Hp 5342a & the 1826-0372

HP 5342A Repair Log — Restoring the High Frequency IF Path (1826-0372 Replacement)

 

  1. Instrument & Failure Description

Instrument: HP 5342A Microwave Frequency Counter Frequency Range: 0–18 GHz Unit age: ~1980’s

Observed failure:

  • Direct count (low band) still worked
  • High frequency path (>500 MHz) intermittent
  • Unit could operate for ~1 hour, then fail
  • After failure, sensitivity dropped to zero on high band
  • IF output from A25 IF/driver board disappeared
  • No overload (OVL) indicator, no counting on high band

Classic symptom cluster → IF path failure, not sampler or prescaler.

 

  1. Initial Diagnosis

Key findings during troubleshooting:

✔ IF present after sampler (A26)

✔ IF present into U1 (1826-0372 limiter #1)

✔ IF present out of U1 (≈ 300–400 mV @ test)

✔ No IF into U2 output (same device type 1826-0372)

✔ U2 output stuck at ~3.7 V DC, no IF

✔ U2 pin 5 & 7 bias values correlating to “failed limiter”

Conclusion: U2 hybrid (1826-0372) was dead, U1 was still healthy.

This is consistent with known field failures: on 5342A, U2 runs hotter and drives the IF output to A11.

 

  1. Thermal/Intermittent Behavior

Notable behavior:

  • When cold → sometimes works
  • After ~1 hour → U2 quits
  • Letting unit cool → recovers temporarily

Indicates classic hybrid thermal degradation (bond wire / die attach / metallization deterioration).

 

  1. Known Issues with 1826-0372

Searching HP groups.io and service reports confirmed:

  • 1826-0372 failures are common
  • Most failures are on U2, rarely U1
  • NOS replacements are unreliable due to age

 

  1. Verification Measurements Before Repair

Before U2 death (when it managed to run 1 hour), unit sensitivity was:

  • 19 GHz @ –20 dBm — exceptionally good

This confirms:

  • Sampler on A26 healthy
  • U1 healthy
  • Prescaler healthy
  • A11/OVL healthy
  • Timebase OK

 

  1. Repair Approach Options

Option A — NOS hybrid replacement

Pros: drop-in

Cons: NOS degraded, expensive, unobtainium, sometimes dead-on-arrival

Option B — Rebuild U2 using discrete components

Pros: new silicon, thermally stable, repeatable

Cons: requires RF knowledge, hand-tuning

Chosen solution: Option B

 

  1. Custom U2 Replacement Implementation

Discrete U2 built according to my ugly hand-drawn schematic:

  • 2× BFR93 RF transistors
  • 3× 120 Ω resistors (bias + limiter shaping)

Functionally equivalent to:

  • limiter stage
  • IF driver
  • bias/threshold

Soldered into position of removed U2 hybrid. I recommend to cut the old 0372 instead of try to de-solder it. This because old FR-4 can be quite brittle. But do what you think is best.

Schematic of the hybrid:

 

  1. BAL Adjustment

BAL trimmer controls:

  • limiter symmetry
  • IF clipping level
  • OVL threshold
  • high-band sensitivity

With custom U2 installed:

✔ limiter waveform adjustable

✔ clean clipping achieved

✔ no DC offset anomalies

✔ no oscillation

 

  1. Waveform Verification

Test signal: 50 MHz into RF input Observed on IF:

  • clean sine-to-limiter shape
  • ~1 V p-p
  • symmetric clipping
  • correct bias point

50 MHz works because direct-count path feeds limiters without sampler, making it ideal for BAL tuning.

 

  1. Sensitivity After Repair

Measured with HP generators & power meters:

Band

Measured sensitivity

10 GHz

–33 dBm

18–20 GHz

–24 dBm

These values are better than HP factory spec (typ. –20 dBm at 18 GHz).

 

  1. Thermal Stability After Repair

Long-run test:

  • 8 hours warm
  • No sensitivity loss
  • No IF dropout
  • No OVL malfunction

Confirms custom U2 solution is thermally superior to aging hybrid.

 

  1. Conclusion

The high-band failure of the HP 5342A was caused by a dead 1826-0372 IF/driver hybrid (U2) on A25 board.

Replacing U2 with a discrete design restored:

✔ high-band sensitivity

✔ limiter function

✔ OVL behavior

✔ long-term stability

Performance now exceeds original HP specification.

 

  1. Notes for Other Owners
  • If direct count works but high band dies → suspect U2
  • If high band works cold but fails hot → suspect U2
  • U1 rarely fails but U2 often does
  • NOS hybrids are hit-or-miss due to aging
  • Discrete replacement is viable & robust

 

But of course, sometimes it is the sampler that has a failure. A measurement at the A25 inlet from the sampler with a scope will tell if there is the high and low side from the sampler. Then out from the first BFR-90 amp => second BFR-amp and so on. There were more problems except the 1826-0372 present. I replaced 8 electrolytic capacitors on various places. One trimpot had a problem at the oscillator. Extra soldering was needed at the A25 pre-amp board. But actually, if I have replaced the 0372 first maybe that would fix it?

The 0372 was standard in IF chains from 1978 to 1990. As in instruments:

HP 5342A, HP 5343A, HP 5345A (not exact but works), HP 5370A/B (depends on REV, but works), HP 5350/5351/5352-series (B/C uses 0372), HP 8440/8441 Sampler front-end, HP 8566/8568 spectrum analyzers, HP 8620/8660/8662/8663 generators (not as limiter, input protection & sampler gates)

 

New hybrid can be seen in the middle (U2).

 

 

 

 

 

Welcome

Welcome to sm6why.n.nu.

Facebook

My Newsletter

Links