 82, M = 77
    Middle of board – S = 80, B=79, M = 78

    Strange thing when I measured U8 – left side measured 90 and right side measured 83.

    Bob

    #16 4 years ago
    Quoted from Quench:

    What about when the board is booted, what voltage do you measure across the power-supplies banana connectors (the connector outputs for your 5 volt supply)?

    Wouldn't that be the same as measuring at the alligator clips? I can't measure at the banana connectors unless I pull them out.

    Bob

    #17 4 years ago
    Quoted from oldschoolbob:

    Strange thing when I measured U8 – left side measured 90 and right side measured 83.

    Might be stray heat coming from the 1 watt 82 ohm resistor near it.

    Are the warmer ROMs all ceramic packaging?

    #18 4 years ago
    Quoted from oldschoolbob:

    Wouldn't that be the same as measuring at the alligator clips?

    Ideally yes, but you're getting loss somewhere so we need to see if it's in the leads or the power-supply.

    Quoted from oldschoolbob:

    I can't measure at the banana connectors unless I pull them out.

    Just pull the banana connectors out 1/16" so you can get your multi-meter leads on the connectors to measure.

    #19 4 years ago
    Quoted from Quench:

    Might be stray heat coming from the 1 watt 82 ohm resistor near it.

    That could be because I was shooting from the left over R11.

    Quoted from Quench:

    Are the warmer ROMs all ceramic packaging?

    How can I tell the difference? See photos:

    Bally
    IMG_3801 (resized).JPGIMG_3801 (resized).JPG

    Stern

    IMG_3756 (resized).JPGIMG_3756 (resized).JPG

    I'm guessing the Stern U2 is the only ceramic.

    Quoted from Quench:

    Just pull the banana connectors out 1/16" so you can get your multi-meter leads on the connectors to measure.

    I'll try that tomorrow. I think I may have a banana connector with another receptial for a second banana connector. I'll see if I can find it.

    Thanks.

    Bob

    #20 4 years ago

    The purple looking chips are ceramic packages, all the black ones are plastic. Looks like the CPU (Or is that a PIA?) on EK is ceramic, and the ROM on Nugent as well. Everything else is plastic.

    Ceramic packages are where the housing for the chip is actually made of a ceramic material.

    #21 4 years ago
    Quoted from oldschoolbob:

    I'm guessing the Stern U2 is the only ceramic.

    Yep, the purple chips are ceramic: Your Stern U2 ROM and also that Bally U9 CPU. You might see some white Fujitsu EPROMs on these early boards that are ceramic too.
    Ceramic is a better thermal conductor than plastic so should be better at dispersing heat away from the internal silicon die but it can depend on the silicon technology/construction used.

    #22 4 years ago

    Not much shop time today. Tonight I booted the board and checked voltages.
    Power supply meter shows 5.32 VDC.
    Then I booted the board and checked TP5 = 4.71
    I pulled out the banana plug enough to get my probe on. 4.96 VDC.
    I checked those voltages twice. Same each time.
    Then I removed the clip from TP5 and read 5.21 at the clip and at the banana plug.

    Is my power supply not keeping up with the power requirements?

    Why does the power supply meter not show the power loss? It read 5.32 through all the tests.

    Thanks

    Bob

    #23 4 years ago

    Bob,

    The problem you're going to encounter is manufacturer's rate their parts (operating, not storage) for the die temperature, not the case temperature. Now, this is where it gets complicated: To determine the die temperature, you need to determine the power dissipated in the part, multiply that by the thermal resistance of the part, do the same for the thermal resistance of the part to the PCB board too, and also figure in the ambient temperature. Yikes! I know that's not the answer you are looking for. Point is, there's a lot a viables to factor in here and as Quench eluded too...some parts are better thermally than others.

    In the end, if you think a part is operating correctly (i.e. dissipating the correct amount of power) and everything is working OK but just warm (maybe from a poor design), you can always try adding a thermal pad under the part which will help wick away the heat by lowering the 'case to PWB' thermal resistance. Look up 'sil pad' or 'gap pad'. Not the ideal solution, but it may help.

    Sorry for the long winded answer.

    #24 4 years ago
    Quoted from oldschoolbob:

    Then I booted the board and checked TP5 = 4.71
    I pulled out the banana plug enough to get my probe on. 4.96 VDC.

    So you're losing voltage between the output connectors of the power-supply and the alligator clips on the end of the cable when loaded. Didn't you have some connection issues with those alligator jumper leads back when you were building your bench power-supply? You might want to check each end of those leads and make sure you have good solid connections to the alligator clips/male banana plugs.

    #25 4 years ago

    most alligator clips use piddly tiny wires with a fat insulation jacket to make it look like gauge. They are also usually just crimped down which may not make a good connection. Maybe put a molex plug on it to plug into mpu j4 or replace the wire in the gator clips and terminate them better than how they come when you buy them.

    #26 4 years ago
    Quoted from barakandl:

    Maybe put a molex plug on it to plug into mpu j4

    Good idea, I've done this too.

    #27 4 years ago
    Quoted from Quench:

    Didn't you have some connection issues with those alligator jumper leads back when you were building your bench power-supply?

    Quench, your memory is amazing. Yes, I did have some problem with the alligator jumper leads but that was with the old power supply and old leads. This seems to be a similar situation.

    Looking over the schematic, the 5 volt comes in at J4 and goes through L2 before it gets to TP5. If connecting to TP5 directly would L2 not be connected? Would that make a difference? What if I connect to the left side of L2 instead of TP5?

    Also C5 is broken (haven't replaced it yet). Would that make a difference?

    I like the idea of a plug for J4 (Quench suggested that a long time ago). That would also help eliminate the chance of me hooking up the test probes wrong. But unfortunately most of the time when I work on a board I don't have the headers installed. That's usually the last thing I install.

    Thanks

    Bob

    #28 4 years ago

    Not much shop time again today. Tonight I installed C5. Made no difference in the voltage. Then I moved my 5 volt clip to the left side of L2 (that would be the same as J4 pin 16 and 17). Still no difference.

    Then I inspected the alligator clip and wire. Andrew was right. Small wire and large insulation - plus the wire looks like it's just bent down and crimped with the insulation. Crappy test leads. I may try to solder the wire to the clip but that won't fix the wire gauge. And I can't take the banana plug apart to inspect that end.

    Any suggestions on where to get a decent set of test leads. I think I got these from Jameco.

    https://www.jameco.com/z/TLM3-Velleman-Set-of-3-Alligator-Clip-Test-Leads-with-Booted-Banana-Plugs_2212293.html

    Thanks

    Bob

    S20191009_0003 (resized).jpgS20191009_0003 (resized).jpg
    #29 4 years ago
    Quoted from oldschoolbob:

    Not much shop time again today. Tonight I installed C5. Made no difference in the voltage. Then I moved my 5 volt clip to the left side of L2 (that would be the same as J4 pin 16 and 17). Still no difference.
    Then I inspected the alligator clip and wire. Andrew was right. Small wire and large insulation - plus the wire looks like it's just bent down and crimped with the insulation. Crappy test leads. I may try to solder the wire to the clip but that won't fix the wire gauge. And I can't take the banana plug apart to inspect that end.
    Any suggestions on where to get a decent set of test leads. I think I got these from Jameco.
    https://www.jameco.com/z/TLM3-Velleman-Set-of-3-Alligator-Clip-Test-Leads-with-Booted-Banana-Plugs_2212293.html
    Thanks
    Bob[quoted image]

    Just make your own!

    #30 4 years ago
    Quoted from mbwalker:

    Just make your own!

    I considered that. And I might just have to do it. Seems a shame you can't just buy a good set without taking out a loan. Mark, where did you get your test leads?

    I dug into my parts drawer and found this set. The wire size seems heavier (at least the insulation is thicker). The insulation on the crappy set measures 0.11 - these measure 0.155. (Can't see the wire gauge without cutting). I didn't like them because the alligator clips are big, bulky and stiff.

    I connected one to the 5 volt line and booted the board. It shows a little better - 4.93 volts - both at the clip and the banana plug.

    I'm thinking about cutting off the alligator clips and putting on smaller clips.

    By the way, Why does the meter on my power supply read 5.32 volts and when booted I read 4.91 at the banana plug? Shouldn't the meter and the plug read the same?

    Thanks

    Bob
    IMG_3842 (resized).JPGIMG_3842 (resized).JPG

    IMG_3846 (resized).JPGIMG_3846 (resized).JPG
    #31 4 years ago

    I'd just google 'test lead wire'. I'm sure Amazon has something appropriate. If you're just measuring voltages (i.e. no current), then a thinner gauge (18 or 16 gauge, flexible) would be OK since it's a high impedance measurement. For higher current measurements, you'll need thicker, short wire to minimize voltage drop in the test leads. Solder the connections too. I typically make high current measurements with a semi-custom set of wires. Short as possible if a critical measurement.

    One other neat thing you could do is use a clamp on current probe for your scope. That lets you look at current in the time domain - very useful! The voltage drop due to the current meter and leads are no longer an issue. I added a SniptIt of a time domain current measurement that you couldn't measure w/a meter.
    pasted_image (resized).pngpasted_image (resized).png
    pasted_image (resized).pngpasted_image (resized).png

    #32 4 years ago

    I always thought test lead wire was like that on purpose. The extra-thick insulation is there for safety with high voltage circuits measured in the thousands of volts, and the thin conductor acts like like a fusable link to protect against huge amperage spikes.

    -Hans

    #33 4 years ago
    Quoted from HHaase:

    I always thought test lead wire was like that on purpose. The extra-thick insulation is there for safety with high voltage circuits measured in the thousands of volts, and the thin conductor acts like like a fusable link to protect against huge amperage spikes.
    -Hans

    The silicone insulation is great for both high voltage and high temperature, plus it's fairly flexible. I worry more about a soldering iron inadvertently touching the insulation than high voltage.

    Never heard of the 'fusible link' comment. If that was a concern, I would think the leads would incorporate a real fuse. You definitely don't want thinner leads on a current measurement since it would result in a voltage drop. A fair amount of meters incorporate an internal fuse for the current measurements.

    #34 4 years ago
    Quoted from oldschoolbob:

    Not much shop time again today. Tonight I installed C5. Made no difference in the voltage. Then I moved my 5 volt clip to the left side of L2 (that would be the same as J4 pin 16 and 17). Still no difference.
    Then I inspected the alligator clip and wire. Andrew was right. Small wire and large insulation - plus the wire looks like it's just bent down and crimped with the insulation. Crappy test leads. I may try to solder the wire to the clip but that won't fix the wire gauge. And I can't take the banana plug apart to inspect that end.
    Any suggestions on where to get a decent set of test leads. I think I got these from Jameco.
    https://www.jameco.com/z/TLM3-Velleman-Set-of-3-Alligator-Clip-Test-Leads-with-Booted-Banana-Plugs_2212293.html
    Thanks
    Bob[quoted image]

    Yeah there is no way that alligator clip could handle much current. The termination is literally stripped wire folded back under the insulation jacket and then crimped down. If you want to stick with alligator clip leads replace the wire. Crimp the insulation jacket and solder the wire leads to the gator clip. Use a fatter gauge of wire.

    I am pretty sure the gator clips are made that way just to be cheap and the fat insulation is just a way to make them look beefy when they are actually cheap ass thin wire.

    I wonder what the metal teeth of the gator to a PCB test post current rating really is. I don't see anywhere where they rate the current of these things.

    #35 4 years ago

    I tried to solder the wire to the clip but it's not easy to get in there. Just melted the insulation and made a mess. I think I'll replace those monster clips with some smaller ones. Will see how that works.

    Quoted from barakandl:

    I wonder what the metal teeth of the gator to a PCB test post current rating really is. I don't see anywhere where they rate the current of these things.

    That's a good point. Besides the test point wire isn't much to attach to. I think when I was taking my voltage readings I was probably probing at the test point. Next time I'm in the shop I think I'll probe on the clip.

    Bob

    #36 4 years ago
    Quoted from oldschoolbob:

    Why does the power supply meter not show the power loss? It read 5.32 through all the tests.

    Can anyone answer the question from post 22 ? If I'm reading 4.96 VDC at the banana plug, Shouldn't the power supply meter be reading the same?

    Quench and Mark, you guys know more about this power supply than anyone - including me.

    Thanks

    Bob

    IMG_3073 (resized).JPGIMG_3073 (resized).JPG
    #37 4 years ago

    Post the schematic for the power supply again. Suspect it could be because of the current meter on ground.

    #38 4 years ago

    Thanks Quench,

    In the photo above - the white test lead is negative, red is 5 volts, green is 12 volts.

    The 5 volt meter only has 3 wires. red is power, white is sence, black is ground. White and red are tied together and connected to the back of the 5 volt banana plug. Black is connected to the ground banana plug (along with the 12 volt meter ground).

    I think this is the latest schematic (not real sure). But the PC board layout IS the latest.

    12 and 5 volt power supply r6a (resized).jpg12 and 5 volt power supply r6a (resized).jpgFinal layout (resized).jpgFinal layout (resized).jpgpcb layout 15 (resized).jpgpcb layout 15 (resized).jpg
    #39 4 years ago

    Hi Bob,

    Still sorting out the posts, sorry if I missed something.

    Mind putting your scope on the 5V line when you are connected to the board and drawing current? You would be making sure there's nothing funny going on w/the power supply 'folding back', and that the VDC is nice and consistent (i.e. a flat line on the scope, no ripple, etc.).

    If the supply is good, then the issue would be resistance somewhere between the power supply and the board under test.

    Don't forget the ground wire too. The same current going out the +5V wire is also being returned in the ground wire. So any resistance there will also show up as a voltage drop too.

    Lastly, looking at your picture...starting at the power supply: Is the voltage meter the first thing, then the current meter? Ideally, the voltage meter would be attached at the output side of the current meter, not the input side. Then you can compensate for any voltage drop due to the current meter.

    I think you need to move the voltage meter red wire *IF* the big fat red wire on the current meter is the output. Like this:
    pasted_image (resized).pngpasted_image (resized).png

    #40 4 years ago

    Thanks Mark, putting this on the scope is a great idea (I should have thought of that). Maybe it will tell us more of what's happening. Unfortunately I'm rebuilding my banana plug test leads and can't test anything right now. Once they're done I'll put it on the scope. I'm also rebuilding the ground lead as well.

    The big red wire you pointed to actually connects to the ground banana plug. Remember the current meter reads on the negative side.

    I tried to sketch out where all these wires are going but it got really confusing. So I broke it into three drawings. First is the 5 volt meter. Second is the 12 volt meter. Third is both meters. I hope this helps us understand what's connected.

    Thanks

    Bob

    meters 02 (resized).jpgmeters 02 (resized).jpgmeters 04 (resized).jpgmeters 04 (resized).jpgmeters 03 (resized).jpgmeters 03 (resized).jpg
    #41 4 years ago

    What a difference a good set of test leads made. I cut off those bulky alligator clips and checked the wire size. Big difference in sizes - see photo.

    Then I soldered on new smaller alligator clips. With all three leads done (5 V, 12 V, ground), I hooked them up to the power supply and tested with the scope.

    First scan is at the banana plug with nothing connected.

    Second scan is at the alligator clip with nothing connected.

    Third scan is booted up at the banana plug.

    Fourth scan is booted up at the clip.

    All four scans are the same. After thinking about it for a while I thought my DMM must be off. So I tested everything again this time using my DMM instead of the scope.

    At banana plug with nothing connected = 5.31 V

    At the clip with nothing connected = 5.31 V

    (power supply meter was showing 5.32 V)

    At banana plug - booted = 5.26

    At clip - booted = 5.26

    (power supply meter was showing 5.28 V)

    Everything seems to be in order - lesson learned = don't use crappy test leads.

    Thanks everyone for all the help.

    And another big thank you to all that helped me build this fantastic power supply last year.

    Bob

    IMG_3868 (resized).JPGIMG_3868 (resized).JPGpic_71_1 (resized).jpgpic_71_1 (resized).jpgpic_71_2 (resized).jpgpic_71_2 (resized).jpgpic_71_3 (resized).jpgpic_71_3 (resized).jpgpic_71_4 (resized).jpgpic_71_4 (resized).jpg
    #42 4 years ago

    Well done, great to see you guys sorted it out!

    #43 4 years ago

    By the way Quench, I got the wire colors right this time. Thanks.

    #44 4 years ago
    Quoted from oldschoolbob:

    What a difference a good set of test leads made.

    Way to go! Case closed.

    BTW, you 5 V meter ground lead is in the right place...after the current meter. Good.

    Reply

    Wanna join the discussion? Please sign in to reply to this topic.

    Hey there! Welcome to Pinside!

    Donate to Pinside

    Great to see you're enjoying Pinside! Did you know Pinside is able to run without any 3rd-party banners or ads, thanks to the support from our visitors? Please consider a donation to Pinside and get anext to your username to show for it! Or better yet, subscribe to Pinside+!


    This page was printed from https://pinside.com/pinball/forum/topic/ics-operating-temperature?hl=skidave and we tried optimising it for printing. Some page elements may have been deliberately hidden.

    Scan the QR code on the left to jump to the URL this document was printed from.