Author Topic: Tektronix 2465B oscilloscope teardown  (Read 439510 times)

0 Members and 1 Guest are viewing this topic.

Offline Miti

  • Frequent Contributor
  • **
  • Posts: 662
  • Country: ca
Re: Tektronix 2465B oscilloscope teardown
« Reply #1250 on: November 04, 2018, 04:34:37 pm »
I tried to remove two of the knobs from the front panel and both broke, that's how brittle they are. Without removing those knobs there's no way to get to the main board. Anyone has a set of brightness, focus, etc knobs?

I read somewhere that carefully warming up the knobs with a heat gun really helps to remove the stuck ones (after so long most of them will be) without breaking them. I have not tried this yet, and hope for now I will not need to. Knock-knock...

That worked, thanks! :phew:
Now I have work to do.  ;D
« Last Edit: November 04, 2018, 04:36:36 pm by Miti »
That big spark at power up was by design!
 

Offline MarkL

  • Supporter
  • ****
  • Posts: 1627
  • Country: us
Re: Tektronix 2465B oscilloscope teardown
« Reply #1251 on: November 05, 2018, 03:51:43 pm »
I was not sure yet that it had been verified. It seems to be a touchy subject for some members over at TekIO when asked, as can be seen on some past messages. In a situation where a scope lost the calibration constants, some of them will tell you to rather invest the time in getting the scope properly calibrated than to thinker trying to rebuild the cal data with exer02 values, as it might not work, or at best be no longer accurate and on the dot anyhow. I get their point, but then again, my scope is intended for hobby use only, not on a bench at NASA. Also, I can understand the interest in using Exer02 data in a situation where someone lost the cal data, and has no intentions to send the scope out for calibration, or does not have the means to calibrate it himself. 10x better, maybe even 100x better to have your own Exer02 data to play with than having to use a third party bin file from another scope just to bring yours back to life.
I agree with you.

I'm willing to bet the vast majority of these scopes (that are still working) are now in the hands of hobbyists who don't care or need up-to-date calibration.  Having a way to backup and restore the existing cal data is all that's needed.  And the backup part is easy by just taking a video of flipping through EXER 02.

Why calibration backup and restore is not implemented and openly documented for every piece of test equipment, both old and new, continues to mystify me.

On these scopes, I find it strange that Tek would provide a way to display the cal data, but not have some easier way to put it back.  The data itself is meaningless to even a service tech, so why bother displaying it in the first place?.  I continue to be suspicious that there's a hidden way to do it from the screen, but I haven't had time to dig through the code to look for it.
 

Offline MarkL

  • Supporter
  • ****
  • Posts: 1627
  • Country: us
Re: Tektronix 2465B oscilloscope teardown
« Reply #1252 on: November 05, 2018, 04:22:16 pm »
My Tek 2445B developed an interesting belly dance at first power up when cold. I think it is the capacitors on the main board. What do you masters think? It stabilises after 2 - 3 minutes and it is stable after that.
...
After the already mentioned power supply checks, maybe try reseating the horizontal output hybrid U800 (no don't do this - see followup post).  You could also try some freeze spray on components in that area to try to find something sensitive to warm-up.

If it's affecting the whole screen, including the readout as you say, I would highly suspect something around U800 since it's in common to the readout and the traces.  At least it's the place to start.

EDIT - Strike out bad brain advice.
« Last Edit: November 05, 2018, 05:45:43 pm by MarkL »
 

Offline tggzzz

  • Super Contributor
  • ***
  • Posts: 11333
  • Country: gb
    • Having fun doing more, with less
Re: Tektronix 2465B oscilloscope teardown
« Reply #1253 on: November 05, 2018, 05:24:55 pm »
My Tek 2445B developed an interesting belly dance at first power up when cold. I think it is the capacitors on the main board. What do you masters think? It stabilises after 2 - 3 minutes and it is stable after that.
...
After the already mentioned power supply checks, maybe try reseating the horizontal output hybrid U800.  You could also try some freeze spray on components in that area to try to find something sensitive to warm-up.

If it's affecting the whole screen, including the readout as you say, I would highly suspect something around U800 since it's in common to the readout and the traces.  At least it's the place to start.

Be very cautious with U800. There are credible accounts that its mounting bracket can easily be overtightened, destroying that excessively expensive ic.

Personally I'd rule out all other options and positively identify U800 as being the problem before fiddling with it.
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less
 

Offline MarkL

  • Supporter
  • ****
  • Posts: 1627
  • Country: us
Re: Tektronix 2465B oscilloscope teardown
« Reply #1254 on: November 05, 2018, 05:46:29 pm »
My Tek 2445B developed an interesting belly dance at first power up when cold. I think it is the capacitors on the main board. What do you masters think? It stabilises after 2 - 3 minutes and it is stable after that.
...
After the already mentioned power supply checks, maybe try reseating the horizontal output hybrid U800.  You could also try some freeze spray on components in that area to try to find something sensitive to warm-up.

If it's affecting the whole screen, including the readout as you say, I would highly suspect something around U800 since it's in common to the readout and the traces.  At least it's the place to start.

Be very cautious with U800. There are credible accounts that its mounting bracket can easily be overtightened, destroying that excessively expensive ic.

Personally I'd rule out all other options and positively identify U800 as being the problem before fiddling with it.
Oops - tggzzz is absolutely right.  I was thinking of one of the plug-in hybrids.  Don't mess with U800 but you can certainly start your probing in that area.

Previous post updated.
 

Offline tggzzz

  • Super Contributor
  • ***
  • Posts: 11333
  • Country: gb
    • Having fun doing more, with less
Re: Tektronix 2465B oscilloscope teardown
« Reply #1255 on: November 05, 2018, 06:10:25 pm »
My Tek 2445B developed an interesting belly dance at first power up when cold. I think it is the capacitors on the main board. What do you masters think? It stabilises after 2 - 3 minutes and it is stable after that.
...
After the already mentioned power supply checks, maybe try reseating the horizontal output hybrid U800.  You could also try some freeze spray on components in that area to try to find something sensitive to warm-up.

If it's affecting the whole screen, including the readout as you say, I would highly suspect something around U800 since it's in common to the readout and the traces.  At least it's the place to start.

Be very cautious with U800. There are credible accounts that its mounting bracket can easily be overtightened, destroying that excessively expensive ic.

Personally I'd rule out all other options and positively identify U800 as being the problem before fiddling with it.
Oops - tggzzz is absolutely right.  I was thinking of one of the plug-in hybrids.  Don't mess with U800 but you can certainly start your probing in that area.

Previous post updated.

In other areas it is worth being careful not to slip and short out IC pins.

I absolutely refuse to discuss why I mention that :(
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less
 

Offline MarkL

  • Supporter
  • ****
  • Posts: 1627
  • Country: us
Re: Tektronix 2465B oscilloscope teardown
« Reply #1256 on: November 05, 2018, 06:56:16 pm »
...
In other areas it is worth being careful not to slip and short out IC pins.

I absolutely refuse to discuss why I mention that :(
And U800 is especially fun since it has +87V coming into it (pin 7), which is not shown on the A1 schematic.  It's worth looking at the power distribution schematic to know where this and other power pins lie in wait.

It's also good place to use one of those plastic probe tip covers so it can only contact one pin at a time.
 

Offline AMR Labs

  • Regular Contributor
  • *
  • Posts: 142
  • Country: an
Re: Tektronix 2465B oscilloscope teardown
« Reply #1257 on: November 05, 2018, 09:51:52 pm »
And U800 is especially fun since it has +87V coming into it (pin 7), which is not shown on the A1 schematic.  It's worth looking at the power distribution schematic to know where this and other power pins lie in wait.

Uggghhh, yes why did Tek, in such a detailed service manual, not include any meaningful voltage readings in the schematics of at at least some of the more important nodes or power to IC pins??!!

I imagine any of the former Tek guys will always have a good answer/excuse for this. For me a service manual should have absolutely  -everything- I might need to troubleshoot the equipment, and it should rather have too much than too little information. Then again, by looking at some of the horrible mistakes (mixed up power supply caps in wrong positions and polarity, etc) and wonder if these manuals where just rushed out to rather concentrate on equipment sales, and not that much on after sales service literature. Maybe nobody was assigned  to proof read them.

Then again, I've seen some pretty obvious mistakes in the Motorola radio equipment service manuals. You would think that a huge company like them that makes absolutely top quality equipment would not make any/much mistakes of this sort, but I've seen it happen quite often. Makes you wonder how these mistakes still managed to slip through. Then again, those manuals where/are usually at least an inch (or more) thick.

Absolutely best service manuals I have ever come across, and so detailed and brimming with information to the level that virtually nothing was left out, where the service manuals for all the Kenwood ham radio equipment from the 80s I used to work on back then. You really got spoiled with those manuals. But I digress....
« Last Edit: November 05, 2018, 09:54:19 pm by AMR Labs »
 

Offline AMR Labs

  • Regular Contributor
  • *
  • Posts: 142
  • Country: an
Re: Tektronix 2465B oscilloscope teardown
« Reply #1258 on: November 05, 2018, 10:18:41 pm »
I agree with you.

I'm willing to bet the vast majority of these scopes (that are still working) are now in the hands of hobbyists who don't care or need up-to-date calibration.  Having a way to backup and restore the existing cal data is all that's needed.  And the backup part is easy by just taking a video of flipping through EXER 02.

Why calibration backup and restore is not implemented and openly documented for every piece of test equipment, both old and new, continues to mystify me.

On these scopes, I find it strange that Tek would provide a way to display the cal data, but not have some easier way to put it back.  The data itself is meaningless to even a service tech, so why bother displaying it in the first place?.  I continue to be suspicious that there's a hidden way to do it from the screen, but I haven't had time to dig through the code to look for it.

Could not agree more, that would really be something if you would discover some hidden code to do this. I mean, how is it that Tek did not anticipate that a good portion of these scopes would still be in use after the supposedly 10 year NVRAM backup battery expiration, specially at the price level they where sold back then. Build in corporate obsolescence? -Bad NVRam? -> buy a newer model scope! And I don't mean they did not use something better because back then it probably did not exist (EEPROM, FRAM, etc). But at least make it easy to replace the damn chip and reload the data.

Can't blame them for 30-40 years later, people like us still pushing the envelope, though.  ;D :-BROKE

Although I think that someone verified  (maybe it was you) that one can get the NVRam data dump via the GPIB? And even restore it (not so sure about this last part) the same way? But obviously this is not so easy and straight forward unless you know how to set it up and the proper commands, and not all scopes have this bus build in. My scope has GPIB but I would not even know where to start. (Read the book!, read the book!).  :-X
« Last Edit: November 05, 2018, 10:25:33 pm by AMR Labs »
 

Offline MarkL

  • Supporter
  • ****
  • Posts: 1627
  • Country: us
Re: Tektronix 2465B oscilloscope teardown
« Reply #1259 on: November 06, 2018, 02:12:20 am »
...
Although I think that someone verified  (maybe it was you) that one can get the NVRam data dump via the GPIB? And even restore it (not so sure about this last part) the same way? But obviously this is not so easy and straight forward unless you know how to set it up and the proper commands, and not all scopes have this bus build in. My scope has GPIB but I would not even know where to start. (Read the book!, read the book!).  :-X
Yes, it's possible to do the dump and restore via GPIB.  The commands are here (but there's no GUI or wrapper program for it that I'm aware), and it works on all of them in the series:

  https://www.eevblog.com/forum/testgear/tektronix-2465b-oscilloscope-teardown/msg1167050/#msg1167050

It's not a popular method because most scopes in this series are not equipped with GPIB.  And for good reason; there's not much you can do with the GPIB option.  The one thing most people would want is to capture the waveform, but being an analog scope it's just not possible.  The option was targeted at computer assisted setup and manual measurement sequences, and in the "B" case (I'm pretty sure) you can also control the scope's built-in automated waveform measurements.

GPIB control is worth setting up if you have other equipment that supports it.  Lot's of info on this forum on hardware, drivers, and application SW.  Good search terms also include IEEE-488 and HPIB.
 

Offline Miti

  • Frequent Contributor
  • **
  • Posts: 662
  • Country: ca
Re: Tektronix 2465B oscilloscope teardown
« Reply #1260 on: November 06, 2018, 02:30:20 am »

In other areas it is worth being careful not to slip and short out IC pins.

I absolutely refuse to discuss why I mention that :(

Ouch...  :-BROKE
That big spark at power up was by design!
 

Offline AMR Labs

  • Regular Contributor
  • *
  • Posts: 142
  • Country: an
Re: Tektronix 2465B oscilloscope teardown
« Reply #1261 on: November 06, 2018, 02:37:08 am »
Just out of curiosity I searched on ebay for a "USB-to-GPIB converter", and only two results came up, both from the UK. One priced at around $1200, and the other for a little over $1000. Both units look exactly the same. Is this really what an interface of this type costs, or did I miss something?
 

Offline Miti

  • Frequent Contributor
  • **
  • Posts: 662
  • Country: ca
Re: Tektronix 2465B oscilloscope teardown
« Reply #1262 on: November 06, 2018, 02:37:45 am »
...
In other areas it is worth being careful not to slip and short out IC pins.

I absolutely refuse to discuss why I mention that :(
And U800 is especially fun since it has +87V coming into it (pin 7), which is not shown on the A1 schematic.  It's worth looking at the power distribution schematic to know where this and other power pins lie in wait.

It's also good place to use one of those plastic probe tip covers so it can only contact one pin at a time.


I measured the voltage on pin 7, it is 86.18V sharp on DMM and scope on AC 500mV, straight line, no move whatsoever. I'll do more tests in the weekend when I can find some time. It looks like U800 has been replaced or at least resoldered judging by the amount of solder and flux around the pins. It is a Tek chip, not Maxim and after 5 - 6 minutes it is barely warm, I don't think it needs a heatsink.
That big spark at power up was by design!
 

Offline AMR Labs

  • Regular Contributor
  • *
  • Posts: 142
  • Country: an
Re: Tektronix 2465B oscilloscope teardown
« Reply #1263 on: November 06, 2018, 02:42:41 am »
I measured the voltage on pin 7, it is 86.18V sharp on DMM and scope on AC 500mV, straight line, no move whatsoever. I'll do more tests in the weekend when I can find some time. It looks like U800 has been replaced or at least resoldered judging by the amount of solder and flux around the pins. It is a Tek chip, not Maxim and after 5 - 6 minutes it is barely warm, I don't think it needs a heatsink.

I might be wrong, and please do correct me if I am, but I think I read somewhere that the temperature of U800 will vary depending on the "load" it is put through. On an idle scope it might remain cool to the touch, but might get hotter while trying to handle more complex or fast signals. Correct?
 

Online med6753

  • Super Contributor
  • ***
  • Posts: 5724
  • Country: us
  • Tek nut
Re: Tektronix 2465B oscilloscope teardown
« Reply #1264 on: November 06, 2018, 05:45:09 am »
I measured the voltage on pin 7, it is 86.18V sharp on DMM and scope on AC 500mV, straight line, no move whatsoever. I'll do more tests in the weekend when I can find some time. It looks like U800 has been replaced or at least resoldered judging by the amount of solder and flux around the pins. It is a Tek chip, not Maxim and after 5 - 6 minutes it is barely warm, I don't think it needs a heatsink.

I might be wrong, and please do correct me if I am, but I think I read somewhere that the temperature of U800 will vary depending on the "load" it is put through. On an idle scope it might remain cool to the touch, but might get hotter while trying to handle more complex or fast signals. Correct?

I believe that is a true statement.
An old gray beard with an attitude.
 

Offline AMR Labs

  • Regular Contributor
  • *
  • Posts: 142
  • Country: an
Re: Tektronix 2465B oscilloscope teardown
« Reply #1265 on: November 07, 2018, 03:07:19 am »
Are you sure that counter is resettable? It is the ON counter and probably it is stored in the same infamous NVRAM that I replaced this summer.

I got a confirmation today that yes, those 2465 total hours and power on/off counters are not only resettable to zero, but they can also be set to anything you want. CAL05 is the routine that can read/set/reset the counters. Default is to reset. Requires to select the increment using the X10 button, and to rotate the Delta Ref, and the Delta controls to make the appropriate counter change.

Third party information. I have not played with these counters as of yet on my own 2465B.
 

Offline james_s

  • Super Contributor
  • ***
  • Posts: 11662
  • Country: us
Re: Tektronix 2465B oscilloscope teardown
« Reply #1266 on: November 07, 2018, 07:56:53 am »
Just out of curiosity I searched on ebay for a "USB-to-GPIB converter", and only two results came up, both from the UK. One priced at around $1200, and the other for a little over $1000. Both units look exactly the same. Is this really what an interface of this type costs, or did I miss something?

Maybe if you buy a commercial one? I don't know. I think I spent about $10 building mine, most of that was a GPIB cable which I chopped in half and wired to a Arduino nano clone. I found the firmware online somewhere, it's not something I need often but I've used it to clear the error log and enable options on my scope.
 

Offline Gixy

  • Regular Contributor
  • *
  • Posts: 211
  • Country: fr
Re: Tektronix 2465B oscilloscope teardown
« Reply #1267 on: November 07, 2018, 08:34:35 am »
The procedure is official and perfectly described in the service manual.
 

Offline AMR Labs

  • Regular Contributor
  • *
  • Posts: 142
  • Country: an
Re: Tektronix 2465B oscilloscope teardown
« Reply #1268 on: November 07, 2018, 12:21:32 pm »
Looking at a 28 pin SOIC to DIP converter board to mount the FRAM chip and be able to plug it into the 28 DIP machine tooled socket I plan to install at some point once I desolder the original DS1225. Recommended converter board seems to be Aries 28-650000-10 (Mouser PN 535-28-650000-10 $18.55) but the part is currently out of stock until mid December, and then they only ordered 13 pieces. Digikey has the part, but its usually not my favorite go-to supplier, besides I already have all the other parts I wanted to order in the cart at Mouser. I don't know of other main stream suppliers that might have this adapter in stock.

Anyway, at this point mainly just wanted to know if I am on the right track, can anyone confirm that the Aries 535-28-650000-10 is the correct part to mount a FM16W08? Or if Mouser might have a viable alternative in stock? I searched at Mouser for other similar options but found none where available. The machined socket I'm planning on installing is Mouser PN 571-2-1571552-9  $3.43.

On ebay there are plenty of these chinese 28 pin SOIC-DIP adapters but its only the plain board, they don't have any pins soldered onto them, so not the ideal part, as I would prefer a ready-made solution like the Aries converter above anyhow.  And it usually takes around a month or more for the chinese stuff to arrive.
 

Online macboy

  • Super Contributor
  • ***
  • Posts: 2008
  • Country: ca
Re: Tektronix 2465B oscilloscope teardown
« Reply #1269 on: November 08, 2018, 02:54:09 pm »
Just out of curiosity I searched on ebay for a "USB-to-GPIB converter", and only two results came up, both from the UK. One priced at around $1200, and the other for a little over $1000. Both units look exactly the same. Is this really what an interface of this type costs, or did I miss something?
On ebay, you can easily find Agilent 82357b USB/GPIB adapters for well under $100. Beware that most are clones/counterfeit, not Agilent-made, especially those sold as New. Fortunately, the clones are straight copies with the same 75161 and 75160 proper GPIB bus driver ICs, the same CPLD and USB interface ICs, and copied PCB layout. They use the Agilent I/O suite drivers.  The inside of the clones' plastic case is usually not plated for RFI shielding, probably not a big deal. Mine is a clone but works perfectly for me. Some people complain about issues especially with many instruments on the bus... GPIB can be finicky and not all instruments play nice, e.g. I have one older instrument which always "listens" so the drivers think there are 30 devices connected, since something always handshakes for any listen address.  Since GPIB-to-GPIB cables (for daisy-chaining devices) cost nearly as much as these adapters, don't bother, just get one per instrument that you must have connected at the same time, and/or swap the adapter(s) between instruments as needed.  Having one device per bus also greatly simplifies scripts you write since you can hog the bus instead of trying to write fully asynchronous scripts.


FS: Tek 2465-2467 front cover and probe pouch
« Last Edit: November 08, 2018, 03:13:07 pm by macboy »
 

Offline Miti

  • Frequent Contributor
  • **
  • Posts: 662
  • Country: ca
Re: Tektronix 2465B oscilloscope teardown
« Reply #1270 on: November 10, 2018, 06:27:12 pm »
Next step then is to monitor the supply voltages at J119 on the main board and see if one of them is bouncing around. And with that many hours have the Inverter and Regulator boards ever been recapped? If not I highly recommend it be done.

I recapped the power supply few years ago with good Nichicon PW series caps.

OK, but I would still monitor the PSU voltages at initial power up and see if something is amiss. Rule that out first before digging into it further.

Ok, I think I found the reason for the jitter. It is +15V that bounces around for a minute or so and then stabilizes. The question now is if the LM317 - U1260 is defective or the problem comes from the input. The difficulty with this defect is that it goes away after about 1 minute and I need to wait for hours before it shows up again for a minute or less.
One thing that I've noticed is that the unregulated voltages are a bit low. For example +15V Unreg should be 19.2V according to the schematic but I measure only +18.4V, +5VD should be +5.1V but in my case it is +5.00V.
Maybe  I'm on the edge of the minimum VIO for LM317.
That big spark at power up was by design!
 

Offline tggzzz

  • Super Contributor
  • ***
  • Posts: 11333
  • Country: gb
    • Having fun doing more, with less
Re: Tektronix 2465B oscilloscope teardown
« Reply #1271 on: November 10, 2018, 10:02:24 pm »
Next step then is to monitor the supply voltages at J119 on the main board and see if one of them is bouncing around. And with that many hours have the Inverter and Regulator boards ever been recapped? If not I highly recommend it be done.

I recapped the power supply few years ago with good Nichicon PW series caps.

OK, but I would still monitor the PSU voltages at initial power up and see if something is amiss. Rule that out first before digging into it further.

Ok, I think I found the reason for the jitter. It is +15V that bounces around for a minute or so and then stabilizes. The question now is if the LM317 - U1260 is defective or the problem comes from the input. The difficulty with this defect is that it goes away after about 1 minute and I need to wait for hours before it shows up again for a minute or less.
One thing that I've noticed is that the unregulated voltages are a bit low. For example +15V Unreg should be 19.2V according to the schematic but I measure only +18.4V, +5VD should be +5.1V but in my case it is +5.00V.
Maybe  I'm on the edge of the minimum VIO for LM317.

I had a vaguely similar problem on my Tek 485. It would take 60s to startup, and then work. If I switched it off for a few seconds, it started immediately. If I switched it off for a 30 mins, it would restart after 15s. It would only tale 60 to startup if it had been off for >6hours. That made fault location slow!

I eventually traced the problem to an old electrolytic in the startup circuit. It felt as if it took 60s for the dielectric to reform, and several hours for the dielectric to decay. So, if there is an old electrolytic "near" that LM317, it might be worth replacing it.
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less
 

Offline Miti

  • Frequent Contributor
  • **
  • Posts: 662
  • Country: ca
Re: Tektronix 2465B oscilloscope teardown
« Reply #1272 on: November 11, 2018, 01:35:07 am »
Unfortunately there are no electrolytic caps around the LM317 so it's either the input or U1371. I'll keep you posted.
That big spark at power up was by design!
 

Offline Miti

  • Frequent Contributor
  • **
  • Posts: 662
  • Country: ca
Re: Tektronix 2465B oscilloscope teardown
« Reply #1273 on: November 11, 2018, 02:45:28 am »
Could anyone measure the unregulated voltages in a working 2465B? My scope is 2445B but I think the PS is identical to 2465B. Not sure about 2467.
That big spark at power up was by design!
 

Offline AMR Labs

  • Regular Contributor
  • *
  • Posts: 142
  • Country: an
Re: Tektronix 2465B oscilloscope teardown
« Reply #1274 on: November 11, 2018, 03:14:51 pm »
Ok, I think I found the reason for the jitter. It is +15V that bounces around for a minute or so and then stabilizes. The question now is if the LM317 - U1260 is defective or the problem comes from the input. The difficulty with this defect is that it goes away after about 1 minute and I need to wait for hours before it shows up again for a minute or less.
One thing that I've noticed is that the unregulated voltages are a bit low. For example +15V Unreg should be 19.2V according to the schematic but I measure only +18.4V, +5VD should be +5.1V but in my case it is +5.00V.
Maybe  I'm on the edge of the minimum VIO for LM317.

I am assuming that any of the unregulated voltages may vary 10% from what is stated in the manual, in accordance to the mains voltage variation at your location. Also in general, maybe not directly related to your problem but good to keep in mind is that all regulated voltages of the power supply are referenced to the +10 supply, and any deviation there will reflect on all others. This power supply should measure between +9.99 and +10.01V. Anything outside is considered out of spec. Problem is, if you attempt to readjust it, you might affect a whole bunch of other internal scope calibrations as well, so best to leave as is, unless its seriously out of spec, or you intend to perform a full scope calibration.

I would try to measure the input voltage to the LM317, if you also see variations there you know the problem is originating before the regulator. Also check the reference input of the regulator. I would think that an input of +18.4v as opposed to the expected +19.2v should make no difference whatsoever to get a stable 15V output as it is more than 3v above which I would think is still very much ok.

Maybe solder a couple of temporary flying wires at the points of interest that you can easily access to quickly make the measurement as soon as you turn on the scope without having to move things around too much.
« Last Edit: November 11, 2018, 03:24:33 pm by AMR Labs »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf