#30 - Keithley 614 electrometer repair and calibration

hello repair time again this time I have a vintage Keithley 6:1 for electrometer I picked it up on eBay as non-working for parts and it's quite old from mid 80s I believe but it still can be a very useful instrument because it can measure current down to Pico amps and the resistance up to hundreds of Giga ohms it can also measure voltage and charge and the range of voltages is not very wide from 0.2 volts to 20 volts but the input impedance is remarkable more than five theorems according to specs and because it is so sensitive a triaxial connector is used for the input and these are rare and expensive and unfortunately this connector is an old one with two locking lugs on it newer connectors were changed intentionally so that regular B&C connectors won't fit I happen to have this cable by Keithley you see on this sticker but it has a newer connector with three working lugs so it's not compatible with this instrument so I will have to do something about this connector get some adapter or maybe replace the connector on the front panel or the easiest way would be perhaps to remove one log so that any connector would fit I don't quite like the destructive approach but it is certainly the cheapest and easiest but let's not worry about this for now let's see if we can manage to repair the instrument first let's have a look at the back side for a moment just for completeness it's made in the USA and there is a line voltage switch here a fuse a standard connector and some sort of analog output and preamp output it was described as not powering up but this description was not accurate it does power up but it shows some random numbers in the display were just flashing zero sometimes which means over range another problem with it is sticky switches on the front panel for example the function switch when I press one button another one should pop out and sometimes it happens sometimes it doesn't like this current button gets stuck and I need to help it pull it out and I think they are getting better when I used them a few times and this which was also sticky but it is also getting better I believe this one was fine so they definitely need some cleaning but that should be easy have a look at this for example it's in the voltage mode and there is some random number on the display and the range switch doesn't change anything at all and the zero check button which shorts the input also doesn't affect anything so something is definitely wrong here fortunately I have a service manual with all schematic diagrams let's go just two screws on the back and we are in very easy and look at this thing I think it's an optional battery and these were led acid batteries and it must be completely dead a long time ago but not to worry it doesn't affect anything the unit should work just fine without it I looked around this unit and I don't see anything wrong so far but I do see some test points here marked minus 5 plus 5 minus 32 and plus 32 so that's a good starting point and measure those voltages let's do it this is the ground minus 5 no problem at all plus 5 fine plus 32 slightly lower but should be okay - 32 about the same should be just fine no problem the most sensitive part of the instrument must be here under the shield so let's take it off and have a look I think we can see some interesting design decisions and I need to be super careful not to touch sensitive parts of the circuit even smallest contamination from my fingers can create problems there look at this the most sensitive input part is mounted on special Teflon standoffs this is the schematic for the main unlock part of the unit and some voltages are marked here on the schematic so I suggest to start measuring them like this - 32 and corresponding + 32 here - six point four plus six point four and on this op-amp - five point eight plus five point eight here so let's go this op-amp is here on the board and these two transistors are these guys so let's start measuring voltages around them oh look how convenient we should see all three marked voltages around each transistor - 32 volts on the collector - six point four on the base and minus five point eight on the meter and the same thing on the positive side the ground is marked here on the board let's use this point and this transistor has collector here 29.5 problem the base should be at six point four whoops and it is at 28 something is wrong and the meter should be at five point eight and it is a twenty seven point four and the next one collector is minus twenty nine point eight no problem and the base should be minus 6.4 and it is plus 15 something is wrong and emitter 15 again it should be minus five point eight so something is wrong around these transistors and look at this interesting arrangement as if two parts are missing here I looked in the service manual carefully and I found that these are jumpers which can be removed during the calibration or can be left in place depending on particular measurements so this is not a problem at all let's check the voltages across this Zener diodes we should see six point 4 volts on each side on each Zener diode and here they are on the board let's measure 16.4 no problem at all and another one six point four so using our diodes are fine and let's check the transistors I turned off the unit and the emitter is in the diode checking mode this device on the negative side is PNP device so we should have negative on the base and we should observe two diodes collector problem emitter problem and the other device is NPN so we should put positive on the base and we should see two diodes problem at all so this looks fine on the first glance let's check voltages across these two things see are 101 and CR 102 these marks on them mean that they are current regulators and the purpose must be to have stable current through Zener diodes and we should see about 25 volts or so across them we have around 30 volts at this point instead of 32 and we have six point four across Zener diodes so let's check this here they are on the board look like transistors but they have two pins let's see across this one we have 1.5 volts and across the other one we have 45 volts and here they are by the way in the parts list just called diode J FET and they are hard to test in circuit but should be trivial to test out of circuit so I am tempted to just solder them and test with just power supply pass some small current through them and see what happens I removed the board so I can this older than your current regulators and by the way I looked at the date codes on the chips and other parts and they range from 84 to 86 so this thing was made around 1986 and have a look how this zero check button works how this gold-plated spring touches this gold-plated pin and here is the other side here I'd assaulted one of them from the negative side and the marking on it is CR 120 and after a lot of searching I found this data sheet from international semiconductor and here it is C R 120 nominal current is 1 point 2 milliamps here is my test set up precision HP lab supply I'm going to pass current through this device and the test and I set the voltage to 25 volts limited the current to 10 milliamps just in case and resolution here is in hundreds of my cramps but it's not accurate enough at the low end so we need this trusty fluke 87 to measure the current let's go 1.2 million no problem at all and here is the second one from the positive side 1.2 million this device looks fine as well it occurred to me to check the switches and I found more problems with them let's have a look so these short switches have two groups of three pins and these longer ones have four groups of three pins and in each group the middle pin can be connected either to this one when the button is not depressed or to this one when the button is depressed so in this case we should see open between these pins and short between these and we do but this other group we see open here and open here and if I wiggle the button so there you go and I found a few switches like this so we definitely need to fix this first so can this problem with switches lead to this other problem with voltages here I'm not entirely sure it's possible if we look at the schematic there are these switches all over the place and if some of them are open who knows what might happen I didn't wanna lies the schematic deep enough and which switches are where to reach any conclusion I think we need to fix the switches first before we proceed I really hope that this contact cleaner should help us to fix the problem and the way to use it is to apply small amount right into the switch like so and then operate the switch before the skin achill evaporates I'll do the same to all the switches and check with the multimeter again the contact cleaner worked really well I checked all the switches in both positions after the cleaning and they work perfectly and is the last step I applied tiny amount this cleaner with lubricant the lubricant is not conductive it's used in the industry for decades and has a lot of research behind it very good stuff I highly recommend it now let's check if it affected the other problem we had with voltages we don't really though I display for this I connected the power here and by the way according to the service man al to check these voltages we need to select voltage function here and 20 volt range and I believe I have done it the same way before so let's turn this thing on so here is the ground let's check collector 29 volts B is 6 point 4 as it should a meter five point seven nine no problem at all the other one - 29 - six point four and minus five seven nine no problem at all I put the board back into the case for convenience but I didn't put the shielding back on top of that sensitive input circuitry yet and now the readings on the display look much more sensible to me at least on the first glance with zero check depressed we have zeros on all the functions and without the zero check it picks up some noise which it should without the shielding so let's try something let's try measuring the stenwick resistor for example we are reading ten point two six on the fluke so the wires from the input tray axial connector are still disconnected here and I attach these grabbers to the same pins they were connected to and here is our 10 Meg resistor reading 10.2 no problem at all here I'm using Agilent precision power supply and the voltage is set to 10 volts no problem at all let's set 1 volt 1 volt let's switch ranges here no problem at all let's go to point 1 volt no problem at all let's go to this range works fine now I am trying to measure current and I'm afraid I don't have a proper current source for the job this thing can go maximum up to 2000 micro amps which is 2 milli amps and I think it's about the noise floor for this supply I set it to limit the current at 1 milliamp at the moment and we are reading 1.2 1.2 7 to be precise which looks sensible close but I'm not sure how accurate it is let's try changing this current here let's go to 2 milliamps okay the range makes sense let's go to 1 and again it's reading a bit higher and about point five higher than before which makes sense but I'm not sure if it is off because of the supply or because of the meter and let's make sure because we can this six-and-a-half digit hewlett packard meter agrees with Keithley one point seven eight and let's go back to one milliamp again it agrees with Keithley one point two seven so the meter was spot on I am quite happy about that I think it's time to take care of our input connector I managed to pick up this free lock panel mount tri-axial connector in a local electronics store for just seven dollars fifty I think it's a very good price so I removed this to lock connector from the chassis and the ground is still attached here we are going to reuse it here I have a piece of thin quarks with Teflon and solution which should be great for this application and here I have these crimp pins which are very similar in size to these and I've tried they fit over the corresponding pins and the instrument just fine the trick is not to contaminate the surface of the Teflon between the central conductor and the outer shield I'm going to make a fresh cut and wear gloves here we are a fresh cut and the shield is out of the way all right the center pin is soldered and I put a small piece of heat shrink over it now the question is how do we seal this area properly it seems to me that here some additional shielding is used what I am going to do is to solder this piece of copper foil and wrap it around maybe a couple of times and then I put a heat shrink over it here is a step towards the goal and now I just need to wrap the foil around cut the excess and put a piece of heat shrink over this whole thing right almost done just need to crimp these pins at the ends and remember this and is in the shielded area of the instrument so no additional shielding is required here crimps and I wanted to show you before I put heat shrink on the last one all right looks as good as Keathley ready to go hopefully I didn't contaminate it too much whoops I didn't notice so far that the new connector has slightly bigger diameter here so it seems to me that I need to enlarge the mounting hole in the front panel a tiny bit and that's unfortunate because I wanted to have an easy option to go back and put the original connector if I really wanted to for some reason and after enlargement I think I still will be able to with some water and such nothing is perfect and before I enlarge the hole which is irreversible I would like to test the performance a little more so I put the shield back and they connected the ground here with this alligator clip and the top cover doesn't have any shooting so it doesn't make any difference at all so let's try something when I was shopping for the triaxial connector I also picked up a couple of 500 mega resistors 1 percent tolerance let's try as you can see fluke 87 cannot measure 500 Meg's in the regular resistance mode but there is a way to measure even more than this but we need to switch it to conductance which is done by changing range here to knowledge Siemens and here we are almost to nano Simmons - naina Simmons would be exactly 500 Meg's but according to the manual it's only accurate if we zero out the meter first which is done by disconnecting the resistor and then we press relative and now we should be good there we go exactly to nano Siemens which corresponds to 500 Meg's and here is our key limiter spot-on well at least within 1% for sure and the device and the test is not properly shielded but it just have a giggle not hundreds of Giga ohms and here I connecting both resistors so we should have a Giga ohm and we have it slightly higher about 1.8 percent higher but according to the specs accuracy in this range is too cent so we are within the spec and here is an interesting experiment and the demonstration why these meters with very high input impedance are useful sometimes here I connected two resistors in series to form a voltage divider and I applied 20 volts from this precision power supply across one Giga ohm and I'm measuring voltage across one of the resistors so we should measure 10 volts and we are measuring exactly that well almost exactly this is not possible using meters even with the Giga ohm input impedance and here I found a precision current source of 500 nano amps let's have a look at the specs of the Agilent meter and here in resistance mode on high ranges the current source is specified at 500 nano amps so I manually selected the high range and here we have it 503 and let's look at the specs of this chief limiter here 2000 nano ampere inch accuracy half a percent plus one digit and half a percent above would be five oh two five plus one digit so we are within the spec well right next to the limit but not beyond the limit and the instrument is thirty years old so I am quite happy about this and let's not forget that the Agilent meter can be off a bit too I fitted the new connector to the front panel here is the dead battery for two volt LED as it sells I'm not going to worry about it at this time maybe I'll come back to that sometime in the future and make another video battery powered isolated instrument like this might be hindi let's make some adjustments I found the calibration procedure in the manual and it suggests to start from selecting volt point 2 volt range depress 0 check and then adjust the trim port through this hole so that display reads 0 plus minus 1 I made the adjustment and now we read exactly zero the next step is to have 19 volts from a voltage collaborator and I have one here which is five digit collaborator and I'm very fine using this six-and-a-half digit meter that it is exactly 19 now we are supposed to connect this to the chief limiter and select 20 volt range so I connected the output of the collaborator to the input of the chief limiter and we are supposed to connect another precision multimeter to the back of the Keef limiter to that analog output we've seen before which I've done for the rear terminals of this meter and I switched to the rear terminals and we are supposed to adjust a trim port inside the chief limiter so we read minus one point I've done that but it made the reading on the display here less accurate in fact I'm not sure what's going on but let's proceed now we are supposed to dial one point nine volts here and select two volt range and adjust another to import so we see minus 1.9 here this is also done aha and now we are supposed to adjust another trim port which makes this reading on the display agree with the external measurement well supposed to reverse the polarity and the calibrator here and make sure that the reading is the same plus/minus free counts and our reading is exactly the same now we are supposed to go back to positive here and dial point one nine volts and select point two volt range here and we are supposed to adjust one more report so that these readings are point one nine here we are no problem at all now we are supposed to go into the current mode and select 2000 peak amps and then measure voltage in a particular test point marked TP one I found it connected the meter to it and now we need to adjust one motoring port so that we read exactly one world I've done my best this should be good enough now we are at the last step of our calibration procedure and it's a bit tricky we are supposed to set the calibrator to 19 millivolts and then connected through very high precision 10 Meg resistor to the input of the Keithley in the current mode and measure 1909 ramps and here is the list of recommended equipment we are supposed to use this decade resistor by ESI model DB 62 and the accuracy is 0.03 percent I don't have such precision resistors around and that's why I didn't seriously consider this way of creating current sources for calibration before but now we have to do something now let's see what we can do by the way this company ESI or electro scientific industries was founded in 1944 and it is still around but it turns out that in 1990s they sold this line of precision resistance products to another company called the T gam or something like that I'm not sure how to pronounce it properly and in 2006 IET labs acquired the rights to this line of products and they still sell this DB 62 series of precision decay resistors with improved accuracy which is 0.01 percent and they also claim improved long-term stability now let's say in the ten Lagrange they claimed 25 ppm three-year stability which is amazing so what I think we can do is use my best meter in terms of measuring resistance which is this one hewlett-packard or Agilent six and a half digit meter and measure some 10 Meg resistor let's say even this carbon 5% resistor we used before and then adjust our target reading of 1909 amps so if we plugged in the actual resistance here we can recalculate what we should read Henry Keithley and let's see what the accuracy of the meter is in this 10 Meg range let's use one year a figure and it is 0.04 percent which is not too bad I was not quite happy about the idea of using this five percent carbon resistor so I ran out to a local electronics store and picked up five precision 1% 10 Meg resistors and I measured them already and it turns out that the closest one at least a queen to my meter is this one and it is about 0.05 percent out assuming the meter is perfect as you can see the last digit is jumping around a bit so let's assume it is 5 so we divide 19 millivolts by the ten point zero zero five five and we get 1899 if we round up so we really need to adjust by one count and another tricky part of course is the voltage calibrator because we are using the last two digits it's not going to be terribly accurate I connected it to the meter and look it's slightly higher but fortunately we have this adjustment here so it can adjust it alike so it's a bit touchy but here we are it's perfect and here we are the last ring port is adjusted which completes the calibration procedure and here it is measuring our 19 millivolt source 10 volt and the 509 amp current source in the Agilent meter here it is repaired and calibrated should be good as new if you liked this video give it thumbs up and don't forget to subscribe to the channel thank you for watching bye

Channel: FeedbackLoop

Views: 14,723

Rating: 4.9733777 out of 5

Keywords: electronics repair, test equipment repair, keithley, electrometer, high impedance, low leakage, diy electronics

Id: txH6EpqN5V0

Length: 40min 58sec (2458 seconds)

Published: Wed Sep 09 2015