Keyboard Shortcuts
ctrl + shift + ? :
Show all keyboard shortcuts
ctrl + g :
Navigate to a group
ctrl + shift + f :
Find
ctrl + / :
Quick actions
esc to dismiss
Likes
Search
My 2465B has arrived + A5 Board leaky caps cleanup
|
This is a 2465B Sn 55K purchased off ebay, very good external cosmetics, almost pristine interior with only very few traces of dust and so far have not seen any signs of having been disturbed or repaired in the past. Seller was stating that a Test 05 Fail 40 error was coming up upon power up, and that the traces where showing a lot of vertical jitter and eventually would move off screen towards the bottom. I figured that price and shipping cost numbers where decent for a B version, with the front cover and storage pouch (<$250), given the description, which I figured could be related to the well known leaky caps problem, so went for it and won. Oh and somehow the blue filter was loose and almost fell out when I took off the front cover. Have to check the manual, but I figure the filter just needs to snapped back into the groves that seem to hold it in place? If it is so easy to fall off, no wonder so many scopes are being offered without the blue filter, gone missing.
But today upon arrival when I powered up the scope it does not show a trace at all, or even a readout, nothing at all on screen, so not sure how long ago it still had a live screen. The power up self test does stop at the "ADD" label, so as I understand this seems to match the reported Test 05 scenario error. In any case pulled out the A5 board and got started on the leaky cap mess cleanup following some very good suggestions from Chuck not to try to desolder the caps in one piece, but rather carefully chop them up to get to the pads. Well, I kind of cheated and very carefully lifted the single cap in the corner by desoldering it bit by bit by extremely gently rocking it side by side very little at a time while heating the (cleaned up) pads, as it looked easy to do it on that location. It came out fine. But then decided to heed the recommendations as the other remaining 3 caps looked more difficult to get at due to less space around them. I think in the end all went quite well, there was not too much corrosion damage on the pads, and I could not see any traces that where eaten away. Off course still quite a few other affected parts with dull/greenish pads that will need to be cleaned or replaced (see pictures liked bellow), and I am guessing that in the case of the IC legs that show corrosion they will have to be cleaned thoroughly. Also once I am done removing the rest of the affected SMD resistors and caps, will wash both the affected areas of the board with dish soap and water with a soft toothbrush. Then its into the oven at 375 deg for 45 minutes. Just kidding... its really only 40 minutes that are needed. Some A5 pictures here: /g/TekScopes/album?id=76874 BTW the scope has a TEK U800 (not the less desirable Maxim part), and most ICs show a 1990 date code. The NVRam has a 9020 date code. Once I can sort out the most pressing issues with the inoperative display I'm thinking of going the FM16W08 route, to do away with the need of a backup battery for good. I only hope getting the A5 board back in shape will bring back the screen to life. If not once I'm done with it will start checking voltage rails at J119 as well. Comments and suggestions very welcomed about any follow up tests to make sure I have possibly not missed anything on the A5 cleanup, any know suspect VIAs to check, or hidden traces to verify they have continuity to another part, etc. Any other obvious possibilities besides the A5 mess on why the screen is non-op? BTW, noticed that one of the 3 caps had two very fine traces between the pads to which the cap is mounted. I see these traces on one side go to what looks like two test points, but on the other side they seem to end nowhere, and don't seem to be connected to the capacitor pads either. Out of curiosity, anyone know what they are meant for? Thanks for any input. Alex |
|
Alex,
Very nice cleanup job. I'd just like to jump in here about removing SMD caps with minimal chance of pad damage. Every technician has his own favorite method of removing SMD capacitors. My favorite method is to grip the capacitor from above with a pair of pliers, push DOWN, and simultaneously firmly twist the capacitor in the horizontal plane, about 120-180¡ã. NEVER pull the capacitor in an upwards direction. The correct horizontal twisting action will break the capacitor leads at their weak point, and the capacitor will separate from the PCB. If this method is done correctly, it will never cause damage to the solder pads on the PCB. |
|
Alex,
There is a vast amount of information around about the A5 corrosion problem. The corrosion usually causes two 10k precision resistors to change their value. These set the voltage reference for the DAC that controls most of the operating points for the scope, including trigger levels (that causes the Test 5 failre) and also the brightness levels of trace and readout. There is a high probability that the screen will be OK after you replace the damaged components. Before I did my A5 repair I also had nothing visible on screen, I adjusted the Grid Bias pot before repairing the A5 board as a confidece measure to see that the rest of the scope was OK. Roger |
|
On Fri, 26 Oct 2018 at 01:13 <tekscopegroup@...> wrote:
Comments and suggestions very welcomed about any follow up tests to makeA5 problems are a sufficient condition for no screen. On my 2467 a clamping zener on the -1.25V reference had gone short, causing no display at all. You've probably read up on this already, but the resistor network that sets the DACs reference current is your prime suspect for the no display condition. In your shoes, I'd buzz out all the traces and the resistors involved there - this is everything connecting to pins 14 and 15 on the DAC. Note that if you need to replace R2012/R2013 that they're precision, low tempco resistors, so ideally you'd want to replace them with like - especially if you intend to calibrate the scope. The little trimpot is a frequent casualty of capacitor leakage, so give that a good look. I wouldn't worry about jiggling it or replacing it, you'll have to go through the DAC calibration procedure either way. It's a simple enough procedure, and I think it's safe to assume that bringing the DAC into spec will bring you nearer calibration than wrecking it. There was a recent thread here discussing the +1.36, -1.25 reference voltages and how they derive directly from the DACs reference current. I don't think on your A5 board they will be brought to a test point, but if you hunt them down, they allow you to assess the wellness of the DACs reference, which is otherwise difficult to measure. |
|
Gentleman,
Thank very much to all of you for all the information provided. I managed to get the scope screen back to life by replacing the two 10.0K 0.1% resistors (R2012 and R2013) in the DAC resistor reference voltage chain. Also replaced R2016 1.0K as it was pulled out by mistake but tested good. Both 10K resistors where open, and I had to use 5% 1206 parts I had on hand just to test if I was going to be successful to fix the DAC volatges, altough the resistors I used are nowhere near the required specs but close enough for the scope to boot without any errors. Everything done by hand soldering with a T7 fine tip and a Weller 35W pencil iron dialed down to reduce the tip temperature. I do quite a bit of 1206 and SOT-23 parts installations, so the experience gave me the confidence to proceed here as well. In the end those resistors where practically trying to jump off the board with the iron by just reflowing them with fresh solder on both ends. So all in all they where very easy to remove, and zero damage to any pads. I did not attempt to correct the DAC reference voltages yet, no point doing so until the proper parts arrive. They where at +1.3848v and -1.2688v, not even near on-spec. Posted a couple of updated pictures on the photo album: /g/TekScopes/album?id=76874 Checked a bit the scope accuracy and it was not too bad at all, given the ref voltage errors, but I am not sure this would also affect the accuracy of the cursors as well? On the new picture you will see a 500nS/Div sweep speed but cursors set at two divisions read 973nS. Of course I now realize that I should have spread the cursors out further, in which case the error seemed to get numerically much smaller. Also ordered a new DS1225AD-150 and a 3M socket, although at some point eventually would very much like to go the FM16W08 way. My Mouser parts order went out earlier today to get these two resistors and also other parts involved in the general exposed area to the leaking caps, just to have them handy. Of course I also ordered the replacement for the 4 filter caps. I chose to order SMD Aluminum Organic Polymer Electrolytic Capacitors which look very much like the original ones. Also ordered the two power line safety caps, in case mine are of the "fishy smell" type. I have not yet checked the PS board to see what type I got. Recommended subs where to get Kemet Polyester 0.1uF 300V Safety Capacitors. Will also check ripple on all the main DC power rails at J119. Scope will probably need some help with the filter caps, as traces seem a bit fuzzy at times, although they look very much sharper when the 20MHz limit is enabled. Just in case last night I also copied the EXER 02 values by taking a video while scrolling through all the values 00-FF, and even took them down the old fashion way on paper and pencil. EXER 05 indicates this scope has only seen 360 hours of On time, and 125 power cycles (those last 25 are probably mine). From the writings and labels on the front cover and one side of the scope it looks like this was a designated field instrument that probably did not go out that much. The insides are pristine, and cosmetically the outside and front panel controls are in very good shape. Maybe the A/B timebase Sec/Div concentric knobs feel a bit wiggly, maybe a nut or something will have to be tightened inside? I am also glad the so very important front panel cover was included, and the always good to have pouch. No probes unfortunately. Once the parts order arrives, and I get a chance to throw them on the board, will update this thread again. Any comments/suggestions always appreciated. Alex |
|
Chuck Harris
For what it's worth, the power cycle counter, and on time counter
toggle quoted message
Show quoted text
are meant to be reset after every calibration. They are there to let the technician know how long it has really been since the last calibration... as opposed to whatever the customer says.. They don't mean much to a customer. If you see really low numbers, it simply means the scope saw little use since its last calibration. -Chuck Harris tekscopegroup@... wrote: ... EXER 05 indicates this scope has only seen 360 hours of On time, and 125 power cycles (those last 25 are probably mine). From the writings and labels on the front cover and one side of the scope it looks like this was a designated field instrument that probably did not go out that much. The insides are pristine, and cosmetically the outside and front panel controls are in very good shape. Maybe the A/B timebase Sec/Div concentric knobs feel a bit wiggly, maybe a nut or something will have to be tightened inside? I am also glad the so very important front panel cover was included, and the always good to have pouch. No probes unfortunately.
|
|
On Tue, Oct 30, 2018 at 01:11 PM, Chuck Harris wrote:
Hi Chuck, this certainly makes a lot of sense. Thanks for clarifying it, never heard this simple and logical explanation from anyone else, other that people being very happy when they see low numbers on those counters on a newly acquired scope. This one does not have any past cal stickers at all, so wonder if it ever was calibrated after being purchased. Since I am new to this model, I noticed that when pressing any front panel switch, all the other indicators seem to very briefly flash in unison before the requested change is executed. Is this normal? Also, under certain conditions, specially when using the x10 horizontal magnifier and on the lower vertical v/div ranges (10mV and bellow), the traces seem somewhat blurry. But without changing the v/div range they clear up nicely and look very sharp when I engage the 20MHz BW limit, tough. Would this possibly mean HF power supply noise or power line ripple creeping into the circuits? (aging filter caps that need to be replaced?). I have not yet gotten to checking all the power rails at J119, will do so when the Mouser parts order arrives and the cover needs to come off again. The other issue I've seen is that the automatic frequency measurement (press the measure button and then select 1 for frequency) is off usually by at least 10% or so, sometimes even more. Accuracy of the internal 10MHz reference oscillator not really being that accurate? If so wish there was a way to retrofit an external 10MHz ref input, which I understand is one of the standard options. I read somewhere that it is fairly easy to do it, but found no detailed information. As a comparison, my 2247A will always read the frequency very accurately, and its off by just 71Hz tested against a freshly calibrated rubidium oscillator. Lastly, I figure you're probably the best person to ask this. On the installed options sticker on the back, the only field marked (punched out) is 46 (between the 22 and 1E fields) on this rear aluminum sticker. So far I have not been able to find any reference at all for an Option 46 on these scopes, and nobody that I asked so far seems to know. Are you by any chance familiarized with what this option 46 entails? Thanks much. -Alex |
|
On 31/10/18 21:48, tekscopegroup@... wrote:
On Tue, Oct 30, 2018 at 01:11 PM, Chuck Harris wrote:Yes. IIRC all the indicators are driven by a shift register; one changes and all stages of the shift register have to be loaded serially.For what it's worth, the power cycle counter, and on time counterHi Chuck, this certainly makes a lot of sense. Thanks for clarifying it, never heard this simple and logical explanation from anyone else, other that people being very happy when they see low numbers on those counters on a newly acquired scope. This one does not have any past cal stickers at all, so wonder if it ever was calibrated after being purchased. Also, under certain conditions, specially when using the x10 horizontal magnifier and on the lower vertical v/div ranges (10mV and bellow), the traces seem somewhat blurry. But without changing the v/div range they clear up nicely and look very sharp when I engage the 20MHz BW limit, tough. Would this possibly mean HF power supply noise or power line ripple creeping into the circuits? (aging filter caps that need to be replaced?). I have not yet gotten to checking all the power rails at J119, will do so when the Mouser parts order arrives and the cover needs to come off again.Measure the ripple and replace the capacitors where beneficial. Replace them one at a time; in some manuals two caps are swapped compared to the actual boards. That is well documented elsewhere. Don't forget that there is a lot more thermal noise in a 350MHz bandwidth than a 20MHz bandwidth - you may simply be seeing that thermal noise. The other issue I've seen is that the automatic frequency measurement (press the measure button and then select 1 for frequency) is off usually by at least 10% or so, sometimes even more. Accuracy of the internal 10MHz reference oscillator not really being that accurate? If so wish there was a way to retrofit an external 10MHz ref input, which I understand is one of the standard options. I read somewhere that it is fairly easy to do it, but found no detailed information. As a comparison, my 2247A will always read the frequency very accurately, and its off by just 71Hz tested against a freshly calibrated rubidium oscillator.Highly unlikely the 10MHz is incorrect - but you can always measure it on your other scope :) What's more likely is that the A5 voltages and/or calibration values are invalid. Before tweaking anything, make sure the reference and DAC output voltages are as specified in the service manual. |
|
Chuck Harris
Hi Alex,
toggle quoted message
Show quoted text
The design of the front panel is such that all of the displays states are in registers that get strobed whenever anything is changed on the front panel. That shows up as a brief flash of all lights. The 2465B uses 74F logic for those registers, the older 2465 used 74LS logic, and fairly dim LED's. The blink is scarcely noticeable with the older 2465. The 2465B uses much brighter LED's, and the blink is quite apparent. With 350MHz bandwidth comes a lot of ambient noise. Cut the bandwidth, and you will cut the noise. That is why the scope has the 20MHz bandwidth limit switch on the front panel. Working amounts of power supply ripple are not generally noticeable. Large amounts, usually cause erratic operation. An example of erratic operation would be trigger jitter. The 2465B is not a frequency counter, and the A5 board's 10MHz reference has nothing to do with sweep or frequency accuracy. The way the 2465 family creates its sweep is by using a miller integrator, and precision integration capacitors. The way it measures frequency is rather convoluted, but depends entirely on the sweep accuracy. Think of it this way, the sweep is generated by charging a capacitor with a constant current. It is correlated to both the graticules, and the cursors during calibration. The frequency measuring method uses the trigger hybrid and a DAC to measure the zero crossing points of the waveform, and uses the delay sweep to measure the time between the zero crossing points. The CPU then performs the inversion, and the display shows that number as the frequency. It is not a frequency counter! It is standard practice of most sellers to remove all calibration stickers, and stickers showing previous ownership before the sale. Although I don't sell a lot of scopes, when I do, I blow out the dust from the inside, clean the outside, and the CRT, make any repairs, and calibrate the scope, complete with zeroing the counters. It would look as you are describing your scope, regardless of how old it really is. If I cannot repair it, and clean it up to that degree, it becomes a parts source. According to Manuel Maseda, Option 46 is the military version of the scope (OS-288/G) that includes option 10 (IEEE-488), and 2 probes. -Chuck tekscopegroup@... wrote: On Tue, Oct 30, 2018 at 01:11 PM, Chuck Harris wrote:Hi Chuck, this certainly makes a lot of sense. Thanks for clarifying it, never heard this simple and logical explanation from anyone else, other that people being very happy when they see low numbers on those counters on a newly acquired scope. This one does not have any past cal stickers at all, so wonder if it ever was calibrated after being purchased. |
|
Chuck as usual your explanations are outstanding. After reading them, just a few short extra questions/comments, bellow.
On Wed, Oct 31, 2018 at 09:49 PM, Chuck Harris wrote: The design of the front panel is such that all ofI assumed that it was normal, but I was surprised to see the effect as it was never evident on any of the videos I'd seen before. With 350MHz bandwidth comes a lot of ambient noise. CutSo it is normal to see some fuzziness on the traces at full bandwidth? I really have to go through this again, as I seem to remember the amount of fuzziness varied somewhat depending on the time/div setting, not sure that would be normal. I am still awaiting for the Mouser parts for the A5 board to arrive, and will confirm that once I have to put the scope back on the bench to work on replacing those parts. The 2465B is not a frequency counter, and the A5 board's...snip... Very nice explanation of the principle used to make the measurements, now I have much better understanding of the process. It is standard practice of most sellers to remove allIf I may ask, how do those hour and power cycle counters get zeroed out? I don't think I have seen the reset procedure explained in the service manual. According to Manuel Maseda, Option 46 is the militaryIs there a more specific service manual for this variant I should look for? If you happen to know, are there any differences in the military version from the regular version? Perhaps a sturdier build? or what where the military requirements? Thanks again Chuck. Alex |
|
Hi Tom, thanks very much for your input. Bellow just a quick follow up.
On Wed, Oct 31, 2018 at 04:08 PM, Tom Gardner wrote: Highly unlikely the 10MHz is incorrect - but you can always measure it on yourIs there a test point designated on A5 to check on the 10MHz oscillator? Just for fun and while I am at A5, might as well check that, although after reading Chuck's detailed explanation on how the scope actually performs the frequency measurements, the 10MHz is not used in the process, and I suspect that it will be within spec. What's more likely is that the A5 voltages and/or calibration values areYes, at the time of making those measurements both the positive and negative voltage reference values where/are still off. Did not want to touch the adjustment until the two 10.0K 0.1% resistors that where damaged by the leaking caps are replaced with the correct parts. They should arrive sometime this week, and once they are in, I will tweak R2010 to set the reference voltages to +1.36v and -1.25v. Thanks again. Alex |
|
On Thu, Nov 1, 2018 at 01:30 AM, Manuel Maseda wrote:
Hi Manuel, Unfortunately my scope does not include the CT option. Also owning a 2247A, and routinely using the build-in counter timer, the CT variant was something that was desirable and high on my 2465 wish list, but ultimately when the right scope came along it was a straight B version only and I did not have much of a choice anymore. The only option included is 49, which as you and Chuck have pointed out means it is a military version. But as Chuck also very eloquently explained, the accuracy of the 10MHz internal oscillator does not play any role in the process of measuring frequency of a waveform (at least not without a CT option), so adding an external ref input is now a moot point. Thanks for the input. Alex |
|
On 2018-11-06 7:36 a.m., tekscopegroup@... wrote:
...Yes, I think Chuck (or someone else) pointed this out earlier in the thread. The wider bandwidth the more noise in the trace, this is normal on any scope afaik. --Toby ...Alex |
|
On 11/6/2018 4:05 AM, toby@... wrote:
On 2018-11-06 7:36 a.m., tekscopegroup@... wrote:Particularly if you live in an area with nearby FM broadcast transmitters, the average scope probe will pick those up quite well :^) dr |
|
Chuck Harris
Hey! Don't muck with the DAC adjustment, unless you are prepared
toggle quoted message
Show quoted text
to do a full calibration! It is not supposed to be +1.25 and -1.25 volts, but rather a 2.500V span. Sure, if you are a perfectionist, and diddle with the 10K reference resistor values, and put in perfect opamps, it can achieve those values, but that is not a necessary goal: 2.500V span when doing the calibration routine is the necessary goal. -1.25 and +1.36V references that are *stable* is the goal, as any errors in them get stored in the calibration constants, and as such calibrated out, when the scope is calibrated. Any tweaking, or replacement of DAC parts that you do will require a full calibration of your scope. Why? Because you have no idea what they were like when it was last calibrated. If they were on the low (or high) side of spec when last calibrated, and you adjust them to be right on, with a DVM that was on the high (low) side of spec), they errors could add up to be wrong enough to perform poorly. -Chuck Harris tekscopegroup@... wrote: Hi Tom, thanks very much for your input. Bellow just a quick follow up. |
|
Chuck Harris
Noise is endemic in electronics. It is related to the
toggle quoted message
Show quoted text
temperature. The resistors, transistors, etc.. All make a contribution to the noise. The noise of a receiver is set by the noise of the front end of the receiver. Once established, any further amplification simply amplifies the noise... the signal to noise ratio stays basically the same.... well, until overload. When you have the scope in the higher gain positions, the noise is being amplified, and is displayed as more fuzz. Noise of this sort is called white noise. Its distribution is fairly uniform across the frequency band of the scope. If you run the scope at its full 350MHz bandwidth, it is slurping in all of the front end's noise from DC to 350MHz, and adding it to the trace, everywhere... If you reduce the bandwidth of the scope, to say: 20MHz, there is less overall noise to add into the trace, because there is less overall bandwidth... the noise from 20MHz through 350MHz is mostly blocked. About the blink: Sorry, it has always been there with the 2465A and B models. It is not as apparent with the plain model, mostly because the LED's are dimmer. Wnen the LED's registers are loaded with the state of the LED's on the panel, they all go into an LED ON state for the duration of the load. It is a feature of the register used. CAL 05 is the routine that can read/set/reset the counters. Its default position is to reset the counters. You can set them to whatever you want... the technique requires you to select the increment using the X10 button, and to rotate the Delta Ref, and the Delta controls to make the appropriate counter change... I know of nobody that does this. There is no specific manual for this option, but rather there is the main instrument manual, and the GPIB options manual. GPIB on the 2465B is a complete waste for the usual user. It only allows you to set the front panel controls, and display some text on the screen. It does *not* allow you to read anything that is derived from the CH1-4 inputs. It's purpose is for use with the rack panel version of the scope, when using the scope to do mindless calibration and repair routines of military equipment. Basically, it replaces the calibration manual with a computer as the source of calibration information. -Chuck Harris tekscopegroup@... wrote: Chuck as usual your explanations are outstanding. After reading them, just a few short extra questions/comments, bellow. |
to navigate to use esc to dismiss