Skip to content

AN/USM-38 Oscilloscope

April 3, 2014

Here is a Wireless World advert for this instrument that I stole ages ago from another website. I can’t remember where I got it from so if it was you, please let me know and I will acknowledge you for it:
USM-38 3
PREAMBLE
The ebay seller originally had this up at $150 which was very high. He also said that he had turned it on and some smoke came out that he didn’t think was serious. Hmmm. So I contacted him and let him know not to just turn on old equipment because damage may well result, up to and including power transformer failure*. (This author knows, he has learnt the hard way.) I later determined that the smoke had come from the negative rail smoothing resistor that was running into a dry capacitor that was a near dead short. The result was that he re-listed it at $75 with free shipping. I was the only bidder so another relic arrived on my doorstep. I must say, the seller was very prompt in getting this to me so I left him positive feedback.
*Notably, the three power transformer failures that I have precipitated were all in USM oscilloscopes (I told you that I know, damn it). My thoughts on this are that 1/ the USM scopes tend to be rather tightly enclosed, presumably to reduce RFI emissions and 2/ this, probably combined with much duty, resulted in prolonged high temperatures hence degraded power transformer insulation combined with failed capacitors. Also, my experience prior to looking at USM scopes was almost entirely with Tektronix. In their early years, Tektronix experienced a spate of power transformer failures that they addressed by bringing in Gordon Sloat to set up their own transformer manufacturing facility. Tektronix also include 10Ω carbon fusing resistors in each rectifier circuit to protect the transformer and this excellent practice may have been inspired by the intention to make not only the best performing, but the most reliable oscilloscopes possible at the time, so I was spoilt. Encountering the very interesting USM scopes certainly increased my knowledge and experience base!

The general condition of the case is good with the exception of the CRT hood which is showing some denting and loss of the black anodising. The prop tilt stand is missing also the detector probe is missing from the accessories. As usual with new acquisitions, I opened it up, to find it in fair condition but dirty and oily due to oil migration from many leaky paper-in-oil (PIO) capacitors, I ended up replacing most of them including the HV capacitors. Most of the POI caps were superbly neat molded phenolic types, it is a shame that the sealing of the phenolic around the leadouts did not stand the test of time and temperature. All the power supply electrolytics were dry. I disconnected each one and reformed them one-by-one. Only one unit failed (the one referred to as causing the sellers non serious smoke), it did reform but the ESR was so high that it was useless. (I understand that this problem can be due to corrosion of the internal connection between the aluminium foil tails and the solder tab and/or can.) Many of the tube screen cans are exhibiting “season cracking” a phenomenon whereby deep drawn brass will transition from ductile to brittle at low temperatures so this unit apart from being hot, must have spent some time at freezing temperatures too!
At this point, the real work began for it was not happy! A hard-copy manual (that I prefer) was not available, my friend Volker Klocke has the manual in pdf form on his website at
http://www.oscilloscopemuseum.com/scopes/oscilloscope-military-usm-38-sxxx
Here is the backside of the sync amplifier assembly showing replacement (Russian) PIO capacitors:
USM-38 Sync Amp 2

It has early printed circuit boards. The tube sockets connect to the traces by side contacts that do not overlap or engage with the traces physically, instead relying on a solder “bridge” from each contact to the associated trace, a recipe for intermittence! I spent much time re-working these connections, this job was made more difficult by the connections on the sockets which were corroded and would not tin easily, I ended up using flux and then thoroughly cleaning each area using flux-off and a stiff brush. The other task was to test the tubes. I usually do not do this however it has been my experience that the tubes in the AN/USM scopes are often exhausted (which tends to confirm my thought about prolonged hot service). In this case, most of the tubes were strong with only two exceptions including the HV rectifier. It was apparent that somebody had gone through this example at some point, evidenced by some truly horrible soldering and melted insulation. Weller soldering gun anyone? I also found that the HV circuit had been re-wired incorrectly resulting in the unblanking multivibrator (that rides on the negative end of the HV supply to allow DC coupling to the CRT grid) not working.

Failure of PIO caps shows up in two primary ways, low power supply voltage due to a leaky bypass cap(s) or a circuit is drawing too much current due to a grid associated with a cap being high. In this case the HV voltage was less than 1000V, down from 1500V and one of the supply regulator gas tubes would extinguish when I switched in the marker generator. It is worth noting that failure of coupling caps can (and do) cause power transformer failure as well as failed electrolytic caps in the power supply!

PURPOSE and outline description
This OS-57, USM-38 oscilloscope is serial number 527 and was manufactured by the Trad Electronics Corp of New Jersey. Along with many of the vacuum tube era armed service scopes this model falls into the synchroscope category whereby the timebase can be driven by an internal trigger generator that has an output on the front panel that may in turn, be used to trigger the circuit being investigated in sync with the timebase. Applications would include radar circuits and logic circuits. Somebody tried to tell me that I was wrong in using the term synchroscope, insisting that this term refers to the well known AC power phasing device. Well yes it does, and also to the triggering/triggered oscilloscope. In fact Tektronix, in their book “Using your type 535 or type 545 oscilloscope”, refers to the term synchroscope when describing how the gate pulse from the B timebase may be used to trigger both the A timebase and the circuit under examination.
Here is the trigger and marker generator assembly:
USM-38 Trigger & Marker Gen
This model uses the classic (and excellent) 3WP1 CRT with 27 tubes plus 2 hivac neons hence the 30 valve claim in the Wireless World advert above.
The timebase is of the triggered multivibrator type that is ac coupled to a push-pull deflection amplifier that is dc coupled to the X plates. The timebase has 5 ranges, 10mS/in, 1mS/in, 100µS/in, 10µS/in and 1µS/in with the X gain set for a 2.5″ sweep length and the sweep speed turned fully CW. The CRT remains cutoff until the timebase sweeps, to prevent burning of the screen by a stationary dot. A sweep expansion feature is provided that allows 9X magnification of any region of the waveform (why 9X I don’t know).
It has an internal trigger generator, rate variable from 40 to 5000 pulses/S, also a marker generator with settings at 100, 10 and 1µS.
The ac coupled Y amplifier has 5 stages including a pentode long-tailed-pair that provides push-pull deflection, ac coupled to the CRT plates. The front end is a switched attenuator that presents a constant 1M / 40pF load to the input, it has 5 ranges, 1, 3, 10, 30, 100 and 300x corresponding to sensitivity ranging from approximately 200mV/in to >50V/in. A 400nS delay line is also included. A variable up to 1Vp-p calibrated signal is provided that may be switched in; this in combination with a variable Y gain control allows on the spot calibration of the Y channel. The bandwidth of the Y amplifier is 10Hz to 6MHz. A 75Ω dummy load is available that may be plugged into the CF probe socket on the front panel to provide a standardised load to the input.
Accessories that are stored in the front panel cover include a cathode follower probe, a 10x attenuator probe, a detector probe and the aforementioned plug-in 75Ω dummy load.

Working! Here it is displaying a 100KHz sinewave with 10µS markers just visible at the top of each crest:
USM-38 100KHz & 10uS markers

DESIGN
Timebase. The heart of the timebase is a sweep gating multivibrator that may be adjusted from astable (free running) to monostable (triggered) using the stability control that alters the degree of negative bias applied to the multi. When the bias is sufficiently negative, the multi is held in the wait state until triggered, otherwise it will free-run. The multi provides a negative gate (rectangular pulse) to drive a simple capacitor charge sweep circuit. In the wait state, the sweep generator tube is normally conducting, that is discharging the timing capacitor. Upon receipt of the negative rectangular pulse from the gate generator, that is applied to the grid of the sweep tube, the tube cuts off, allowing the timing capacitor to charge or sweep at a rate controlled by the setting (resistance) of the variable sweep speed control. When the gate multi reverts to the wait state, the resulting positive pulse turns the sweep generator tube back on, discharging the timing capacitor (flyback).
The gate multi has a time constant that causes it to pause to allow the sweep to take place before reverting to the flyback and wait state. The gate time constant is switched with the 5 sweep rate ranges and it is also varied in tandem with the variable sweep speed control, holding the gate time constant at approximately 1/10 of the sweep time constant so that the sweep amplitude is limited to about 1/10 of the charging voltage; since the first 10% of an exponential rise/decay is substantively linear this technique results in good (surprisingly so) sweep linearity. It has the further benefit (for a given setting of the X gain) of holding the sweep length constant.
The sync signal is applied to the gating multi via a coupling diode that passes only the negative going signals that are required to trigger the multi while in the wait state. (The multi is anode triggered in case you spot the apparent contradiction between the negative going trigger and increasing negative bias locking the multi into the wait state.) In the sweep state, a positive going signal would be needed to trigger the multi so that as long as the multi is in the sweep state, it is prevented from re-triggering during the sweep by the coupling diode.
The sweep gate is also used to unblank the CRT during the sweep; the positive (inverted) going gate being capacitor coupled to the CRT grid for HF unblanking. Unblanking at low frequencies is assisted by a bistable multi (referred to as the intensity gate shaper) that rides on the negative end of the CRT supply. It is turned on and off by the unblank signal and provides the necessary square topped positive unblank pulse, directly coupled to the CRT grid.
X Amplifier. The output from the sweep generator is buffered by a cathode follower that drives the X gain control. From here, the positive going sweep signal is ac coupled to the long-tailed-pair X deflection amp. The input grid of the X deflection amp is also connected to the X shift control via a diode. The direction of the diode is such that as the sweep moves positive it disconnects, and at the end of the sweep reconnects thereby restoring the grid potential to the shift potential so that the sweep always starts from the same place, reducing jitter on the display.
Sweep Expansion. The X expansion switches in a further gain stage having a gain of approximately 9. In this mode, similarly to the grid clamp above, the start of the sweep is clamped to ground to prevent expansion jitter. The extra gain stage is arranged with a bias control that causes the stage to respond from the start of the sweep and then as the bias is increased the start point moves progressively up the ramp causing the expanded display to move along the waveform under examination. The result is that any 10% of the normal sweep can be expanded and displayed on the screen.
Triggering.
The timebase may be driven by the internal pulse generator or from the signal applied to the Y axis or from an external signal. There is no separate trigger circuit, the sweep gate is triggered directly by an amplified sync signal from the Y axis or from the trigger generator. The way to operate this timebase is to set the sync full ACW then bring the stability down (turning CW) until the timebase free runs, then back up again until the timebase just stops. Then bring the sync back up until the timebase triggers and locks. It may be necessary to repeat this operation if you want a different sweep speed. Triggering from the Y axis or an external signal may be selected from the rising or falling signal.
Synchronisation Amplifier. The sync amplifier consists of two gain stages and the sync selector switch that allows the user to select triggering from from a rising (+) or falling (-) edge from the Y amp or an external signal, or triggering from the internal trigger generator. The gate is triggered by a negative going signal only so the sync amp is arranged to generate a negative going signal from either a negative or positive going signal; + selection causes the sync signal to be inverted and – selection preserves the polarity of the signal.
Y Amplifier and Calibrator. The 1st stage of the Y amplifier is a pentode video amplifier that is followed by the 2nd stage cathode follower followed by a series resistor from the cathode to provide the required 1k source impedance to the delay line that is terminated with a 1k resistor shunt at the input to the 3rd stage that is also a cathode follower. The 4th stage is a second pentode video amplifier which in turn drives a push-pull pentode deflection amplifier that is ac coupled to the Y plates. The delay line drive CF is also used for sync pick-off to the sync amplifier.
In addition to the constant impedance attenuator at the front end, there is a user variable gain control located between the 1st and 2nd stages that allows the deflection to be calibrated using the internal 1Vp-p 60Hz calibration squareish wave. The calibration signal is derived using a + and – diode clipper from the power transformer.
Power Supply. There are 3 DC power supplies, positive, negative and HV (negative). The 250V B+ supply is of the choke input type and this is the first time I have encountered choke input in any equipment! It is vacuum tube rectified. A tap on one side of the B+ winding supplies a half-wave tube rectified C-R-C filtered 200V negative supply, that is used for the timebase including the gate bias and the X amplifier long-tail. The negative 1500V supply is fed from a 1100V winding that is (as is usual) a continuation of one side of the B+ winding, it ends in a winding that feeds the HV rectifier filament. The half-wave rectified HV is smoothed by a simple C-R-C filter. There is a separate heater winding for the CRT and unblanking shaper bistable multi.

Pictures!
Top view:
USM-38 SN527 Top

Bottom view, you may spot the ceramic wirewound resistor that replaced the burnt negative rail smoothing resistor that was caused by the seller “testing” (ho hum) the unit with the failed electrolytic:
USM-38 SN527 Underside

Y channel board and trigger and marker generators assembly:
USM-38 SN527 Y amp, Trigger and Marker Gen

X channel board and sync amplifier assembly, you can see where I cut the failed electrolytic open to instal a new cap:
USM-38 SN527 Timebase and sync amp

Display board with X (top left) and Y (top right) deflection amplifiers, also intensity shaper multi:
USM-38 SN527 PSU & Display Board

Advertisements

From → All Posts

3 Comments
  1. Hi Richard,

    Another very interesting post! How do you like the Russian PIO capacitors? I bought a bunch of them a while ago to use as replacements, but I have yet to try one.

    thanks much,
    ben

    • Hi Ben.
      Thansk!
      I find the Russian PIOs to be of first rate quality, usually better than the ones I am replacing.
      Not all of them are PIO, some are metalised film in a top quality hermetic case. I know this because a friend cut one of the larger types open.
      Richard

  2. Stephen Scott permalink

    A fascinating write up – when I eventually get round to restoring my USM-38, I’ll definitely use the Russian caps through out, agreed they are excellent quality, and look the part too!
    Thanks
    Stephen

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: