HP 1740A on the bench, help please?

[qu1j0t3]

I'm watching this thread with interest as I have a 1703A, 1741A and 1746 with various problems (the 1741A is almost fully functional though).

I have a copy of the 1740A Operator's Guide (see attachment). If you PM me your address I can post it to you.

[Specmaster]

I'm watching this thread with interest as I have a 1703A, 1741A and 1746 with various problems (the 1741A is almost fully functional though).

I have a copy of the 1740A Operator's Guide (see attachment). If you PM me your address I can post it to you.

Thats very kind of you but I already have the owners / operators manual and also a copy of the workshop service manual. It was very nice of you to offer though. If I or anyone else here can assist you with getting yours up and running then I'm sure we will do our best.

[Specmaster]

Ok definitely hit blanking next. The time distortion might be an artefact of that. I think blanking is the most common scope breakage so fingers crossed should be relatively ok to debug.

I'm looking at the A7 board (Horizontal output) and I tried to remove the cooling hat for the transistor Q7 which actually came away with the transistor body and left the legs in the PCB 
So I need to replace this now and the manual shows this as being HP part number 1853-0380 and described it as Transistor PNP SI TO-92 PD=350mw which it clearly isn't as TO-92 is plastic encapsulated and Q7 is a T0-18. Got any ideas as to the equivalent for this?

I should also point out that this is not from the from but the parts one I got and I'm going through those boards at the moment looking for obvious faults so I can perhaps try swapping them for the others to see if there is any improvement and then work on those ones.

Which of the boards would you attack first looking for the blanking issues, Gate Amplifier, Main Sweep, Horizontal sweep or Horizontal output. I cannot locate any reference to blanking in the manual at all but as it effects the horizontal I assume that it is part of one or more of the listed boards / sections.

[bd139]

Oh nice - I hate it when that happens. If you have another one on that parts mule, I'd try that first. If not it's forensics time. 1853-0380 comes up as an MPS-H81 for me which is a PNP 600MHz RF transistor. Suspicious!

Blanking usually works as follows. The horizontal timebase in a scope usually has the following states:

ARM -> TRIGGER -> UNBLANK -> SWEEP -> HOLDOFF -> BLANK -> FLYBACK in a loop.

The blanking might be marked up as "Z-axis" or intensity. When the scope is sweeping the intensity is high. When it's blanked the intensity is low. So I'd start by pumping a signal into the scope's external trigger and getting it to sweep on that. This is usually piped out to a z-axis amplifier from the timebase where it is combined with various signals from delayed timebase etc so there is a composite intensity signal. This is used to modulate one of the CRT grids to control brightness. So you want to start at the horizontal timebase and try and find the output intensity or blanking signal and follow that entire chain to the CRT. Usually if these are broken it tends to be the final transistor which controls the CRT grid or some resistors that are bonked or some of the cables between the two sections.

If you're sneaky you can actually use the scope to debug itself. Connect the cal output to ext trigger via a probe. Then connect the channel 1 and channel 2 amps to things inside the scope. Sneaky but it works. If it's on auto you can also trigger on the timebase as well.

[Specmaster]

Haha, it was the parts mule board that I was working on, thought that I could work my way through them, checking for faults like broken connections, duff resistors, caps etc and refurb them, then swap them over for the ones that are currently in the scope and take it from there. That way the blanking etc. might be resolved as I did each board, recapping etc as I went through them, at least that was the theory. 

The transistor most certainly is not a MPS-H81 as that one I already have replaced on the scope and have a spare, that is also a T0-92 package (nice one HP). I discovered that a few people have been asking the same thing on the web going back years and never got an answer to that either, so that board will not get refurbished just yet, thats for sure.

There is a Z axis input on the rear panel, never had occasion to use that on any of my scopes yet (never really understood what it did and what it was there for TBH).       

The timebase and trigger all sits on A7, A8 and A9 boards which is the area that I was suspecting for the odd length of the trace, all my other scopes are all on or very slightly longer than the 10 divisions on the screen but this one as you saw on the vid is more like 20+ divisions.

On a tantalum what frequency should they be tested at, my LCR meter has 100hz, 1Khz or 7.8Khz options and the choice of frequency will make a difference to the ESR value so the wrong choice would mean that a cap could be left in circuit that had a undesirably high ESR figure.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               

[tggzzz]

On a tantalum what frequency should they be tested at, my LCR meter has 100hz, 1Khz or 7.8Khz options and the choice of frequency will make a difference to the ESR value so the wrong choice would mean that a cap could be left in circuit that had a undesirably high ESR figure.                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                               

For tants that are used as power rail decouplers, I wouldn't worry too much.

The simplest sanity check would be to make measurements of old and new tants, and see how much variation there is.

For tants used elsewhere, e.g. in the timebase sweep circuit, ESR might be more important. However, the capacitance vs DC bias voltage might be even more important, since that would affect the sweep linearity. (MLCC ceramic capacitors are astoundingly bad in that respect!)

[Specmaster]

I notice that on some parts of these schematics it gives a voltage rating into a section and another when the voltage leaves that area, for instance, it might show +15V on part of the circuit and at the end it reads +15VF, particularly on a custom hybrid it shows the supply rails as having the VF suffix. Does that mean the supply rail has been passed through an extra stage of filtering as the hybrid is susceptible to ripple etc? 

[bd139]

Genuinely have no idea at all. There’s usually a sheet of conventions etc before the schematics in HP manuals that may (or may not) have the answer.

With respect to tants they seem to have a binary failure mode I.e open or short circuit. The open ones are harder to find than the shorted ones. I like the shorted ones to be honest as long as they don’t blow up a power supply or something.

[tautech]

So I need to replace this now and the manual shows this as being HP part number 1853-0380 and described it as Transistor PNP SI TO-92 PD=350mw which it clearly isn't as TO-92 is plastic encapsulated and Q7 is a T0-18. Got any ideas as to the equivalent for this?

HP X references:
http://www.sphere.bc.ca/test/hpparts.html#cross

[Specmaster]

So I need to replace this now and the manual shows this as being HP part number 1853-0380 and described it as Transistor PNP SI TO-92 PD=350mw which it clearly isn't as TO-92 is plastic encapsulated and Q7 is a T0-18. Got any ideas as to the equivalent for this?

HP X references:
http://www.sphere.bc.ca/test/hpparts.html#cross

Thanks for this it has helped a bit. Why are these HP manuals so hard to fathom out?

Turned out I was looking for the wrong item as the Q7 transistor I was looking for is located on board A11 not A7 as previously thought, a glimmer of hope I thought and armed with the fact that the transistor bore the new shortened code of 3-0232, which means the HP part number is 1853-0232, yes, I thought I was in luck... Turns out it suffers the same fate, not listed as a spare or any known equivalents. Back to the search again maybe another dead 1740A may yield a working replacement but that can wait a while as it was only a spare board that I broke.

Oh well back to sorting out the blanking and timing issues.

[Specmaster]

Here's where we are at the moment. After reading the manual a bit more (heavy going) I thought before I start pulling boards and then pulling legs out on the parts in order to check their values without others interfering, I spotted a bit about timing controls on board A8 and so set up the calibrator and went through the adjustments.

Sweep length is still a bit long but that might be another adjustment and if I read the manual again, I may spot it. The traces are now much improved and are about 1hz out on the screen and if I can adjust the trace length then it should be possible in conjunction with the trimmers on A8 to get it within spec again. I'm getting the impression that someone has been in here before me and screwed the settings up,

Anyway, as a result of my tinkering tonight, the blanking seems to be working correctly, I get no hint of flyback lines or the trace doubling back on itself so it seems that we're getting there.

I have noticed however that while adjusting these trimmers, the A8 and also the backplane A7 flex a lot and on the slower sweeps the display can flicker and or at times grow dimmer. Gently tapping, or pulling A7 towards me (as I look at the right hand side with screen on the left) with a plastic prodder gives a temporary fix for this. Problem is that there are about 7 plug & sockets as well as 14 switches to be checked and heaps of solder joints that may have to remade to track it down.

So I think that my best plan now is to see if any of the linearity trimers on the horizontal output have any effect and then work from there.

Just as I was writing this report, I turned the scope off (been on for about 2 hours) and switched it back on again, no beam and no sweep vernier CAL light either. Switched it off and left it for approx a minute and switched on again, beam is back, CAL light is off so this may be a duff LED, and everything works as it did before switching off. Does the HV side have a thermal cutout? But why does not switch off when its on? It only fails if I switch it off and back on again?

More work to do another day thats for sure but the light at the end of the tunnel is glowing ever brighter. 

[bd139]

Sounding good! You'd be surprised how much people can f*ck up a scope just whacking the trimmers in it. Does my head in when you get one of them. I think my 465B has been played with in that way.

Unknown with the HV side of things. Might be the feedback network in the inverter that's doing it. I had a similar problem after power cycling on another thing I actually put together myself that happened randomly when it was turned on and off. Very annoying. Turned out to be not enough decoupling.

[Specmaster]

After a couple of days chasing rabbits down holes, I gave in and swapped out the A7 board (main trigger) for the one from the parts mule after I spent a few hours going over it with meters looking to see if I could detect anything out of spec and it all checked out as being reasonably OK.

Powered it up and I now have a trace that is just about there, marginally longer then 10 divisions, 10.5 to 11.5 depending on the range and with my calibrator connected I can my 1 cycle per 10 divisions  and also the 10 cycles, 1 per division on the screen, result. No flyback visible and lots of brightness, which was one of the problems with the other board, it would dim down a lot, poke the board a bit and it would come back for a few seconds. Also after a few minutes, some of the slower speeds would become unusable and only send a sweep every so often, again poking the board would provide a temporary repair.

Over the weekend I'll attempt to calibrate and adjust the input capacitance etc and then this particular scope can out of my repair queue and be put to work on the duty roster along with the others so I can get into the Fluke 8505A a bit and check that out.

I'll post a couple of photos of the 1740A when shes done.

[Specmaster]

Ok the end is here (nearly, more on this later) as you can see from the photo she is performing beautifully a nice steady and sharp trace on the screen, all in cost, includes, scope, a parts mule and other spares £75.

Heres the later bit, the manual is very misleading and following it has got me a trace that is still slightly by 1 division wider then the screen but I now know how to get that back to 10 divisions, if I can be bothered, later. I may just elect to leave where it is, like bd139 said, its good enough considering its age and its probably had a hard life. 

Heres to the next item on the bench, it might be the Fluke 8505A for some tweaks, for some reason, it has a slightly difference of opinion on a voltage depending on if have the black or the red lead connected to ground, only a few millivolts but a difference nonetheless. 

[bd139]

Looking good to me! Nice work

That’s a very sharp trace you’re getting there, arguably better than you get out of a similarly aged tek.

[qu1j0t3]

Heres the later bit, the manual is very misleading and following it has got me a trace that is still slightly by 1 division wider then the screen but I now know how to get that back to 10 divisions, if I can be bothered, later.

Can you describe the problem in the manual for posterity and people finding this thread in future?
Thanks

[Specmaster]

Heres the later bit, the manual is very misleading and following it has got me a trace that is still slightly by 1 division wider then the screen but I now know how to get that back to 10 divisions, if I can be bothered, later.

Can you describe the problem in the manual for posterity and people finding this thread in future?
Thanks

The manual, page 5-11 section 5-52 speaks about setting the beginning of the trace on the far left graticule (far left line) and then using the 2nd graticule line as a reference point for the small highlighted dot during calibration, I took it as meaning the 2nd vertical line, which with 10 columns, or 10 divisions, the 2nd vertical line happens to be the 1st graticule line and the 10th graticule is the last vertical line on the right of the screen.

Given that this is how you set the trace width. You have to set delay on .2 which sets the highlight dot in the first instance 2mm from the far left line, then position this dot, using the horizontal control, on the 2nd graticule, with the delay vernier set on 1.

Then you set the delay to 9, then adjust A7R93 until the dot is on the 10th graticule line (far right), then you set delay to 1 and repeat the process again until delay on 1 set the dot on the 2nd graticule and then setting delay on 9 moved the dot to the 10th graticule. Once done you can move onto other settings etc.

So unless you remember that 10 divisions actually has 11 graticule lines and compensate for that by actually using the 3rd vertical line for delay position 1 setting, you will lime end with 11 divisions in your sweep like I did 

Not really the end of the world because you can compensate using the sweep vernier but you have to remember to this each time your want to approximate a value of the screen (this where digital boys gain over us analogue boys since their scope automatically take frequency readings. I just wonder if any of their scopes will still be working after some 30 plus years, I doubt it. 

I'm going to open her up again and readjust the sweep length.                                             

[Specmaster]

here she is, adjustment this time was spot on and the sweep width is marginally wider, I guess that might be added automatically to allow for drifting slightly? as you can see from the attached photos with various frequencies fed in I was able to get a perfect single sine wave on screen, pixel perfect. The video shows just how much extra wriggle room the scope allows.

I'm extremely happy with the end result, a worthy addition to my scope line up, maybe I'll bring the Hameg 408 back onto the bench and try to resolve its problems? 

https://www.flickr.com/photos/158191416@N07/37942213376/in/dateposted/

[bd139]

Perfect

Go go go with the 408. They're rather nice. Cursors and everything!

[Specmaster]

I know, thats why I got it 

I need to find a cheap source of their header cable though as many of them have suffered and and are not really long enough now to reach.  It had power issues when I got it, its a SMPS and despite many attempts to cure it I couldn't, recapped it, rechipped it etc, nothing doing. Brought a reconditioned PSU for it from their German repair agents, still nothing doing, so I guess there's a real chance that I now have 2 PSU's for it as well.

Its in pieces here at the moment waiting for me to get on with it at component level on the bench to see if I can find any reason for the PSU not firing up, they believe that there's an issue with the feedback circuit preventing it from switching on.

Just been playing around with my other recent acquisition, Hitachi V-525 50Mhz scope which also has cursors and it has this special on Ch1 where it will display either the trace and cursors or the trace as normal and also 3 divisions below, the trace at Mag10. Its a lovely machine, almost new, pristine condition, with its own protective bag and a set of unopened probes, all for £25 and it was local to me, 3 miles. I just did the same test as I did with the 1740A and its spot on spec. 

[Specmaster]

I also checked my Iwatsu SS-5710, no cursors though but does have 2 + 2 channels, a very nice scope actually and it too was spot on as well.

[Specmaster]

Perfect

Go go go with the 408. They're rather nice. Cursors and everything!

This should make you happy, today I ordered some cable that I hope will be what I need for the 408-1A.

[bd139]

Excellent 

Loosely on the subject of Iwatsu, I was looking at Japanese oscilloscopes the other day and Yokogawa have by far the best form-factor I've ever seen for oscilloscopes: http://cdn.tmi.yokogawa.com/DLM2000_Front_image_lg.jpg

Search for more toys continuing here. Nothing yet

[Specmaster]

Wow, that looks great but I bet its a pig to work on.

Talking about working on, I decided that I'd tackle a flickering trace on my Goldstar CRO this morning and when I took the cover off I discovered that a 500 ohm trimmer has fallen apart so thats at the head of the queue for now,  parts have been ordered, thought I had some but they turned out to 5K ones.

[bd139]

I think it's likely disposable if it breaks.

I like the Goldstar/LG units, particularly the OS-9020. Dead simple, really easy to fix and they're slow enough so you can stick pretty much any transistor you have lying around in any hole! Really a design marvel as they used crap tolerance components throughout and still got pretty good performance.  I used one exclusively for about 3 years until I got my hands on a broken Tek 453 in about 1999.