Restoration Idea 14

Repair Challenge Tektronix 321 Oscilloscope- The First Transistor Oscilloscope

Tektronix has always been noted as a leader in oscilloscope technology starting with analog and now digital formats.

They are well made and reliable test equipment.

Early Tektronix products are great examples of hand crafted electronics made in USA. Their versatility and superior abilities over their competition makes them complex instruments to service and restore. Marketed in 1960, the Tektronix 321 was the first transistorized and truly portable oscilloscope since it can operate on batteries also. The operating bandwidth of the scope was 6 MHz, which was average in those days. The vertical front end amplifier had one tube, a Nuvistor by RCA so it really was a hybrid. The 321 would be replaced by the 321A . The 321A had improved circuitry and eliminated the Nuvistor tube with a new transistor called the Field Effect Transistor or FET in 1964.

Article is on The Tektronix 321 Serial Number 2149. Later productions made changes to the cabinet and went to a round CRT bezel over the rectangular one shown. Some panel changes were made to the charging switch and knobs for vertical input and time base selector. Incremental circuit changes were done thoughout the model run including newer and improved transistors.

REPAIR date May 2015

I own two Tektronix 321's and a 321A. The 321A is totally transistorized. Most of the transistors used are germanium type as silicon transistors were not widely available. Two of these oscilloscopes had no sweep and displayed a dot on the CRT face. This failure could lead to a CRT face burn if the intensity is too high. Repairing Tektronix products can be difficult because of the complex circuitry that provides the high performance offered. This scope has been in storage for many years and when things really settled down it was time to take it on.
The time base generator circuit looks simple but servicing is complex because the the generator is essentially a precision oscillator (multivibrator and Miller integrator with feedback) and triggering or gating system. If one part fails the entire system is disabled. Usually when a radio or TV is repaired we look for open or shorted transistors (complete failures). However in a Tektronix time base generator, if a transistor starts to show a slow failure like the transfer gain (hfe) the system will fail but the transistor looks good on the good/bad transistor test and even on no signal DC measurements. Unless you have in circuit dynamic test set up repair can result on trial and error and educated hunches.
Shown on the RIGHT is the time base generator used on the 321 oscilloscope. The top circuit is the CRT unblanking amp to control the retrace location and sync it to the time base generator. Below it is the gated multivibrator which develops the sweep and sets the timing to blank/unblank the CRT. The circuit to the right is the Miller integrator and driver that sets the timing and sends a feedback to the multivibrator to oscillate or stop.
1. THE FIRST EDUCATED HUNCH: The two multivibrator transistors were GE 2N169A NPN audio transistors. They are rate grown transistors and have low transfer hfe (30-50). The DC voltages were good at the emitters and base but off at the collectors. So I first suspect them. The diode test show the forward drop around .31 V for each junction (BE and BC). Germanium transistors at low base currents should be around .25-.27 V. So I replaced them with two newer silicon transistors 2N2369A which have high transfer hfe (over 100). You can use silicon to replace germanium transistors if you can afford the higher junction drop of 0.63-0.67 V.
2. SILVER PLATED CERAMIC TERMINALS- WATCH OUT!: 1950-mid 1960 Tektronix oscilloscopes were point to point wiring and done on custom made low loss and heat resistant ceramic terminals that had slots for drop in and solder (like lap soldering). However for reliability and and longevity, Tektronix plated the terminal mounting slots with silver amalgum material (like a silver tooth filling) and required higher temperature silver bearing solder for soldering components. They warned if you don't follow instructions the plating will come off. From experience I have seen this happen if you continually do this on a slot. Since I don't have any silver bearing solder, I tried to mount the new parts on existing wire leads where possible with conventional solder. If I had to directly solder to the terminal slot, tin the lead and heat the slot allowing the new lead to solder in.
3. HUNCH CORRECT?: Well not exactly. I did get a nicer spot on the screen but still no sweep. So time to make another educated hunch. The next step was to again look at more DC measurements and even use an oscilloscope to check an oscilloscope. If the multivibrator was working but there was no feedback to gate the multivibrator from the Miller circuits I would get no reset and therefore no sweep. I checked the DC levels from the Miller circuit driver transistor 2N671 (the large GOLD color one in the picture to the left). It looked fine but the base was hung up one one level (no signal from multivibrator. I also did static tests and the junction voltages look good for a germanium transistor. I decided to go further back to the beginning.
4. THE SECOND EDUCATED HUNCH: The input PNP transistor (2N1997) to the Miller integrator could be low gain unable to provide a large enough output to the integrator. Again the junction voltages were checked and they were a bit high like the multivibrator transistors. This transistor was another low hfe audio transistor but capable of higher output voltage and current. I decided to use a 2N3906 PNP silicon that had better specs than the 2N1997. After installation turned it on and I got a crisp sweep on all time settings. I think that all the transistors I replaced based on both hunches was the right thing to do because the oscilloscope seems to responds faster and sharper than my other 321 which has all original parts still in.











5. CLOSE UP OF CHANGES (ABOVE)
 Upper part are the two 2N2369A NPN silicon transistors replacing the 2N169A germanium type. On the lower right is the black square shape 2N3906 PNP silicon transistor replacing the 2N1997 PNP germanium type. NOTES of INTEREST- the oval shape transistor above the 2N3906 is an early TI silicon transistor J509 (2N1592) and the big gold one is the output transistor of the Miller integrator circuit the 2N671.
6. SUMMARY:
How times have changed. If you looked at the parts list of the 321, Tektronix used many types of transistors. Today only a few are needed and can be used all over this oscilloscope. Also there is a price difference. The total cost of the transistors I put in is less than a dollar, Back then it would be over ten dollars. Good chance that the new transistors will be more reliable that the originals and perform better.

Here are some interior views of the Tektronix 321:


1. The CRT- Tektronix designed and built their own cathode ray tubes for consistency and highest performance.


View of the T321P31 CRT. Tektronix used a special phosphor P31 which is more pleasing color, shorter persistance for better response than the common P1 phosphor in most oscilloscope CRTs. Notice the pins just before the CRT bell. Those are the deflection plate connections. Instead of using only a gray coating inside the tube for shielding, Tektronix used that and spun wire inside for shielding and strength. Tektronix oscilloscopes were rugged and reliable.

2. The Vertical Amp and CRT circuits: All hand wired and notice the only vacuum tube (looks like a thimble) on the right photo.

3. Tektronix 321 S/N 2149 Calibrated and Functioning Again 4. Two versions of the 321. The left is the earliest model

Return to Restoration HOME Page: Restoration HOME

Return to PHOTO COLLECTION INDEX: PHOTO COLLECTION