This article will discuss repair and calibration of the Jackson 648-series tube tester. It will also discuss the procedure to eliminate the 1S5 tube from models 648 and 648A. High voltages are present, repairs should only be attempted by a qualified technician. Copyrighted, all rights reserved.

Introduction

The Jackson 648 was available in several versions, including 648, 648A, 648P, 648R, 648S, 648-1, 648-1T. The basic circuitry is fundamentally the same in all versions, with only minor changes (socket layout/configuration, adjustable grid leakage circuit in 648-1T).

I am a big fan of the Jackson 648. It is definitely one of my favorite tube testers. While it does not employ a mutual conductance test circuit, it does use an advanced emission circuit that applies separate element voltages to each tube element, and separate load circuits are also used. “These voltages and loads have been carefully selected for each tube to meet most ideally the normal operating condition of the tube.” (Source: 1950 Jackson catalog).

In my experience, the Jackson circuitry is very effective, and often will surprise you with accuracy that approaches Hickok’s coveted “mutual conductance” test method.

Example: testing a weak 6N6-MG (metal envelope) triode tube:

(as tested with calibrated Jackson 648A)

(above: the same 6N6 triode, as tested on the Good-Bad scale of my calibrated Hickok 6000 mutual conductance tube tester)

Here is another comparison:

(Jackson 648A testing a very weak triode section in a Mullard 12AX7)

(Calibrated Hickok 6000 testing the same triode section of the Mullard 12AX7, using the Hickok good-bad scale.)

As you see from the above examples, the Jackson 648 is a very effective tube tester.

Another great feature of the Jackson 648 series is the huge amount of tubes that can be tested.

With a modern chart, and possibly the addition of some sockets/adapters (depending upon your version of the 648), you can test just about every tube — everything from an early 01-A triode, to a Western Electric 300B, to a modern M2057 compactron.

This flexibility cannot be overemphasized, as many people do not want to own multiple tube testers.

Build quality is very high. Beefy transformers, metal/wood cabinet, solid push buttons, quality switches and potentiometers, and smooth meter action all speak to the quality of the components.

Physical size is somewhat large: 16.5 x 14.5 x 7 inches (42cm x 36.5cm x 17.5cm), weight 18-lbs (8.2Kg).

REPAIR AND CALIBRATION

Tools needed: calibrated Fluke digital meter (capable of volts, ohms, milliamps), Variable DC power supply, a Variac, Caig DeoxIT cleaning solution, various Test Sockets, jumper wires, 1-megohm resistor, another tube tester to check the 1S5 tube that is inside your 648.

To begin, I always remove the roll chart. This protects the chart from chemicals used during the cleaning process. Next, I carefully clean the entire tester, inside and out, and use my garage air compressor to blow dry. I observe whether any wires move from the high pressure of the compressed air (likely a broken connection), and follow that up by using needle-nose pliers to give a gentle pull on each connection.

Inspect line cord, and replace if not excellent.

Remove the internal 1S5 tube, clean its pins, and test this tube on another tube tester. Replace if weak or defective. Reinstall tube.

—-

[Update: May 24, 2008. See the end of this article for procedure to eliminate the 1S5 tube, and "upgrade" the older Jackson 648 and 648A models to the solid-state circuitry found in the newer models, such as 648R, 648S, etc.]

—-

Next, check all resistors for accuracy, and replace those out of tolerance. To check many of these resistors, you will need to remove one side of the resistor from the circuit to achieve an accurate reading. I also replace the 0.1uF capacitor as a matter of habit.

Next, you must clean the Plate potentiometer, which is probably the MOST COMMON source of inaccuracy and inconsistent readings. Remove the large round knob from the “Plate” potentiometer. Remove the nut affixing the potentiometer to the chassis. Once loose, look under the chassis and you will see an opening into the potentiometer, as shown in the next photo:

Into that opening, apply Caig DeoxIT (which is available in either drops or spray can). Turn the potentiometer clockwise-counterclockwise a number of times, then repeat once more. The DeoxIT will clean the potentiometer, which will then be extremely accurate. Reassemble.

Treat all remaining potentiometers, controls, and switches with DeoxIT (including the push button contacts, the Grid Leakage potentiometer, Line control, Shorts switch). Treat R15 and R5 with DeoxIT also.

Treat all tube sockets with DeoxIT. For the large-pin sockets (4-pin, etc…), I find that you must “scrub” the socket to truly remove all of the dirt and oxidation. While there are many ways to accomplish this, I find that a .177 caliber gun cleaning brush, with the DeoxIT, works excellent as a gentle abrasive (you may need to crimp the brush hairs a little).

The small pin sockets are just treated with DeoxIT, and worked gently with a clean tube. If the socket does not respond favorably, it must be replaced.

CALIBRATION

With power off, make sure that the panel meter is resting at exact zero, and adjust the setscrew if necessary. The following photograph shows the location of the components necessary to calibrate the 648A model; the controls will be similar for other versions of the 648, but may be located different:

Line Control Calibration. You will see a series of numbers on the Line Control, and these numbers correspond to the “mains” voltage (ie, your current line voltage). Setup your Variac for one of these numbers (such as 100), and plug the tester into the Variac. With the Line Control at 100, the meter should read perfectly centered. If not, disconnect power and adjust R15. You can check the accuracy at other settings, such as 120 volts, to confirm your calibration.

Plate Current Calibration. You will need a Variable DC power supply, and an accurate (calibrated) milliammeter such as an American-made Fluke. I have found that the modern mass-produced Chinese digital meters do not have the required accuracy to perform calibration work. They apparently are not factory calibrated, because it is not uncommon for a Chinese meter to read as far off as 27ma or even 30ma due to poor quality control, whereas a calibrated USA-made Fluke or analog USA Triplett meter will read accurately at 25mA. So, vintage (calibrated) USA meters are required.

Keep the tester power off for this entire test. The only item that you will apply power to, here, is the variable DC power supply. Connect the variable DC power supply, positive lead to the junction of R5 and R11, negative lead to the pushbutton X clip of switch S4 (this clip has either a gray/white wire, or black/white wire, attached to it.). You need to apply 25mA thru the circuit. The circuit resistance is approx 2160 ohms, therefore usings Ohm’s Law (V=IxR) you can calculate that V = (.025) x (2160) = 54 Volts. So, your DC power supply must be capable of more than 54 volts, to give you room to maneuver. Commercial variable DC power supplies capable of higher voltages are quite expensive, so I just made my own. Set Plate control to 10. Connect your Fluke meter in series with the circuit, push V button, and turn up the voltage on your dc power supply until your Fluke reads 25mA. Adjust R5 for exact mid scale reading.

SHORTS TEST

You will power-on the tester for this test. The Shorts Test information found in the Complete Jackson 648 Tube Tester Manual is completely wrong. Nothing about it is correct (otherwise, the book is very good). I contacted the publisher to explain the issue regarding the Shorts Test information, but I did not receive any reply.

I have created new charts to provide accurate checking of the Shorts Test functionality:

These tests are easiest to perform using a Test Socket and working from the front panel. You can quickly and easily move your jumper to each setting.

When you turn the Shorts Test switch, the Shorts light should illuminate in the positions noted in my Shorts Test chart above. If not, you have a continuity problem somewhere that you need to track down.

UNDERSTANDING HOW THE SHORTS TEST IS MADE, AND LIMITATION THEREOF

To further understand the Jackson 648 shorts test, you must realize that this tester analyzes each section (of a multi-section tube) independently, without reference to the other section(s). What does this mean? Each section is analyzed for shorts or leakage among only its own elements. For instance, in a dual-triode tube, the elements of Triode #1 (cathode, grid, plate, heater/filament) will be analyzed against each-other, without regard for the elements of Triode #2. Thus, if a tube has developed shorts across sections, the Jackson will not detect those shorts. The same limitation applies to cross-section leakage.

For example, I will simulate a cross-section short of cathodes in a 6SN7GT. (I could have chosen any elements, such as plate #1 to plate #2, grid #1 to cathode of section #2, plate #1 to grid of section #2, etc). A 6SN7GT tube has 2 triode sections, with pin 3 the cathode pin for Triode #1 and pin 6 the cathode pin for Triode #2. Using an octal test socket, I have jumped together pins 3 and 6 to simulate a cross-section cathode-to-cathode short.



As you see in the photos (click to enlarge), cross-section shorts will not be detected (shorts light not illuminated in either Triode #1 or Triode #2 setup). That said, cross-section shorts are very rare, and if the tube was previously operated with this type of short, it will often fail the main test anyway.

LEAKAGE TEST

You can check the accuracy of the leakage control by slightly changing your approach to the Shorts Test above. Instead of jumpering a solid wire between the pins as indicated on my Shorts Test chart, you will instead connect a 1-megohm resistor between the pins.

When turning the Shorts knob to those positions where my chart indicates “Short”, what you will see will depend upon the setting of the Leakage control. Starting at the 0.25 meg position, you should NOT see any glow in the Shorts light. As you slowly turn the Leakage control clockwise, you should start to see a very dim glow (which may be intermittent flashes) just past the 0.75 meg position,

and a dim glow should solidify at the 1 Meg position.

You can see the difference between how the bulb indicates Leakage (photo above), vs. a full short (next photo below):

As you continue to turn the Leakage control clockwise, the glow should become brighter. If your results are materially different, you need to recheck the resistors in this circuit.

VOLTAGE READINGS

Finally, you will need to perform a series of voltage tests, testing for proper voltages of Plate, Screen, and Grid. There are too many of these to list here, but they are explained in the “Complete Jackson” manual.

FINAL TOUCHES

Polish all screw heads, knobs, and buttons. This really makes your tester stand out. Clean the Plastic window for the roll chart so that visibility is excellent. If the window is cloudy, you can replace it with a new piece of thin plexiglass, which will really make the roll chart look excellent.

Reinstall roll chart. Install new equipment feet if the old ones are not excellent. You can also touch up the white lines on the knobs, and install socket-savers (especially in the high-traffic 9-pin-miniature socket, using an SS-9 socket saver).

Finally, we achieve a properly repaired, accurately calibrated, beautiful test instrument. One that you can be proud to own.

FINAL THOUGHTS

Pros: As discussed above, the Jackson 648 series is a Quality test instrument. While my tube tester collection consists of more than 30 unique models and quite of number of premium Hickok, Weston, and B&K models, I have always considered the Jackson 648-series as one of my top 5 favorites when all factors are considered.

1. High Build Quality
2. Capable of testing a large variety of tubes (from antique 01-A to modern M2057 compactron)
3. Respected test method
4. Effective shorts test (with minor exception that rare cross-section shorts are not detected)
5. Excellent (and adjustable on-the-fly) leakage test.
6. Intelligent calibration procedure allows consistent results across units.

Cons: No tester is perfect, so I offer a list of “cons” to consider. In my opinion, none of these cons are significant, although depending on your needs for a tube tester, these cons may be important to you.

1. Line control adjustment steps in 2.5V increments. This is less than ideal, because if your present line voltage is between settings (for example, your line voltage at this moment is 121), you must choose to set the line at either 120 or 122.5. This will cause a different tube test score (although not by a large variation) than when your line voltage can be set “perfectly” (such as 120 line voltage can be set to 120 Line on the tester). Hence, you will only achieve “perfect” results when the line voltage can be set identical to one of these 2.5V incremental settings. As a workaround, you could run the tester from a variac, and make any slight adjustments. Personally, only someone being very anal would fret about this issue and the (rather modest) test result variation, but it is worth mentioning.
2. The unit is not fused (possibly some versions of the 648 are, but some are not), and therefore the transformers are not protected from overload. This seems like a bizarre oversight, although easily corrected (you could add a 1A fuse if you feel the desire).
3. Not the best choice for someone who works primarily with sweep tubes. For example, most HAM rigs use compactron finals, so Ham radio guys would be better served with a tester that offers a series of quick test sockets for compactrons (such as many Sencore, Accurate Instrument, and B-K models).
4. Size and weight make it less than ideal for persons doing repairwork on-the-go, or persons who want a very small footprint on their workbench. Much smaller and lighter choices are available (Accurate Instruments models, Sencore Mighty Mites, LaFayette models, etc.)
5. Not all 648 models have the same socket configuration, which means that earlier models require socket adapters (or self-installation of new sockets) to test newer tube configurations, such as for novar and compactron tubes.

Food for Thought: Over the years, I have only met (in person) three guys who were truly world-class technicians. The real deal — guys who could design and fix just about any tube gear, regardless of whether it was sophisticated laboratory recording gear, military radar gear, whatever. Two of these guys worked in the top-top design labs of America’s old companies (Westinghouse and US Steel.) Fact is that 2 out of the 3 guys relied upon a Jackson 648A as their primary tube tester in the lab, and the 3rd guy used a B-K. Yes, there was other laboratory-grade tube test equipment available to them, cost was not an issue, but the Jackson is what received the overwhelming majority of use. In fact, the 648A’s that I bought from those guys were so worn out that almost all of the printing on the faceplates had worn away. On one unit, the shorts switch (which has a snap-action for each position) was actually worn smooth — it just twisted to each position without any snap action at all. And yet, while those Jackson’s looked like crap, they were still working perfectly. I can’t think of any better compliment for the Jackson 648.

How Long Does a Calibration Last? (update 4/29/2008)

Since posting this article, many people have asked how long a calibration lasts (before the tester will need recalibrated). This is an impossible question to answer, but I can explain which factors most influence the calibration.

For the early 648 models (such as the 648A used as an example in this article), the condition of the 1S5 tube directly relates to the accuracy of the Line Control Calibration. As the 1S5 tube ages, or develops a defect, the Line Control will become inaccurate. This is the most common source of inaccuracy after a complete calibration has been once performed. This problem is easy to spot (the Line control will not correspond to your AC line voltage), and easy to recalibrate. Just keep in mind that installing a new 1S5 will not solve your problem, you will always need to adjust R15. [See below for discussion to eliminate the 1S5 and upgrade to solid-state circuitry.]

Next, component failure could occur. Typically, this would be resistors changing value. Obviously, you will not have any notice when this is going to happen. Assuming diligence in your restoration, this is a less-likely situation.

Finally, continuity issues relating to the switches/buttons and sockets. At a minimum, you should clean switches and sockets at least once a year (more often if you are a power user) and treat them with Deoxit. This is good practice and should be considered part of regular maintenance.

Reduce the wear on your tester sockets — Get in the habit of never testing tubes with dirty or rusty pins until you have completely cleaned the tube pins first. This will also increase the accuracy of your tube testing, because dirty or rusty tube pins can significantly affect the accuracy of the test result.

———-

[UPDATE: May 24, 2008. Procedure to eliminate the 1S5 tube from models 648 and 648A, upgrade to solid-state circuitry.]

Newer versions of the “648″ (648R, 648S, etc…) use a solid-state circuit instead of a 1S5 tube. The 1S5 tube is found in the earliest models, “648″ and “648A”.

The following discussion will show how to eliminate the 1S5 tube and upgrade to the solid-state circuit found in the newer models. This project is not for amateurs, will require work, and will require recalibration. The benefit is that the tester calibration should remain accurate much longer with the solid-state circuit. If you are diligent in checking your calibration on a regular basis, then there is no reason to perform this upgrade. Test results are not changed (assuming proper calibration). This upgrade is more suited to power users.

Parts needed: 5000-ohm variable resistor, 1N4007 diode, piece of wire (the 648S uses a green coated wire, if you want to be consistent), and a copy of the 648A and 648S schematic.

In summary, the newer models replace the 1S5 tube with a diode, change R15 variable resistor value to 5000 ohms, and have wiring changes in the “X” and “Y” button switch banks.

To begin, I remove the 1S5 tube, and then remove the original 400-ohm variable resistor (R15) from the terminal strip. Leave the terminal strip in place. A white-wire-with-black-trace is attached to the center terminal of this resistor, and this wire goes to the 1S5 tube socket. Cut this wire at the tube socket, and remove old R15 and this white/black wire.

(Removing old R15)

Take the new 5000 ohm variable resistor and solder the banded side of 1N4007 diode to the center terminal. Observe on the schematic that the one end of R15 is attached to the (-) from the transformers. If you follow this circuit, you will see that it also connects at Pushbutton “C” on the upper “WVU” pushbutton bank. In fact, the black (-) from the transformer actually is connected here. Therefore, this Pushbutton C junction will be a very convenient location to install the new (R15) 5000-ohm variable resistor. Solder one end of this variable resistor here.

(Good location to install new R15 and Diode)

The other end of the diode connects to pin 4 of the tube socket (R16 resistor is also connected to pin 4).

Run a wire from the remaining end of the 5K variable resistor to the 30V transformer lead (white-wire-with-brown-trace), which is found in the pushbutton “X” bank, the terminal below the white-wire-with-gray-trace. (The Jackson 648S uses a green color wire.) Examine the “after” photo below to see where this Green wire is attached on model 648S; it is on the bottom right side the photo, inside the red-circled area.

At this point, you need to study these two photos to show how the wiring must be changed. I will describe the changes to the best of my ability, but examine the photos and schematics. I have circled in red the changes that must be made. The photo on the left is “before” — what your 648/648A looks like now. The second photo is “after” — from a model 648S, which has the updated wiring for the solid-state circuit, and what you must achieve. Click each photo to enlarge.

I would describe the change as follows: on the same terminal where you just connected the “green” wire that you ran from the new 5K variable resistor, there is a short jumper wire with black heatshrink going to the junction of R14 resistor and a solid yellow wire. Remove the black jumper wire from the terminal where it meets the yellow wire and R14. Move this end of the black jumper wire one step to the right, to connect to the white wire with brown trace.

That concludes the solid-state circuit conversion, and now recalibration must be performed.

regards,

Bob Putnak.

eBay ID = rjputnak

[Above: Jackson 648A, after final restoration (cleaning, repair, calibration, knob indicators repainted, SS-9 socket saver installed) ]

Article Tags: 648 | 648-1 | 648-1t | 648a | 648p | 648r | 648s | calibration | Jackson | rebuilding | repair | restoration