Tube-era capacitor analyzers
Nearly all of the tube-era capacitor analyzers measure resistance and capacitance with a direct-reading bridge circuit. A current-limited AC voltage is used on the bridge. The eye tube indicator opens when the bridge is balanced with the correct resistance or capacitance selected by way of the main dial potentiometer. Accuracy is mainly dependent on the precision of the standard capacitors and resistors in the device. Measurement of the caps is not the best use of the cap analyzers, however, especially at the extremes of the ranges. Such measurement is more accurate with a modern digital meter. Even my inexpensive Harbor Freight digital multimeter ($20 0n sale, shown here) will easily outshine any of these tube era cap analyzers in cap measurement. However, except for expensive specialty meters such as Sencore Z meters, typical digital meters cannot test for leakage at rated-voltage.
Nearly all bridge capacitance meters use the same three values of capacitor for standards; those are 200 pF, 0.02 MFD, and 2 MFD. The 200 pF standard is usually a mica cap but could be a mica in parallel with a compression trimmer for accuracy or to account for internal wiring capacitance.
Since caps can be wired in parallel to add capacitance, some of my replacement standards can be two or three selected capacitors which add up to the standard with a goal tolerance of 1% or less. The big 2 MFD cap in these testers is usually easy to spot. It's either one big cap or sometimes a pair of 1 MFD caps in parallel. The 0.02 MFD and the 200 pF can be spotted wired one terminal each away from the 2 MFD cap on the range switch.
Older Knight and Heathkit cap testers
I consider the Sprague Tel-Ohmike series to be the Cadillacs for tube-era cap testers, largely because of the built-in meters and hand-picked standards. However, the less expensive cap testers, often sold in kit form when they were new, are perfectly acceptable for reforming caps or checking for leakage when used with a cheap digital meter to monitor voltage or leakage current.
Here's a Knight (Allied Radio) R/C tester model 83Y124. It's simple, basic and with replacement caps, especially the 0.02 and 2 MFD caps, and a couple of resistors, works reasonably well. Note its claimed accuracy of + or - 10%.
The Heathkit C-3 is a very similar circuit to that of the Knight, complete with the spring-loaded leakage control. The Heath substitutes a 1629 and uses a 1626 surplus 12 volt filament triode as rectifier. I have a basket-case C-3 but have not yet decided whether it is worthwhile repairing. I consider the earlier Heathkit C-1 to be more versatile for reforming caps than the C-3. Click on this page for my notes on restoring that Heathkit C-1.
Lafayette Radio was not about to be outdone. Combining two testers in one, the Lafayette model LC-4 combined essentially the same circuit as the Knight with a short / open cap tester circuit.
Enter the Eico cap testers, 950A
My favorite Eico cap tester is the blue-face Eico 950A, one of their first. It is still quite usable. Its voltage is totally variable by way of a potentiometer. Electrolytic leakage, if 2 mA or more, is visible on a simple neon bulb circuit like that of the Heathkit C-1. The bulb lighting at 2 mA is a useful indicator when reforming an electrolytic cap since the voltage control can be advanced in stages as the neon bulb first lights up and then goes off as reforming current drops below 2 mA. With a voltmeter connected as well as the cap, reforming progress can be readily viewed. In fact, a voltmeter is almost a requirement with the potentiometer voltage control since its voltage markings are not all that accurate.
I completely rebuilt a 950A, replacing the electrolytics and all other caps. It had arrived with a shorted cap and an open voltage control pot. Because of the stacked series caps in the power supply, one of which turned out to be shorted, the B+ voltage relative to the chassis went very high. I was slowly powering up the 950A with my variac when I noticed the eye tube getting quite bright, even with only 70 volts or so of line voltage from the Variac. I immediately shut power down to find out why.
A quick view of the power supply schematic, shared by the Knight, Lafayette, and most other cap testers, revealed why that would happen. The Eico uses a high voltage supply of about 700 volts with two electrolytics in series. That voltage is split into two parts with the chassis in between the two parts. One electrolytic has its negative terminal to chassis, the other has its positive connected to chassis. The load of the tuning eye tube and the 82K ohm resistor balance the voltages so that the eye tube (and the one cap) can function at a design B+ of about +200 volts relative to the chassis. The other cap sees B- at about -500 volts relative to chassis which is used for leakage tests and for reforming electrolytics. If the 82K resistor opens or goes high, the B+ on the tuning eye increases. If the cap seeing the most voltage shorts, as happened to my 950A, all that voltage is dumped on the remaining cap with the tuning eye seeing up to 500 volts or more which can destroy it. Luckily, the eye tube in my 950A was in good condition.
The 950A uses the eye-tube for the leakage test for mica-paper capacitors. I measured 8 uA to close the eye tube on that test.
Like the Knight and the Lafayette, the Eico 950B substitutes eye-closure for both leakage tests. It has a more substantial voltage pot than that of the 950A. It probably rates as the all-time best selling tube-era cap tester made from a kit. I see more of these at swap meets than most of the rest combined. They are usually priced right, rarely costing more than $10 in an as-found "afraid to plug it in" condition. They use the surplus 1629, so they are not usually harvested for the eye tube. At last count, I had accumulated 5 of these because I find it hard to resist repair-challenge bargains or devices with eye tubes.
I repaired 4 of the 5. The 5th had a defective power transformer making it an organ donor. Surprisingly, none of the 4 needed their electrolytics replaced. Each 100K ohm high-voltage-control pot was also intact. Cosmetics ranged from very good to just fair, but a thorough scrubbing on the bare aluminum face usually solved most of the cosmetics issues. The 200 pF micas were in good condition. I replaced the 0.02 standard in the sets. For fear of depleting my small stock of 1 MFD and 2 MFD caps, I did not automatically replace the big 2 MFD if I was able to get a partial opening of the eye tube at the proper setting on the two highest cap ranges which use the 2 MFD as a standard. Since the best use of one of these testers is for high voltage electrolytic cap reforming and leakage tests, that was my main target for accuracy as opposed to basic cap measurement.
The handy comparator function
Note that the Eico 950A and 950B each have an extra pair of terminals on the left side. Those are a very handy addition to a basic capacitor-resistor bridge. You can simply connect a standard capacitor, resistor or inductor to those terminals, connect an unknown to the other terminals and dial for a ratio between the two. That feature alone makes an Eico a desirable device. My very expensive Sencore LC-75 Z-meter (which I obtained used for relatively cheap) will test power supply chokes for inductance up to about 10 Henries. So what to do with chokes higher than 10 Henry? I use a known power supply choke of 5 or 10 Henries and compare it to an unknown. The Eico provides the ratio between the two, a handy feature indeed! The ratio test can also be used to determine turns ratio of laminated transformers such as audio and interstage transformers.
With two identical caps or resistors, the comparator function should be dead center at a ratio of 1. If the pointer does not show 1, the 10 K pot used for the main dial may be off a bit. Add a small resistance to one side of the pot or the other. A small variable resistor can be used. I just cheated with one of the 950B sets and moved the pointer just a bit to the center since the extremes of the dial are not that accurate anyway.
High voltage and leakage tests
Like the 950A, the Eico 950B uses a high voltage supply of about 720 volts with two electrolytics in series. If its 68K resistor has drifted high, the tuning eye will see voltage higher than 200 relative to chassis and the test voltage (the high B-) will see that equivalent amount subtracted from 520. You will be unlikely to find the correct voltages even with an exact 68K resistor since there are variations in the power draw of various tuning eye tubes. A tuning eye should not be subjected to more than about 250 volts. I tried to balance the voltages in one 950B and settled on a resistor of 62K to achieve a compromise to reduce the tuning eye voltage to about 220 volts. I was able to bring it down to 200 by using about a 51 K resistor but decided to stay within 10% of Eico's design. The voltages are, of course, partly dependent upon line voltage.
The Eico 950B has a 470K resistor in the circuit to measure leakage of "paper and mica" capacitors. I found that the 950B closes its eye tube at an average leakage current of about 15 uA (microamps) with variations ranging from 13 to 17 uA. The leakage figure is partly affected by variations among eye tubes. I measured the 470K resistor in each of the 950B. If that resistor had drifted high (typical), I left it in place because that would reduce the leakage current necessary to close the eye. I personally find a leakage allowance of 15uA for non-electrolytic caps to be too high. Later Heathkits (below) use the figure of 2uA maximum as a cutoff point for leaky paper and mica caps and provide an onboard control to set the eye tube closure at that point.
The maximum for electrolytic leakage is typically set at 2 mA in later Heathkits. The Eico 950B uses a 2.2K resistor for setting the eye tube closure for leakage of electrolytics. I found that the average leakage current to close the eye was about 3 mA. Again, if the 2.2K had drifted upward a bit, I left it in place since that would mean less current needed to close the eye.
Testing for leakage current
So how can you check the mica-paper leakage current needed to close the tuning eye of an Eico without using a microammeter? I used my cheap 1 megohm input digital meter (from Harbor Freight) and set it for DC voltage. The highest scale (1000 volts) is fine. I connected the meter across the two terminals in place of a cap and slowly turned up the voltage pot. When the eye just barely closed, I read the voltage. That's the uA leakage. The cheap meter read the microamps directly because it presents a constant 1 megohm load.
My higher quality 10 megohm digital meter was connected the same way. In that case, I simply divided the voltage by 10 to get the uA reading.
The electrolytic leakage point was determined by resetting the digital meter to milliamps (such as the 20 mA scale on the low-end Harbor Freight meter). I adjusted the voltage upward until the eye just closed and then read the meter making sure that the meter never exceeded 4 mA to avoid overloading the Eico's voltage control.
Don't know whether your digital meter is 1 meg input or 10 meg? Put the digital meter on its high DC range and use the cap tester to read its resistance on the bridge directly or use that external comparator input against a 1 meg or 10 meg resistor!
Caution on testing aluminum shell electrolytics with all of these capacitor checkers that use the split voltage
The positive connection to a capacitor under test is at or within a few volts of the chassis but the negative side of the test voltage can be hundreds of volts negative with respect to the chassis. That means when leakage testing or reforming those aluminum shell electrolytics, there can be several hundred volts between that aluminum shell and the cap tester chassis. . You can verify this with a voltmeter connected between the negative test terminal and the chassis. I use a plastic jar for insulation to contain the aluminum shell electrolytic to keep my fingers away from such a hazard.
Enter the 1960s Heathkits
The Heathkit IT-11 and its near-identical twin, the IT-28 are by far Heathkit's best design in a cap tester, a major step up from the early C-2 and C-3. The IT-11 was introduced in 1961. The IT-28 followed in 1968 and was reportedly available until 1977, a 16 year run for the overall design.
The Heathkit IT-11 and IT-28 sold for the same amount or a little more than the Eico 950B but the design is far superior. The Heathkits limit the bridge voltage to 12 volts AC versus the Eicos at 55 volts. Heath uses a third tube to amplify and rectify the bridge voltage. The Heath design also has an additional leakage setting for low voltage electrolytics of the transistor era. The Heath has three on-board pots to adjust the leakage settings for accuracy. The Heath uses divider resistors for accurate preset voltage as compared to a single pot with marked but often inaccurate voltage settings. The Heath has a lever to select for bridge or leakage but adds that all-important discharge setting between the two. That setting dumps the electrolytic charge into a 10K ohm 10 watt resistor as opposed to turning down the voltage pot and using its resistance to discharge the cap. The Heath uses a single high voltage with the chassis at B- ground thus avoiding the hazard described above that is inherent in the split voltage testers. Also, a shorted cap will not burn out the eye tube but may blow the fuse which the Heath has but the Eico does not. The Heath has the comparator connections like the Eico but also has terminals for connecting an outboard signal or function generator for more accurate readings of small caps and comparator-connected inductances.
Differences between the Heathkit IT-11 and IT-28
The IT-28 is a bit more than a cosmetic upgrade of the IT-11 (although I prefer the looks of the IT-11). The IT-28 upgrades the power cord to a three wire grounded version, has a spring-clamp on the tube socket for the 6AX4 rectifer, uses modern caps for standards, and has a power transformer with dual primary windings that can be rewired for 240 volts. However, it also has push-on knobs that cannot easily be re-adjusted as compared to the IT-11 with set-screw knobs. To reset the pointer just a bit on the IT-28 main dial, I found it easier to loosen the main pot and tighten it again rather than to change the pointer on the knob.
Repairs on the Heathkit IT-11 and IT-28
The first act on either of these after safety checks is a good application of deoxit to all the controls and switches and the 6BN8 tube socket. I did not immediately apply deoxit to the 6BN8 pins in the IT-11 and was greeted with a cold filament in that tube. The cold filament was due to poor pin to socket connections, a problem quickly solved with the deoxit cleaner.
The voltage settings for both were either spot-on or within 1 or 2%, a surprise since I had expected the resistors to have drifted high. The electrolytics were in good shape as well. Maximum voltage read 612 DC with my line voltage, about 2% high.
The 0.02 cap and the 2 MFD standard caps in the IT-11 were badly in need of replacement. I carefully matched a pair of tested 0.01 caps and a tested 2% 2 MFD cap. That brought the capacitance bridge to very good accuracy. However, the range that uses the 2 MFD cap showed only a very narrow opening of the eye tube. I was puzzled by that and tried another known good 2 MFD cap (a pair of 1 MFD in parallel) but no change. Then it dawned on me that the only item in that circuit that was not part of the other ranges was the power factor control. Sure enough, the power factor control was open at what should have been the grounded end. More deoxit did not cure it. I opened the control and found the problem. The center contact was not quite level. A quick adjustment with needle nose pliers cured it. The IT-28 did not need repairs other than the deoxit treatment.
Setting the leakage eye-close points
Both the IT-11 and the IT-28 were well off the mark for the leakage set points. The three on-board controls that adjust those points must be set in the proper order with the 2 mA setting for electrolytics first (adjustment pot is at the top) followed by the small electrolytic adjustment for 15 uA (pot is at the bottom) and finally the 2 uA setting for paper and mica caps (pot is in the center). Accuracy of the adjustments can be partly affected by changes in line voltage.
The Eico 950A and 950B are more versatile and useful than the early Knight, Lafayette and Heathkit offerings. I especially like the very simple 2 mA leakage circuit of the 950A using just two resistors and a neon bulb. The 950B has a more robust voltage control than the 950A. However, the later Heathkits are obviously more capable than the Eicos and a bit safer when testing aluminum shell electrolytics. I consider the Heathkits at an original price of $29.95 to rival even the Sprague Tel-Ohmikes in usefulness. The lack of a meter is easily overlooked once one is confident of the repeatable voltage settings and the accuracy of the adjustable leakage settings. The comparator feature is a real plus in either choice, but the ability to hook an external generator is unique to the Heathkits. I consider the Heathkit IT-11 and IT-28 to be best-buys in 1961-1977. They are still very usable if performing at their full capability. If you can find a functioning one near its original price, grab it.
The Eico and other testers can be modified with larger resistors or even an on-board control pot to make them more sensitive for leakage. Here is a link to Bob Putnak's useful web page where he has done exactly that.
The Eico 950B and Knight manuals are available on BAMA. The Heathkit IT-28 manual is still protected by copyright since it was written after 1964. The IT-11 manual, however, had a publication date that was obviously prior to 1964. Like most of its manuals, Heath never renewed the IT-11 manual copyright, a requirement for continued copyright protection for publications prior to 1964. The manual is therefore in the public domain. One possible source is mods.dk, a free membership site in Denmark
I have added a supporting webpage covering Heathkit's instructions for calibrating the IT-11 and IT-28.
Power Factor and ESR
Looking through the manuals of tube-era capacitor analyzers, you quickly realize that the term "power factor" is in fact a way of expressing ESR (equivalent series resistance) at a frequency of 60 Hz. A few ohms of equivalent series resistance for an electrolytic in a vacuum tube power supply circuit will likely not be a problem. The advent of switch mode power supplies (SMPS) with their filtering requirement of up to (and more than) 100 KHz ripple, has eclipsed the "power factor" term for small capacitors. An ESR meter is a useful device for all electrolytic caps and essential if you service switch mode power supplies. Use it instead of a tube-era cap analyzer for low voltage electrolytics on the secondary side of a SMPS.
ESR can also be checked with an oscilloscope, an audio or function generator, and a few resistors. The test is based on the fact that an electrolytic with low ESR is essentially a short circuit to a high frequency low voltage AC (for example, half a volt peak to peak) such as produced at 10 KHz or so by an audio or function generator.
Two more pages on capacitor testers
Part One Click here for the previous items on the bench: Sprague Tel-Ohmike Capacitor analyzer examples
and also Part Three - What about those "Open/ Short" capacitor testers? Knight, Sprague and Cornell-Dubilier
date 6-20-11, 7-1-11