Vactrols

Vactrols have gone through a number of manufacturer changes.  Vactec is the first manufacturer of vactrols and who originated the name.  In 1983 they were acquired by Perkin Elmer.  In 2010 this group in Perkin Elmer was spun out and became Excelitas Technologies who manufactured them until 2015.  CoolAudio is manufacturing the VTL5C3 and Xvive is manufacturing a complete line.  I find a number of  vactrols in my repair work where the range is just too high for the circuit to operate properly. In some of my builds I match vactrols and notice that they vary wildly.  I decided to document random samples of different brands I have available.

 

VTL5C3/2

I only had two samples of the original Vactec and they differed at low currents but matched up nicely at higher currents and the resistance side to side was also well matched.  Both the Excelitas and Xvive varied widely in performance.

For the Xvive, #1 and #3 are the closest matched.  For the Excilitas, #2 and #3 are closest matched.

VTL5C3/2

Brand

Current

           #1

            #2            #3          #4
Left Right Left Right Left Right Left Right
Xvive 1 mA 59K 52K 39K 38K 60K 54K 123K 90K
Excelitas 22K 21K 86K 78K 68K 63K 52K 48K
Vactec 16K 16K 26K 27K
P-E Spec 55K 55K 55K 55K 55K 55K 55K 55K
Xvive 2 mA 23K 21K 14K 14K 21K 19K 29K 24K
Excelitas 9.3K 8.7K 31K 29K 25K 23K 19K 17K
Vactec 8.4K 8.3K 11.8K 12.5K
Xvive 5 mA 9.0K 8.5K 5.6K 5.3K 8.4K 7.7K 8.9K 7.6K
Excelitas 3.3K 3.0K 10.6K 9.7K 9.0K 7.5K 6.3K 5.8K
Vactec 3.7K 3.7K 4.7K 5.0K
Xvive 10 mA 5.2K 5.0K 3.3K 3.1K 5.0K 4.6K 4.5K 4.0K
Excelitas 1.5K 1.5K 4.8K 4.5K 3.7K 3.5K 2.9K 2.7K
Vactec 2.2K 2.2K 2.5K 2.7K
Xvive 20 mA 3.3K 3.2K 2.1K 2.0K 3.3K 3.0K 2.6K 2.4K
Excelitas 880R 830R 2.6K 2.5K 2.0K 1.9K 1.6K 1.5K
Vactec 1.3K 1.3K 1.4K 1.5K
Xvive 40 mA 2.2K 2.2K 1.6K 1.6K 2.2K 2.3K 1.5K 1.6K
Excelitas 600R 560R 1.8K 1.6K 1.3K 1.2K 1.1K 1.0K
Vactec 910R 890R 920R 950R
P-E Graph 1.5K 1.5K 1.5K 1.5K 1.5K 1.5K 1.5K 1.5K

The Perkin-Elmer specification for the VTL5C3/2 is a bit confusing.  The characteristics table shows a 2R on resistance at 40 mA which is extremely low for a CdS photocell.  However, the graph shows a more reasonable ~1.5K.  I don't understand why such a large difference unless they dropped a "K".

 

VTL5C3

Both the Excilitas and Xvive varied widely in performance.  For the Xvive, #2 and #4 are the closest matched although they all tended to converge at the higher currents.  For the Excilitas, #3 and #4 are closest matched and likewise they tended to converge at at the higher currents.  In general the Excelitas were lower resistance than the Xvive at currents above 5 mA.

VTL5C3

Brand

Current

#1

#2 #3 #4
Xvive 1 mA 32K 26K 21K 29K
Excelitas 35K 24K 14K 18K
P-E Spec 30K 30K 30K 30K
C-A Spec 10K 10K 10K 10K
Xvive 2 mA 14K 13K 8.8K 12K
Excelitas 14K 9.5K 6.3K 7.5K
Xvive 5 mA 6.1K 5.5K 3.5K 5.2K
Excelitas 4.5K 3.0K 2.3K 2.5K
Xvive 10 mA 3.5K 3.3K 2.0K 3.0K
Excelitas 2.0K 1.3K 1.1K 1.3K
P-E Graph 4K 4K 4K 4K
C-A Spec 1K 1K 1K 1K
Xvive 20 mA 2.2K 2.0K 1.2K 2.0K
Excelitas 1.1K 730R 647R 730R
Xvive 40 mA 1.5K 1.4K 800R 1.3K
Excelitas 700R 467R 647R 730R
P-E Graph 1K 1K 1K 1K
C-A Spec 500R 500R 500R 500R

The Perkin-Elmer specification for the VTL5C3 is likewise confusing.  The characteristics table shows a 1.5R on resistance at 40 mA which is extremely low for a CdS photocell.  However, the graph shows a more reasonable ~1K.  Again, I don't understand why such a large difference unless they dropped a "K".

 

I didn't have a CoolAudio part to test.  The characteristics table shows a 500R on resistance at 40 mA and the graph shows a more reasonable ~1K.  At 1 mA the characteristics table shows a 10K on resistance and the graph shows ~50K.  They still don't match but are more similar.

 

 

On/Off Characteristics

I hand match my vactrols by applying 5 mA of current and immediately taking the first resistance reading. I chose 5 mA because the turn-on and turn-off specifications for some vactrols is at 5 mA. I bin them according to the first two digits, so 2K4, 4K3, or 0.89R. Then I match trying for the same group or the second digit +/-1. There was a discussion on the net about building a LFO using a vactrol and a op-amp or comparator. I designed and measured such a circuit but it seemed overly complicated. I decided a better method was to drive the LED with a low frequency square wave and monitor the voltage using the vactrol in a resistor divider. I chose to do just one of each type as I was more interested in qualitative than quantitative data.

My setup consisted of my square wave generator driving the vactrol with a 1K series resistor to ground. I monitored the voltage across this resistor with my oscilloscope (yellow trace) and adjusted the generator output for a 5V drop for 5 mA which is what I hand match at.

I then powered the resistive side of the vactrol with 5V with a 10K to ground and monitored this voltage (cyan trace) on the oscilloscope. I had to change time base settings because of the different delays. I tested an Excelitas VTL5C3 and VTL5C4 and a surplus Silonex NSL-32SR3.

When hand selecting vactrols at 5 mA the resistance readings can vary from <1K to over 6K. The issue with selecting a resistance is not all vactrols may reach that value. Instead the specifications are to 63% of final value. This is the Excelitas VTL5C3 turn-on to 63% of 2.5K in 2.4 mS. The typical specification is 2.5 mS.

 

This is the Excelitas VTL5C4 turn-on to 63% of 500R in 80 S. The typical specification is 6 mS.

 

This is the Silonex NSL-32SR3 turn-on to 63% of 640R in 26 S. The typical specification is 5 mS.

 

The turn off measurements are more critical for most applications. I chose 100K which corresponds to 0.45V. This is the Excelitas VTL5C3 turn-off curve to 100K at 18 mS. The specification is 35 mS maximum.

 

This is the Excelitas VTL5C4 turn-off curve to 100K is about 180 mS. The specification is 1500 mS maximum

 

This is the Silonex NSL-32SR3 turn-off curve to 100K is 120 mS. The specification is 10 mS maximum.

 

I suspect the VTL5C3 turn-on time is a bit of an outlier. In all cases the turn-on time was much faster than the turn-off. The VTL5C4 which is specified slower than that VTL5C3 was indeed about 10X slower. The Silonex was between the VTL5C3 and VTL5C4.

I had a "bad" VTL5C3 that I had removed from a repair and tested it. The turn-on was much slower at 12 mS and it only reached an on resistance of 5.6K. Turn-off was fairly similar.

With a sample size of just one not a lot of conclusions can be made other that the VTL5C4 is about 10X slower for turn-off than the VTL5C3 and the Silonex is in between.

 

 

Buchla Vactrol Specification

In working with the M.E.M.S. team, they supplied me this Buchla Associates Series 200 Vactrol Selection Specification dated January 12, 1971. I believe this was intended to be driven by a 284 as it could be setup to supply the four 15V timing pulses. I designed a test circuit that would drive it directly. Note the Buchla page has been reformatted to letter size.

Buchla Vactrol Selection Specification

 

After I built this, I added four diodes not shown in the schematic in series with the four lower resistors (180K, 39K, 7K5, and 1K5) to prevent summing with the resistor voltage for vactrols with very low resistance. It does reduce the LED current just a bit which I felt was negligible.

This tester displays the voltage attenuation at four different currents (10 mA, 2 mA, 400 A, and 80 A) and resistances at 3K3 @ 10 mA, 22K @ 2mA, 220K @ 400 A, and 180K @ 80 A will result in a mid-range 7.2V output signal. This scope image is a LED and 3K3 resistor to verify the tester. The high peak corresponds to 10 mA and measures 7.2V. This must have been a challenge on a non-storage oscilloscope due to the low repetition rate.

 

Note that the LED voltage drop is roughly 1.6V and a 284 can only go to about 14.7V so the currents will be less than those shown on the drawing. This scope image shows the LED voltage (cyan) testing an arbitrary vactrol (yellow). For the four steps the LED voltage drop measures 1.76V, 1.68V, 1.6V, and 1.56V which corresponds to 8.6 mA, 1.7 mA, 336 A, and 73 A current steps.

 

This is a Vactec VT2C2 can vactrol which measured quite low. This would meet the specifications for Selected and Gate.

 

This is a Vactec VTL2C2 vactrol which is similarly low and meets the specifications for Selected and Gate.

 

This is another Vactec VTL2C2 vactrol which measures similar to the other two.

 

This is an Excelitas VTL5C3 which I had measured 1K5 at 5 mA. You can see the slow rise-time (decay). This would meet the specification for Selected and Gate.

 

I measured 10 XVive VTL5C3. They have a wide variation in performance, matching, and decay.

 

 

I tested 10 of the M.E.M.S single vactrols. Their part to part matching is quite good.

 

 

For comparison, I tested a M.E.M.S. vactrol using my original method at 5 mA and also three different currents.

mA Resistance
2 4K2
5 1K9
10 1K1
20 750R

Using the test fixture, the first stage measures about 3.9V at ~10 mA. That corresponds to a resistance of 1K2 which is very close to the above measurement.

 

 

 

home