HP 1740A Oscilloscope Delayed Sweep Fault

[__i4cy__]

I have an HP1740A oscilloscope that I was lucky to acquire recently, though it did need a good clean and a service. I have given the front panel and case a good clean and used a switch cleaner on the switches and pots. The scope seems to work very well in that the trace is bright, sharp and stable, when the MAIN or A VS B mode is selected. However, there seems to be a problem when the DELAY pot is turned as an intensified area cannot be set in the centre of the display. Turning the DELAY multi-turn pot moves the intensified area from the left side towards the right side of the trace, but the right side remains always intensified (see first photo). Turning the outer DLY'D knob has no effect.

If either the DLY'D or MIXED mode is selected there is no horizontal sweep, and the trace is confined to the left side of the display as if there is no timebase (see second photo).

Having never used this scope before, I'm not sure exactly how the delayed sweep display works. I have followed the steps in the operating manual so I'm guessing that something is wrong with the scope. Hopefully someone can confirm that the scope is faulty, and perhaps suggest something to check and point me in the right direction.

Many thanks for reading this far.

[Radio Wrangler]

I'll try to find some time over the weekend to get one of mine fired up and check what you're doing. It's easy to get things in a bit of a muddle because both timebases have trigger systems so that the main timebase can run and then arm the second timebase to fire when a trigger condition is met. what you're looking for are the settings to auto tun the delayed timebase after the main one has run to the percentage of its sweep set by the 10-turn pot.

I used one of these scopes for about 15 years, but can't remember the timebase switches without a panel in front of me!

David

[__i4cy__]

David, thanks for replying and for offering to compare my results with your 1740A scope. Though I do fear that the second timebase or the gating circuitry is not working right, as the beam does not move horizontally at all despite adjusting the other controls. I'll study the service manual a bit closer to see what I need to be checking. The delay sweep miller integrator output may be a good start.

[Radio Wrangler]

I last had power on mine a couple of years ago, and now I've found it's sulking, too.

I knew I needed to replace the 100U tant for the slowest main timebase speeds, but everything else was working fine.

With the calibrator as a waveform source, it runs fine and triggers reliably. Chop and alt work. But oddly the X-shift controls do nothing right up to the end of their travel when the trace winds up against one end of the screen or the other. What's going on here, then?

The delayed timebase doesn't produce a bright-up on the main sweep when the delay time/div big knob is brought out of 'off' (all buttons out except for 'main timebase')
But if I repeatedly push the delay TB trigger +/- button I can trigger the delay timebase into running (Delay timebase button selected for this)

So the delayed timebase is viable, but isn't being triggered. The 10 turn delay pot does nothing and the X position knob does nothing.

Rats!

It's probably a dirty connector, and I do have one of those tants ready to fit (£20 ouch!)

Sorry.

On yours you need only these buttons in:

Channel A on
Trig source channel A
Main timebase selector

Time vernier and sweep holdoff should be in their cal positions (cw and ccw respectively)

Trigger knobs at 12 o'clock

Main timebase 2ms/div, delayed timebase off. 10 turn delay pot about half way.

You should be able to get a good cal waveform trace.

Now back off the trace intensity until it is fairly dim and turn the delayed sweep time/div to 0.2ms/div. You should see a bright band on the trace about 1 div long, somewhere near the middle. The 10 turn pot should move this.

If you have no brightened band, then the delayed timebase isn't running. Put the trace intensity back up, press delayed timebase and you should now have no trace. Start reciprocating the delay timebase trigger +/- button and you may stimulate sweeps at the 0.2ms/div rate id the delay timebase is OK and is only lacking triggers from the main timebase.... like mine appears to do!

David

[__i4cy__]

David, Sorry to hear that your scope is also sick. I did try your setup procedure, I get as far as “You should see a bright band on the trace about 1 div long, somewhere near the middle”. I cannot get a bright band in the middle, and by turning the multi-turn pot the bright section just moves towards the right (as shown in my first photo). Switching to DLY’D timebase and trying to simulating a trigger does not cause a sweep. As you suspected, the second timebase is probably not working or not being gated properly to the X amplifier. So it looks like I’ll have to do some signal tracing. I have read that the timebase selector switches can go faulty and often need to be dismantled cleaned and refitted, though these ones they seem to appear ok.

BTW, You mentioned that a 100uF tantalum capacitor requires changing on your scope, why the expense at £20? Are they special HP capacitors?

[Radio Wrangler]

No, they're metal-clad tantalum electrolytics with glass bead seals. The used to be fifty cents a pop when we designed them in. I have an HP8566B spectrum analyser on the todo pile and one board in it cooks several tant capacitors slowly to death... high ripple current doesn't help. The going rate at farnell for one particular higher than usual voltage part is 80 quid.

The price of tantalum dust has gone through the roof. so proper parts are very expensive. Capacitor makers flog cheap tants, but to conserve pricy material, the parts are VERY iffy on voltage rating. So much so that HP banned the use of some big name capacitor brands.

I'll have a go at the scope when I'm in the mood sometime. I have an IC765 apart that I need to finish first, then there's a 1GHz digital sampling scope. Ah! Stuff expands to fill all available space, plus 10%!

David

[__i4cy__]

I spent a few hours investigating the scope but unfortunately still no luck tracking down the fault. The HP 1740A service manual on page 8-30 has details for checking the delayed sweep output waveform (TP1). This should be a ramp waveform going from 1V to about 12V (see 1st picture a snippet from the service manual). What I measure is a square waveform trough 1V and peak 2V (see 2nd picture). As a comparison I also measured the waveform from the working main timebase which is of similar design to the delayed timebase, and as expected it correctly generates a ramp waveform going from 1V to about 12V (refer to 3rd picture).

Referring to the schematic (see 4th picture), I then attempted to check the transistors on this board but there was no indication that something was wrong i.e. dead shorts across pins, but as they were tested in situ it is never certain if it's a goodun or dud. The DLYD TIME/DIV switch also appears to making good contact with the PCB.

Short of removing the Delay Sweep PCB from the scope which I have been resisting to do, I am fresh out of ideas.

Anyone got any suggestions to try before I take the PCB out? Thanks.

[karesz*]

Do corresponding all DC-levels with values on schematics?

[MotorBikeLes]

I have a 1703A, a 1725A and a 1741A, but so far the only one I have used and repaired is the 1703A, so comments from that may well not "transplant". Other than self inflicted faults, I have had three faults in almost 30 years, but not much use over this period. However every fault has been in the timebase sections. The first was a triggering problem which involved both Q2 and Q5 on the A6A4 delayed integrator board. About 5 years later, a "no display" fault was traced to Q3 on the A6A9 horizontal Preamp board.
A more recent fault of intermittent no sweep was eventually traced to Q7 on the A6A7 comparator and hold off board. I would need to go through and recheck, but I think these were all 2N3906 pnp types, HP No 1853-0036, usually marked 3-036 if I recall.
So, I would check all those, removing from board if necessary.
If you align all the switches carefully, (note positions), you can line up all the flats on the control shafts, and with the backing bush's grub screw slackened off, pull all three shafts out for easy replacement later. (Mark boards with felt tip pen as a "belt and braces). With my 1703A, it is Sweep Mode at Main Timebase, Main sweep at 10msec/div, and Delayed Sweep at 0.5Usec/div. (For 1703 readers, there are later and earlier versions, these switch positions MAY be different, mine is later type with smaller push buttons).
I find it very difficult to work on the boards in-situ, so pulling out and checking things cold becomes the easy way out. Now if I had some extension boards (anybody??), life would be much easier.
Les.

[__i4cy__]

Thanks for the suggestions.

Unfortunately as Les has mentioned it is difficult to get probes onto the board to make DC voltage tests on the board, though I will try again. Also there aren't any test posts on board, there are however test posts on the main timebase board to aid testing. Extension boards indeed would be a godsend.

Les, I must admit I am surprised you have had so many low power silicon transistors in the scope fail even after 40 years of service especially as HP gurus have designed the circuitry. I can understand semiconductors that have been taken beyond or near to their safe operating area (SOA) such as power transistors being driven hard, or as a result of other related passive components failing. I clearly haven't performed enough 1970s restorations/repairs to have become accustomed to this problem. Thanks also for the advise when removing the board such as marking the wiper positions etc.

REF            HP #           PART #
-------------------------------------
Q1, Q2, Q3     1853-0036      2N3906
Q4             1853-0244      MPS3640
Q5             1854-0691      MPSA18
Q6             1855-0081      2N5245
Q7             1854-0019      ?

I may be jumping the gun here, but in case I need to change Q7, does anyone know what part this is? Apparently it's a NPN SI TO18 PD=360mW.

[Top Cap]

Probably a red herring as this cure for a similar problem was with a 1725A.
After much time spent checking everything it was discovered that there was a poor connection on every tag of the 'Starts after delay' control. This was cured by the application of solder onto the four rivets securing the tags to the control. When full re-calibration was attempted the main timebase trigger control was found to be very touchy and the sweep vernier only half worked. It was then noticed that the Sweep Vernier Control was similar to the 'Starts after delay' control. All faults were cleared after applying solder to the four rivets of this control too and they say lightning never strikes the same place twice.

[Radio Wrangler]

1854-0019 was a TO-18 general purpose RF transistor. Not particularly high Ft, not particularly high gain. Commonly used in a lot of telecomms/RF instruments in the 1970s. I'm trying to visualise their data in the old labstock manual and the microwave division preferred parts handbook. I think they were 15-20v collector voltage rating and below 500MHz Ft. Ft and gain were good around the 5-10mA mark. Try 2N2369A

David

[Radio Wrangler]

Those waveforms seem to fit with the behaviour in your first set of photos. Either the delayed TB ramp generator is dead and stuck, or else it is being kept in the held state and not allowed to run.

Triggering in basic delayed mode (to give the bright-up and to give the fast scan - particularly visible in 'mixed' mode is done by a comparator looking at the main ramp and comparing it to a voltage from the 10-turn pot with the turns counting dial.

I think you need to find the comparator, and look at the main ramp reaching it, and look at the voltage from the 210-turn pot (and see if it moves smoothly) then see if the comparator output is switching in the right places.

David

[__i4cy__]

Originally Posted by Top Cap

Probably a red herring as this cure for a similar problem was with a 1725A.

Thanks for noting your experience. My 1740A scope does not seem to have a "Starts after delay" control on the front panel, but there is a "Sweep after delay" switch - is this what you mean? From the schematic diagram I can only find one reference to it labelled A10S1D and it is being used as a simple SPDT switch. I cannot find on the schematic diagram any other used switch poles referenced, so I am assuming this is the only part of the switch being used. When the switch is depressed it does reliably change the triggering mode, so I am assuming that the switch shouldn't require re-soldering.

[__i4cy__]

Originally Posted by Radio Wrangler

Try 2N2369A

I don't think I have a 2N2369A, but Towers' specifies BSX20 which I probably have. But I'd be very surprised if Q7 in the scope has gone to silicon heaven.

[Radio Wrangler]

The 1854-0019 probably has 4-019 printed on it. They were nothing special, just HP's standard transistor when something a bit faster than a 2N3904 was needed. Another common one was 1854-0071 which was a Texas job similar to BC109.

The 1725A is a very different scope and not one of the same family. I think they just share some cabinet parts..

For the 1740, as you have another scope, probe the main timebase where it goes to the delayed timebase trigger comparator and see you get the correct ramp. Probe the other comparator input and screw the 10 turn pot around to see it covers the right range. Now check that the comparator output switches at the right point. From your description I think there's something wrong hereabouts. It's also the best place to start in troubleshooting delayed timebase faults.

David

[__i4cy__]

I removed the Delay Sweep PCB from the scope (see 1st picture) then checked and cleaned the DLYD TIME/DIV switch contacts, as I have heard it mentioned a few times that timebase switches can be problematic. After refitting the board and the switch assembly the fault unfortunately still persists.

While I had the scope covers off I traced some of the signals. A9Q4 emitter waveform is shown in 2nd picture, which I would have thought should be clipped at +0.6v due to forward biasing of the emitter-base junction, not peaking at +5.2v. Tracing through this signal is lost at the collector of A9Q4 giving a reading of -9.2v (the schematic diagram states -0.5v should be correctly read here). Clearly something is wrong around here! Faulty A9Q4 perhaps? The 3rd picture shows other voltage readings I jotted down.

[__i4cy__]

Originally Posted by Radio Wrangler

For the 1740, as you have another scope, probe the main timebase where it goes to the delayed timebase trigger comparator and see you get the correct ramp. Probe the other comparator input and screw the 10 turn pot around to see it covers the right range. Now check that the comparator output switches at the right point. From your description I think there's something wrong hereabouts. It's also the best place to start in troubleshooting delayed timebase faults.

David, thanks for your suggestion. I did try probing the delayed ramp, but I couldn't get it to output anything resembling a ramp waveform, just the square waveform I pointed out in an earlier thread. But I did find something that may indicate the problem.

[Radio Wrangler]

No, you're looking downstream of the fault, so the waveforms will be all wrong and difficult to interpret.

Go back to the MAIN sweep, have it running with a good waveform, triggering regularly off of, say, the calibrator. Now set the delayed sweep speed knob out of the off position, and have it 3 clicks faster than the main sweep speed knob. The pushbuttons should be in main timebase pushed state. Set the 10-turn delay pot to half way

Now, the picture you should (if it wasn't faulty) would be bright-up marker on the main sweep trace starting about mid-screen and extending over 1cm.

So the first port of call has to be delayed timebase arming and triggering circuit.

The delayed timebase cannot free run of itself. It is a one-shot triggered by the coparator looking at the main ramp and the delay pot voltage.

Find the comparator.
Check the main sweep ramp voltages are OK
Check the voltage from the delay pot is OK. Move it around a bit and explore its limits.

If all these are OK, the output of the trigger comparator should be fllipping at about 50% of the way up the min sweep. It will flip back when the main sweep resets.

Only once you're sure this stuff is OK, do you even go near the delayed timebase.

David

[Radio Wrangler]

Look at the block diagram in the service manual, page 4-3

The delay comparator is in box 7,9 bottom right corner using device U4

Then we'll go to the delay timebase control logic, then to the current switch, to the miller integrator, and then the schmitt.

So we're looking for A7 U4. It's a CA3046 transistor array.

There it is in the bottom left of the first portion of page 8-29

Main sweep comes in to pin 9

The delay pot goes to a unity gain opamp U3, and feeds pin 6. The diagram says the delay pot should swing it from +1.3v to +11.3v. The comparator output leaves pin1 "Delayed enable" and goes to the custom IC U1pin 5

There should be plenty of explanation of how everything works in earlier pages in the book. These really are excellent manuals with full explanations.

David