HP 1740A Channel B not working and flickering on A

[lowimpedance]

Certainly looks like something has gone 'pop' !  .
Did you set up the controls as per step 1 and 2 on page 8-26 of the manual before measuring ? , repeat if not.

[PurpleAmaranth]

Yup I set them up exactly as indicated.

[lowimpedance]

 Looking at the voltages I have a suspicion that Q6 is toast and I cant find a cross reference to a recognizable part number   
Maybe email Sphere electronics and ask if they have one using the HP part number 1854-0019, otherwise you may have to make an educated guess at a replacement, but which specs were important  .
carefully remove Q6 and bodge in a generic  NPN, 2n2222 etc , watching out for correct orientation and see if any difference in voltages at least.

Whats your level of experience with electronics, new/hobby etc looking to set up a small collection of gear ?. Maybe look at a digital oscilloscope. Plenty of discussion on the forum for the hobby end, RIGOL / SIGLENT etc
These old boat anchors can take some effort to get running and maintain which of course is a great way to learn but a balance needs to be found between learning
and getting on with other projects.
Not to say this unit is a goner yet !. Perhaps another parts unit from this series might turn up and using bits from each resurrect a good one. I have resorted to this method myself 

[PurpleAmaranth]

I don't have any transistors at all, so I ordered some 2N2222 which I'll go pick up tomorrow, I'll let you know the voltage measurements with that replaced. I'll then email Sphere asking if they have a replacement for 1854-0019.

Whats your level of experience with electronics, new/hobby etc looking to set up a small collection of gear ?. Maybe look at a digital oscilloscope. Plenty of discussion on the forum for the hobby end, RIGOL / SIGLENT etc
These old boat anchors can take some effort to get running and maintain which of course is a great way to learn but a balance needs to be found between learning
and getting on with other projects.

I don't have much experience at all, really new to all this. But I'm studying electrical engineering, so I figure I really need to get acquainted with the practical side of things. I've read lots of good stuff about some RIGOL models, and I'm sure the same is true for SIGLENT. But although they're 'only' about $400 or so, I really can't afford that at this stage.
I'm really enjoying working on this a lot so far, although you've (thankfully!) done most of the actual analysis. But I've learned lots already, from replacing that diode to reading circuit diagrams.

I really hope we manage to fix it.
I have been looking for parts units but haven't found much yet that's not more than triple what I paid for this one.

[PurpleAmaranth]

Alright, so things got busy and I didn't get to replacing Q6 with a 2N222 until today. But now it's in, with a pretty questionable soldering job... but at least it's in.
Measuring the voltages yields the following:

Q1: E 1.652V (1.7V), B 1.316V (1.4V), C 0V (0V)
Q2: E 1.652V (1.7V), B 0.97V (1V), C 0.643V (0.7V)

Q3: E 0.643V (0.7V), B 0V (0V), C -0.738V (1V)
Q4: E 1.601V (…),  B 0.89V (1V), C 0V (0V)

Q5: G -0.739V (1V), S 14.67V (…), D 0.684V (0.8V)
Q6: E 0V (0V), B 0.685V (0.8V), C 0.887V (1V)

Q7: E 13.42V (13.3V), B 12.69V (12.7V), C 0.889V (1V)
Q8: E 1.586V (1.5V), B 0.889V (1V), C 0V (0V)

Q9: E 0.95V (1V), B 1.584V (1.5V), C 14.61V (14.6V)
Q10: E 3.994 (3.5V), B 0.434V (0.5V), C 14.61V (14.6V)

Q11: E 0.164V (0.2V), B 0.382V (0V), C 14.84V (15V)
Q12: E 0.802V (0.8V), B 1.473V (1.5V), C 3.993V (3.5V)

Q13: E -9.68V (-9.7V), B -9.13V (-9.2V), C -0.74V (1V)

Thee almost all look much better to me

Testing the old transistor out of the circuit with the DMM yields the following:
EB OL, CB OL, BE OL, BC 0.63V, CE OL, EC OL
Looks like the Emitter-Base junction is 'dead', right?

There's still no trace on the screen, but the RESET light is now ON when the settings are selected as per the A8 schematic section of the manual.

I'll email Sphere to see if they have a proper replacement for Q6. Anything else to check or do? Or just try to find a suitable replacement for Q6 for now, and then see where that takes us?

[lowimpedance]

Q6 is toast, Having a further look I find some mention that its specs are similar to the 2n2369 which is also obsolete but that gives a starting point to choose an alternative.
See what Sphere recommend but remember they may be rather pricey !. Otherwise look up the 2n2369 and get a similar spec. type. ie, Vceo > 15V (I would aim for 30V and up), approx. 350-500mW PD, an FT~500Mhz and if in a TO-18 package great but I guess a TO-92 will also do.

As for the voltage readings things look better, but noted the -ve collector voltage on Q3 is that correct?.
Does that voltage change when you are sure you have supplied a trigger input?.

Note damage to Q5 may have also occurred which is not as easy to replace (2n5245), may be Sphere can help here as well.

[PurpleAmaranth]

Alright, I'll take a look at the 2N2369, thanks for that.
What do you mean by "when you are sure you have supplied a trigger input"?
And yup, -ve collector voltage on Q3 is correct, I measured several times and it's definitely -ve.

[lowimpedance]

An input signal applied and the trigger source selected for that channel and the trigger level such that the trigger generator will trip thus causing a sweep.
Although in Auto a free run a sweep should be visible. Analysis would be much easier with another oscilloscope to see what the trigger signal is doing P/OXA7 etc

[PurpleAmaranth]

Found someone not too far from me selling 2N2369A left over from a project. I'll see if I can get some there.

Could Q5 be the cause of the -ve voltage?

I see what you mean about the trigger, I'll go check it. I really wish I had another scope to check all this, I'll see if I can borrow one but I seriously doubt it.

[lowimpedance]

I'm not sure the 2369 is the best choice as its Vceo is only 15V which also happens to be the VCC rail !. At this stage I would probably leave the 2222 in for further fault finding.
Its possible that Q5 is busted which would prevent the sweep ramp from happening when a trigger signal arrives at Q1 base. Which is why we need to see it with an oscilloscope.
Being able to eliminate any failures on PCB A7 would be useful/desirable.
I suggest reading the circuit operation details in the manual for further info. Section 4.

[PurpleAmaranth]

Not much luck getting anyone to lend me a scope. However, Sphere replied and they have some 1854-0019 that they can send me to replace Q6 with. I haven't looked at section 4 yet though, I'll do it this evening. Is there anything else we can do without a second scope? Should I ask Sphere for a replacement to Q5 too? Should I try desoldering Q5 and checking it with the DMM out of the circuit?

[lowimpedance]

Not much luck getting anyone to lend me a scope. However, Sphere replied and they have some 1854-0019 that they can send me to replace Q6 with. I haven't looked at section 4 yet though, I'll do it this evening. Is there anything else we can do without a second scope? Should I ask Sphere for a replacement to Q5 too? Should I try desoldering Q5 and checking it with the DMM out of the circuit?

Okay I took the cover off my 1740 and measured the gate of Q5 and mine was also negative !?! so don't worry at this stage with Q5 we'll assume its still ok.
Next thing to try is select single sweep at say 1/2 second TB speed and press reset (LED ON). Connect your DMM to R1 on A8 (which is the Trigger in from A7) and you should read approx. 1.3V while measuring trigger a sweep with the trigger level pot (turning back and forth a good amount  ) to see if the trigger line goes low to around 0V.

[PurpleAmaranth]

Okay I took the cover off my 1740 and measured the gate of Q5 and mine was also negative !?!

That's interesting! I wonder if there's an error in the manual...

Anyway, I selected single sweep at .5sec, pressed RESET, and LED turned on. Then on both sides of R1 I get about 1.32V. Triggering a sweep by rotating the trigger level pot, I see the dot on the screen go from left to right, and both sides of R1 go down to about 0.03V. And once the sweep is done, we're back at 1.32V, and RESET light is OFF. So looks like Trigger in from A7 is fine. Does that mean the problem is necessarily on A8?

[lowimpedance]

 That looks consistent with mine, next measurement would be to look at the collector of Q6,(TP1), to see if there is a positive going ramp up (DMM should show something   , better with slow timebase speed.).
I suspect there is as you state you now see a trace scan, a 'dot' is to be expected (there may be a slight tail depending on the brightness setting). What do you see?....Rememver to press reset between each trigger attempt  .

Does that mean the problem is necessarily on A8?

The problem could still be to do with the switch at higher timebase speeds, so if TP1 is good and you see a sweep then the next thing to do is work your way through each timebase setting one at a time from slow upwards till you no longer get a trace (make sure your brightness is set appropriately, ie more as the sweep speed is increased.).

[PurpleAmaranth]

Alright, so I measured all the DC TP1 voltages, and checked for a sweep. Below are the results, but to summarize quickly, RESET lamp turns ON every time, then OFF once the sweep is triggered, and a trace scan appears on screen for all timebase settings.

2 SEC: DMM starts at 0.885V, goes steadily up to between 11 and 12V, then suddenly switches back to 0.885V. Dot was slowly making its way across the screen from left to right while the DMM reading was increasing.
1 SEC: Same as for 2 SEC, started/ended at 0.890V
.5 SEC: Same, start/end at 0.898V
.2 SEC: Same, start/end at 0.918V
.1 SEC: Start/end at 0.947V, getting too fast to tell how high DMM reading goes during the sweep, but definitely increasing

50 mSEC: Start/end at 0.997V, too fast to even tell if DMM reading is increasing or decreasing, but still changes
20 mSEC: Start/end at 0.882V
10 mSEC: Start/end at 0.889V
5 mSEC: 0.898V
2 mSEC: 0.919V
1 mSEC: 0.948V
.5 mSEC: 1.001V, 'dot' moves so quickly it looks like a line
.2 mSEC: 0.881V
.1 mSEC: 0.888V, had to turn up the brightness setting

50 µSEC: 0.896V
20 µSEC: 0.916V
10 µSEC: 0.944V, turned brightness up all the way
5 µSEC: 0.994V
2 µSEC: 0.880V
1 µSEC: 0.888V
.5 µSEC: 0.897V
.2 µSEC: 0.917V
.1 µSEC: 0.945V, had to lower background brightness to still see the scan
.05 µSEC: 0.997V

Looks like the sweep is working, right?

[lowimpedance]

So at faster timebases you see a normal trace as expected ?. There will be a point reached rather quickly where the DMM will not see anything happening at TP1 and then a oscilloscope is needed.
It therefore looks to be working, can it display the cal signal ?. What is the current fault ?.

[PurpleAmaranth]

No I only see the moving 'dot', as if it's scanning, which looks like a line for fast timebases. But as soon as it's done scanning, the screen goes blank (just the green background), whereas before there used to be just a constant horizontal line when no input was applied. Using the CAL as an input does not change anything, nothing is displayed.

Also, I should be able to get access to a modern and functional oscilloscope for a couple hours next week; if we still haven't sorted this out with just the DMM.

[lowimpedance]

 Is single sweep still set ? put the scope into auto/continuous and apply the cal signal to channel A and trigger source from A, what happens ?. 
BTW the green background is not needed, scale illumination to minimum, unless that is not possible and also a fault.

Its getting rather hard to diagnose remotely. You might need to post another video showing the controls and screen.
 

[PurpleAmaranth]

Well, I feel pretty dumb... It turns out I forgot to put it back in AUTO mode, so it actually works just fine. Looks like it was just Q6 that had to be replaced, sorry for wasting your time with the rest. I'm in the process of ordering an original Q6 replacement from Sphere, but the 2n2222 seems to be working pretty well, the trace looks good! I've attached a picture.

So the one thing we have left is the flickering for timebases larger than 1mSEC. You'd said that there wasn't much more we could do with the DMM about that. What should I try looking at with the second scope which I'll be able to use for a couple hours later this week (hopefully) or next week?

[lowimpedance]

 Ahh looks much better, I would certainly use the second oscilloscope to check all the time base circuit test points comparing to the manual. Do this at time base settings that produce a good scan and those that cause flickering. While monitoring those test points use an insulated tool to prodd/flex the suspect PCB area's and observe what happens on the 1740 screen and on the monitor CRO. Take pictures with your phone for records etc.  Trigger the 1740 on its cal signal BTW. Also use 10:1 probes on the second oscilloscope which have been properly compensated to its input with its own cal signal !.

 This will be a good learning exercise on using a CRO as well, so look at the 1740 manual circuits for the timebase/trigger area etc and read the relevant manual circuit operating principles before hand and make a note of the test point waveforms you want to measure etc so your ready to go when the other oscilloscope becomes available. Have fun !

Once the 1740 is working properly a good clean and it will be as good as new  .

[PurpleAmaranth]

So apparently what I've been calling flickering is actually perfectly normal and to be expected on any analog scope.  I went to the lab to try probing with a second scope, and I was told there was no point at all, and that nothing is wrong with it. So I guess that's good, although I wasted a lot of time on it in that case. But at least I learned a lot. Thanks for all the help! Everything else seems tow work on the scope as well.

[lowimpedance]

Excellent  , I was having some difficulty with the fault description and was thinking along those lines too. Trying to diagnose more esoteric issues remotely can be a challenge!. Nothing beats being able to have hands on.
 While it seems time might have been wasted, it actually was valuable learning of many aspects of the fault finding/repair process which will useful for all future work, whatever that may be.
As for the HP, I would give the Low Voltage Power Supply a one over for dry joints etc particularly around the transformer connections and the bridge rectifiers, generally where you get high thermal cycling and/ or mechanical stress. Also if you ordered a replacement for Q6 put that in too.
And lastly a nice clean and a blue screen filter   . ...........Then put it to good use !.

 

[PurpleAmaranth]

I somehow missed this reply until now, months later. I'll get the blue screen filter, and check for dry joints. I didn't end up buying the replacement for Q6, but it's been working great since June with the 2n2222, so I'll leave it at that for now. Thanks again for all the great help!

[scopeman]

Just a short comment. Enjoy that scope. I have a mint one that was willed to me by a good friend. I was with him when he placed
the order for it with Jack Molcan the local HP Salesman with HP when HP was a real company. I has also bought my first HP calculator an HP 45 (I still have it) from Jack.

My friend paid over 2K for the 1740A and Jack delivered it to his house when it came in.

I pull it out and use it once in a while. It is still in cal and works like new. Real HP quality. I still have the original shipping box.

It's a great scope and a great teaching tool as well.

All the best!

Sam
W3OHM