1S1/Repairs

From TekWiki
Revision as of 02:51, 15 January 2021 by Peter (talk | contribs) (Created page with "==David Wise 1S1 (S/N 002993) repair notes - August 2017== ===SYMPTOM=== Unsynchronized trace - blizzard of samples - no reaction to TRIGGER SENSITIVITY. Slight reaction to...")
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

David Wise 1S1 (S/N 002993) repair notes - August 2017

SYMPTOM

Unsynchronized trace - blizzard of samples - no reaction to TRIGGER SENSITIVITY.

Slight reaction to RECOVERY TIME. Reaction to TIME/CM recovery range.

DIAGNOSIS

Trigger free-running due to slight misadjustment of R420 or R460.

TREATMENT

Adjust INT TRIG LEVEL R420 and CONTROL TD BIAS R460. I did whatever it took to get triggering - need to redo per manual.

SYMPTOM

Vertical shift as SMOOTHING is turned.

TREATMENT

Adjust R110 MEMORY BALANCE.

Step 14, page 5-23.

SYMPTOM

Vertical shift as mVOLTS/CM is turned.

TREATMENT

Adjust R30 BRIDGE BALANCE

Step 15, page 5-23.

SYMPTOM

Vertical shift as mVOLTS/CM VARIABLE is turned.

TREATMENT

Adjust VAR BAL

Step 16, page 5-24.

SYMPTOM

Trace not centered when VERT POSITION is centered.

Adjust R194 POSITION RANGE

Step 17, page 5-24.

PREEMPTIVE TREATMENTS

Adjust +19 V / -19 V

Adjust INT TRIG LEVEL R420

Step 4, page 5-11.

Adjust CONTROL TD BIAS R460

Step 5, page 5-11+.
Readjust R420.

Adjust STAIRCASE DC LEVEL

Step 6, page 5-13.

Adjust INVERTER DC ZERO

Step 7, page 5-13+.

Adjust DELAY ZERO

Step 11, page 5-20.
Used 106 instead of TD Pulse Generator.

Steps 18-19 want 111 Pretrigger Pulse Generator, 20 ns pulses at 40 kHz.

Can I get by with Wavetek 145? 25 ns width, 6.5 ns rise at TTL jack.
How about the 281 TDR Pulser? 460 mV, 5 ns wide, 750 ps fall, about 50-100 kHz.
How about the 106 Square Wave Generator?

Adjust Loop Gain

Step 18, page 5-27+.

Don't need 111 Pretrigger Pulse Generator; 281 TDR Pulser works fine. 460 mV, 15 μs wide, 750 ps fall, about 50-100 kHz.

Double trigger: the goal is to trigger on both the rising and falling edge of the signal. That's twice the signal frequency, and it makes the 1S1 take alternating samples of high and low voltage. This is a stress test on the memory circuit; it will reveal the slightest aberration in loop gain. Just advance TRIGGER SENSITIVITY while tweaking RECOVERY TIME.

Wavetek 145 works, on either FUNCTION or TTL output, but double trigger is harder. SYNC and TTL outputs need attenuators too.

106 works well, and you can test all mVOLTS/CM settings using external trigger. Cable the trigger straight in, and signal from FAST output via attenuator like HP 355D. Set near max frequency, tweak SENSITIVITY on the 1S1 and frequency and symmetry on the 106.

SYMPTOM

Noisy/intermittent changes in loop gain when mVOLTS/CM VAR knob is turned.

Gain settled down to a constant value after manhandling the switch assembly, that is, grasping it and applying twisting and rocking stresses.

Noise reappeared next day.

TENTATIVE DIAGNOSIS 1

Corroded ground path to chassis changing attenuation.

Why use chassis? Recipe for trouble.

Maybe there's a weak or open path somewhere and it's going through the chassis when it shouldn't.

TENTATIVE DIAGNOSIS 2

Dirty switch contacts go partially open when assembly is flexed just right.

SYMPTOM

Insufficient loop gain.

C135 (Memory Gain) at minimum (max gain), R95 (Memory Gate Width) at optimum, loop gain still slightly less than 1.

Less than 1 at all mVOLTS/CM settings except 2 mV inconclusive, too much noise. (Use Wavetek 145 and external trigger.)

Have not messed with Bridge Standoff or Snapoff Current.

SYMPTOM

Wavetek 145 PULSE mode functions intermittently. (FUNCTION modes are okay.)

Wavetek 145 SYNC output functions intermittently. (FUNCTION output okay.)

Reacts to percussion. (Whack it on the side.)

Just what I need, a tool breaking in my hands.

TODO

Adjust Timing Cal

Step 8, page 5-15.

Or:

  1. Feed attenuated 5us time marks to SIGNAL IN
  2. Set TIME/CM to 5 μSEC
  3. Adjust 1S1 for reasonable display
  4. Indicator scope in X-Y
  5. W/Z/1A5 in A-Vc mode
  6. Connect VERT OUTPUT to X
  7. Connect HORIZ OUTPUT to W/Z/1A5 PLUS
  8. Adjust R335 for 1 pulse per volt

Adjust Comparator Level

Step 9, page 5-16.

Adjust 50 ns Ramp Timing

Step 10, page 5-17+.

Adjust Sweep Length

Step 12, page 5-20.

Adjust Bridge Standoff

Step 19, page 5-29+.

Trying it with Wavetek 145 FUNCTION output. Can't get a nice 20 ns pulse, use 50 ns and set 1S1 to 10 ns. (Maybe it's because the 145 has the wrong output transistors right now.) I can see oddness at about 5 Vp-p, but not what the manual pictures.

20 WANTS 067-0513-00 TUNNEL DIODE PULSER, <35ps RISETIME

How about 284, 70 ps?
How about 067-0681-01, 125 ps?
Not 281, 0.75 ns.
Not 065-0544-00, 1 ns.
Not TU-5, 1.5 ns.

Adjust Risetime

Step 20, page 5-30+.

Adjust Vertical Gain

Step 21, page 5-33.

October 2017

Tweaking Bridge Standoff and Snapoff Current can raise loop gain to 1 without - I think - missing the risetime and input voltage specs.

VintageTek Museum 1S1

No trigger. You can test D449 in-circuit if you remove Q254.

I noticed that the bad D449 still had switching action, but with elevated  Vp. This diode is more or less in parallel with germanium transistor Q254. Due to the high  Vp, Q254 was on continuously. I wanted to see if the circuit would work if I popped in a 2N3904.

No go, so I went ahead and installed a new 152-0154-00 for D449. That got reasonable action in the sampler, memory, fast ramp, and staircase! But only if I set Trigger to FREE RUN.

Then I put in a new 152-0771-01 for D430 and got normal triggering. At this point everything seems to be working; it just needs calibration.

I found that the risetime is way slow, almost half a nanosecond. I think it's because the SNAPOFF CURRENT was cranked to maximum. I backed off to 350ps, but I forgot to check BRIDGE STANDOFF (which also affects risetime and loop gain), so it may have been wasted work. If not, then I have solved the riddle of the excess loop gain, because now there's only about 10% spare. That is so close to 0% (my instrument) that unless some new revelation comes along, I think I will leave mine alone. I was all set to borrow sampling diodes from the Museum's instrument and install spare Memory Gate diodes. I think MEM BAL indirectly affects gain, but I haven't wrapped my head around it yet.


Nailed the culprit but don't understand the crime.

Q80, the Snapoff Diode driver, was weak. That is, when I installed a spare (and turned down the Snapoff Current to get back to 350 ps risetime), there was plenty of gain.

I tried a bunch of other transistors, including some HF Germanium types, and MPSH81 and 2N4258. Nothing.

This is a 151-0083-00, 2N964 Selected. But selected for what? I can't find any specs.

Right after I figured this out, one of my sample bridge diodes opened up. I have ordered some HSMS-8202's.

9-October-2017. Probably selected for avalanche performance on some test jig I don't have a description of. The good -0083 avalanches at 20 V while the poor one avalanches at 30 V. That's on my Heathkit IT-1121 Curve Tracer, no steps, 50 μA/step, 10K. Maybe 30 V is too far from 19V, the normal operating point. The manual does say "a few volts short of avalanche."

I have a couple of generic 2N964's coming. I ordered two -0083 from Vernonia Northern but have not heard back. Hakan Hintze of TekScopes is mailing me some too. For free! Plus sampling diodes. What a guy.

15-October-2017. The replacement -0083's performed worse than the original, and the generic 2N964's worse still. Eventually I discovered the secret sauce.

  1. LEAD LENGTH. As you insert Q80 deeper in the socket, loop gain rises.
    Cut the leads to absolute minimum. Also grind the socket down.
    I'm thinking of removing the socket and sinking individual contacts into the board, or even omitting it altogether and soldering the transistor.
  2. LEAD DRESS. Move the T88 red output wire (positive bridge strobe) as far away from the bridge output as possible.

Adjust in the following order.

(a) BRIDGE STANDOFF for 2 V positive or negative input without baseline lift;
(b) BRIDGE BALANCE for no baseline shift across mVOLTS/CM;
(c) SNAPOFF CURRENT for 350 ps risetime;
(d) MEMORY GATE WIDTH for maximum loop gain;
(e) MEMORY BALANCE for no baseline shift across SMOOTHING;
(f) LOOP GAIN to maximum.

Until you've performed the above steps, comparing loop gain is meaningless.

After performing the above, my 1S1 maximum loop gain is 1.18.

I was using three original sampling bridge diodes and one borrowed from the VintageTek Museum's 1S1, which I'm also working on.

When adjusted, VT's 1S1 also topped out at 1.18, in spite of the drastic difference in snapoff current - 13 mA vs 40 mA. That difference only occurs when VT's Q80 and VT's D87 are installed. Substitute either part and you're back to 40 mA.

Same thing happens if VT's Q80 and D87 are in my 1S1. Some kind of magic there, those two parts are just made for each other.

I've been thinking about the 1S1's bridge sample strobe circuit, a snapoff diode driven by an avalanche transistor. Except for a 1 nF coupling cap, there's nothing between the collector and the cathode except circuit board trace.

Between pulses, the transistor is off, collector at -19 V and the cap charged. The diode is forward-biased, so the junction is full of charge carriers. When the transistor avalanches, it wants to go from high-voltage no current to low-voltage high current, RIGHT NOW, but the diode stolidly continues to conduct until the last carrier is out.

It's the classic case of irresistible force vs immovable object. I suppose it's the transistor that gives, via collector region resistance.

I suppose trace inductance does a little bit to isolate the two, but in that case, wouldn't you think that less inductance would mean less stored energy and a smaller pulse? But the opposite seems to be true, since shorter transistor leads equals higher loop gain. I don't know what to make of it.

22-October-2017

I built four HSMS-8202 modules to replace D11. Each HSMS-8202 is two diodes in series which gives about 8V breakdown. A "T" of bare fiberglass 1/16" circuit board material. It has to be clean to avoid leakage. 1/4" x 1/4" trunk and 1/16" x 1/16" arms at one end of the trunk.

File the corners of the arms so the cross section is octagonal. Glue the HSMS-8202 between the arms. Wrap the arms with #30 wire-wrap wire. Solder to HSMS-8202 and solder the wrapped wire. File down to 0.090" diameter. Plug in and enjoy.

I got max loop gain of 1.28 (vs 1.15 with 152-0433) and bridge standoff > 2 V. I didn't have to change snapoff current much.

Two of the modules are leaky, about 25Meg vs about 25 ns. I hoped it was a flux bridge, but when I built a blank under dirty conditions, it was fine. So I damaged the chips, or maybe they were bad already. I was able to use them anyway. I put them on the input corner, which is low impedance.

Tom Lee at TekScopes, 9-Dec-2020

The 8202 is a VERY delicate beast (typical trade off between robustness and speed here). It is easily damaged by ESD during handling. It's a 4 V breakdown device, as compared to the HSMS-2825's much greater 15V breakdown. Less-capable diodes would work as well in this case, and would probably be more robust. As you and Jim Ford have noted, the HSMS diodes are now officially obsolete, so the usual distributors no longer stock them. But I see that Dan's Small Parts and Kits has HSMS-2825 (dual diodes) at 10 for a buck. Hard to beat that. Buy a bunch more stuff to get your shipping dollar's worth. Me: HSMS-2825 is two unconnected diodes. HSMS-2822 is a series pair which crosses to Macom MA4E1338B1. HSMS-2825 has no direct cross. I'd use Infineon BAT68-04, in stock at Mouser for $0.39 each. Or BAT17-04 but it is delicate like HSMS-8202.

SYMPTOM: Drift in MEM BAL, with drastic readjustment needed when switching extenders or mainframes.

CAUSE: The heater current regulator was shorted, maybe since the instrument was new! There was no insulation on the transistor pins, and E was touching the heat sink and shorting to C. The heat sink is anodized, but I guess there's a weak spot in the hole.

POOR FIX: Reposition the heat sink. DONE. BETTER FIX: Put insulation on the pins. Has to endure soldering heat briefly.

ALSO: There was no heat sink grease! I added some, but not everywhere, only what I'd applied to a D44C8 before I realized the old transistor was okay.