HP 1740A on the bench, help please?

[bd139]

I did it. She's going to murder me

[tautech]

Carefully check that jumper/connector bridge board (A12 ?) for cracked joints on pins and sockets and in the boards it mates with. A 1740A that I had some time back had numerous bad cracked joints on that PCB.

[Specmaster]

I did it. She's going to murder me

And will you be going for a Keithly from Germany? [emoji12]

Sent from my SM-J510FN using Tapatalk

[tggzzz]

Yep, the Tek kit seems to be the industry standard through and they have a very active user group but  they always seem to be breaking down judging by the amount of posts on their forum.

Breakdowns = failure rate * number of items. There are a large number of (old) Tek scopes around.

There are two principal reasons there are a large number around: they were the best at the time (so many were sold) and they have good reliability (so, unlike other scopes, they haven't been junked along ago).

Having not ever used or worked on a Tek, I can't really say. Only ever worked and used ones that I own or owned, those being Rapid (Pintech), HP, Hameg, Goldstar (LG), Iwatsu and Hitachi and the general construction of the HP really impresses me.

I used to work for Hewlett-Packard, and, while being generally impressed with their kit and people, some aspects of the 1740 didn't amuse me.

Two examples are the "floating" boards held in place by the timebase "axle" (shades of IBM PC ISA boards), and the timebase switch on the PCB - which abrades the PCB track when rotated. (Fixable by soldering the obvious wire, taking care not to allow solder to spread onto the contacts)

[bd139]

And will you be going for a Keithly from Germany? [emoji12]

Sent from my SM-J510FN using Tapatalk

Passing on that now

[Specmaster]

And will you be going for a Keithly from Germany? [emoji12]

Sent from my SM-J510FN using Tapatalk

Passing on that now

I might go for one, not sure, a bit rich currently as I'm unemployed. That might be about to change so maybe if another pops up?

Sent from my SM-J510FN using Tapatalk

[Specmaster]

Carefully check that jumper/connector bridge board (A12 ?) for cracked joints on pins and sockets and in the boards it mates with. A 1740A that I had some time back had numerous bad cracked joints on that PCB.

The jumper has been checked, I'll do the same on the boards it connects to as well. It actually looks inside as if has left the factory. It is spotlessly clean and the attention to detail on show is exceptionally good.

Sent from my SM-J510FN using Tapatalk

[bd139]

Never seen a single cracked joint on HP kit myself unless it has been repaired or an ECN applied by a chimpanzee. They're usually very well put together.

[tautech]

Carefully check that jumper/connector bridge board (A12 ?) for cracked joints on pins and sockets and in the boards it mates with. A 1740A that I had some time back had numerous bad cracked joints on that PCB.

The jumper has been checked, I'll do the same on the boards it connects to as well. It actually looks inside as if has left the factory. It is spotlessly clean and the attention to detail on show is exceptionally good.

These always are spotless inside which in a large way attests to their popularity.
I can't now remember all the issues with the one I had....jumper cracks and the vertical output PCB wasn't working.
Then I think the timebase stopped working.....nah can't remember much more.
It was saved from a rubbish skip and I had to get a donor unit from the NZ HP agent as I didn't know about Sphere way back then. It fixed up fine, a pair of new probes, manuals on CD and down the road it went, donor and all.
Now I have another to fix that donor could've come in handy. 

Never seen a single cracked joint on HP kit myself unless it has been repaired or an ECN applied by a chimpanzee. They're usually very well put together.

They crack all right, those jumper boards get put under heaps of stress prying them on and off. The factory solder joint balling could be described as miserly. Maybe they were short of solder that shift. 
I'd make a practice of reflowing them as a matter of course.

[Specmaster]

Carefully check that jumper/connector bridge board (A12 ?) for cracked joints on pins and sockets and in the boards it mates with. A 1740A that I had some time back had numerous bad cracked joints on that PCB.

The jumper has been checked, I'll do the same on the boards it connects to as well. It actually looks inside as if has left the factory. It is spotlessly clean and the attention to detail on show is exceptionally good.

These always are spotless inside which in a large way attests to their popularity.
I can't now remember all the issues with the one I had....jumper cracks and the vertical output PCB wasn't working.
Then I think the timebase stopped working.....nah can't remember much more.
It was saved from a rubbish skip and I had to get a donor unit from the NZ HP agent as I didn't know about Sphere way back then. It fixed up fine, a pair of new probes, manuals on CD and down the road it went, donor and all.
Now I have another to fix that donor could've come in handy. 

Never seen a single cracked joint on HP kit myself unless it has been repaired or an ECN applied by a chimpanzee. They're usually very well put together.

They crack all right, those jumper boards get put under heaps of stress prying them on and off. The factory solder joint balling could be described as miserly. Maybe they were short of solder that shift. 
I'd make a practice of reflowing them as a matter of course.

Yes, it would be better to do that while it's open, some of those boards are hard to get out.

Sent from my SM-J510FN using Tapatalk

[Specmaster]

Resistors have been replaced, all the soldered plugs / sockets as suggested have been checked on the boards that I have had to remove, ie. the power, the interconnecting and the vertical output boards, all are OK.

Reconnected and powered up now I have lost all vertical inputs and control so the flat trace cannot be moved up or down by the vertical position pots so naturally I retraced my steps thinking its something I've done wrong, all seems OK, measured volts on the vertical output board and everything there is OK apart from the schematic shows 0.5v on the collectors of Q1 and Q3 and I have 1.1v but have been advised thats OK.

Measured the incoming signal from the end of the delay line and it is constant at 11.4v regardless of the pot positions so my suspicions have swung toward the pre-amp and I discovered that the Q1 and Q3 on that board have their base soldered in a common joint with other parts but the collector and emitter are socketed (see photo) and on channel A the emitter and collector pins are a very loose fir in the socket tubes while on channel B, the transistor has already been replaced with another in T0-18 metal case, pin outs are in a different arrangement and emitter and collector again pushed into sockets but the leads have been bent to suit and were almost touching but loose in the sockets.

Surely this cannot be right? Is it ok to solder these in or not. Note that I will be replacing these anyway with new ones as I have no faith in them being matched. The one in the photo is capable of 100Mhz but cannot vouch for the other as it is crusty I cannot read the part number.

Thoughts please, could these 2 transistors be the root of my problem?

[tggzzz]

Reconnected and powered up now I have lost all vertical inputs and control so the flat trace cannot be moved up or down by the vertical position pots
...
Thoughts please, could these 2 transistors be the root of my problem?

Maybe maybe not. Maybe the BNC connector has become disconnected (I doubt it, but your tests haven't removed that possibility).

You have started at the output end of the signal chain and found the (differential) collector voltages don't change. Conclusion: somewhere between the BNC input and where you are measuring, there is a fault.

Now start at the input end and see where the (differential) collector voltages do and don't change. If the voltages change at a point, the fault will be between that point and the output.

[Specmaster]

The BNC connectors are connected just fine, I've decided to replace both these transistors as the B channel is not the same model as the A and also in trying to make the legs engage the socket better it has become broken not the one in the photo, its the T0-18 cased one. I can't see why the base leg would be soldered and leave the other legs pushed into badly fitting sockets, makes no sense to me. Surely if these transistors were prone to failure etc, that they would have had all legs socketed to allow for easy replacement?

Prior to replacing the resistors on the output board, I had working traces so the only thing I can think of is that I did go over the preamplifier stage with a multimeter checking for any caps short circuiting on the ohms scale and testing connections between any plugs/sockets so I must have disturbed these transistors in the process.

I have now located some of the correct transistors Motorola MPSH81 (HP had their own custom part number printed on these) and these have been ordered. I'll try and replace them in the same manner as they were fitted but the sockets only for 2 legs still mystifies me.

[tggzzz]

I've decided to replace both these transistors as the B channel is not the same model as the A and also in trying to make the legs engage the socket better it has become broken not the one in the photo, its the T0-18 cased one. I can't see why the base leg would be soldered and leave the other legs pushed into badly fitting sockets, makes no sense to me. Surely if these transistors were prone to failure etc, that they would have had all legs socketed to allow for easy replacement?

Personally I prefer to determine which component is faulty before starting to replace components in the hope they might make a difference. Start at the BNC socket and measure the voltages along the signal chain; RTexcellentM.

[Specmaster]

I've decided to replace both these transistors as the B channel is not the same model as the A and also in trying to make the legs engage the socket better it has become broken not the one in the photo, its the T0-18 cased one. I can't see why the base leg would be soldered and leave the other legs pushed into badly fitting sockets, makes no sense to me. Surely if these transistors were prone to failure etc, that they would have had all legs socketed to allow for easy replacement?

Personally I prefer to determine which component is faulty before starting to replace components in the hope they might make a difference. Start at the BNC socket and measure the voltages along the signal chain; RTexcellentM.

The reason I have done it this way is because as you will see from the schematic Q1 and Q3 are the start of the chain and with Q3 already being of the wrong type and broken, there is little to test, so my logic is to correct what I can see as faulty, and as it is effectively the start of the active devices, it will affect everything else downstream from that point? I can't believe that a resistor or cam switch on both channels would fail at the same time, but both of these transistors can be demonstrated to have either no connection or bad connections, depending on which channel you're checking?

[tggzzz]

I've decided to replace both these transistors as the B channel is not the same model as the A and also in trying to make the legs engage the socket better it has become broken not the one in the photo, its the T0-18 cased one. I can't see why the base leg would be soldered and leave the other legs pushed into badly fitting sockets, makes no sense to me. Surely if these transistors were prone to failure etc, that they would have had all legs socketed to allow for easy replacement?

Personally I prefer to determine which component is faulty before starting to replace components in the hope they might make a difference. Start at the BNC socket and measure the voltages along the signal chain; RTexcellentM.

The reason I have done it this way is because as you will see from the schematic Q1 and Q3 are the start of the chain and with Q3 already being of the wrong type and broken, there is little to test, so my logic is to correct what I can see as faulty, and as it is effectively the start of the active devices, it will affect everything else downstream from that point? I can't believe that a resistor or cam switch on both channels would fail at the same time, but both of these transistors can be demonstrated to have either no connection or bad connections, depending on which channel you're checking?

If there are bad connections, fix that. But you said it was working earlier, so why has the connection become bad? (Q1 & Q3 don't need to be matched; they are independent emitter followers)

If you apply 1V to the BNC, what's the voltage at Q2A gate (or Q1 base) and TP9 for various input attenuator switch positions? Do the voltages vary as expected when you change the input voltage to 2V?

[Specmaster]

I'll check that out later but perhaps the reason for the failure is that I did go over the preamplifier stage with a multimeter checking for any caps short circuiting on the ohms scale and testing connections between any plugs/sockets so I must have disturbed these transistors in the process?

It is clearly something that I've done to cause it to fail. I have not removed the board or poking around the panel controls at all but I did check the caps on the component side of the board.

[tggzzz]

I'll check that out later but perhaps the reason for the failure is that I did go over the preamplifier stage with a multimeter checking for any caps short circuiting on the ohms scale and testing connections between any plugs/sockets so I must have disturbed these transistors in the process?

It is clearly something that I've done to cause it to fail. I have not removed the board or poking around the panel controls at all but I did check the caps on the component side of the board.

Randomly checking for caps shorting is something that should only be done after you've observed a specific signal voltage that doesn't vary in a way that might be expected if a cap has shorted.

What's the maximum voltage your multimeter outputs when testing components? How does that compare with the Q2 FETs' max ratings? Could ESD have damaged the FETs. They might be a pain to replace, since they are a dual matched FET.

[tggzzz]

If you apply 1V to the BNC, what's the voltage at Q2A gate (or Q1 base) and TP9 for various input attenuator switch positions? Do the voltages vary as expected when you change the input voltage to 2V?

I've just got out my copy of the service manual, and p8-18 (the one before the schematic you've included) describes how to setup that test.

It is a good manual; use it.

[Specmaster]

Using another scope to observe the waveforms? I thought that page was about checking the calibration?

[tggzzz]

Using another scope to observe the waveforms? I thought that page was about checking the calibration?

Use your imagination, to glean all the info you can; it is telling you what the circuit does when it is working correctly. What more do you need?

Reduce the frequency to 0.01Hz, and use a multimeter.

[Specmaster]

Ok, heres the latest, ordered up some replacements MPSH81 transistors as Q1 has been replaced at some point with another that the part number is not quite readable but looks like BVC 96305 (metal can type TO-18) and the correct one should be a TO-92 package. Anyway, I'm unable to find out any data on it, plus it did not seem to be making good contact with the socket tubes, partly because the legs had been screwed round to try get them in the correct order to match the circuit.

Checking the pre-amplifier board in more detail against the schematic that there was a 12v +&- supply that was not apparent on the power board. This is derived from the +&- 15v rails on the pre-amp board and the -12v was reading -1.89v caused through Q17 (2N3134) deciding that it has more in common with a diode than a transistor. It is listed as obsolete as are the equivalents that I have found for this.

Does anyone know of a suitable replacement , or a supplier at reasonable cost for this?

[bd139]

Assume this is just a TO5 canned pass transistor based on the application and characteristics (I haven't looked at the schematic). 2N2905A is a good sub for them. CVC sell them for a very reasonable £1.90 for two: http://www.ebay.co.uk/itm/262085968304

I just had a dig around in my bags of transistors and I've only got complimentary 2N2219's typically or I'd send you one. Need to buy more transistors!

Edit: just had a look at the SMM and a 2N2905A will be fine. There's only 2v across C-E junction by the look so it'd take a lot of current to blow that. Check that whatever is on the -12v rail is not pulling it down to -1.89v otherwise that will drop quite a few volts and go phut! In circumstances like this I tend to pull the transistor and stick some clip leads across the test points (TP4 and TP3) and connect it to a bench supply with current limit and bring it up slowly just to make sure it's not a problem further up. Otherwise you end up blowing a lot of silicon fuses very quickly. This isn't heavily loaded by the looks so whatever blew it may be present still.

[Specmaster]

I did a check with my multimeter and the resistance on both +12 with the other complementary transistor in circuit is the same as the -12 with the duff transistor pulled, 660 -ohms, so there's no problem on the remaining circuit except for the fact that I have pulled out the BVC one ready for the new one to be fitted when they arrive next week.

I had already checked with CVC as they list the 2N3134 for £4.50 BIN or offer plus £2.50 economy postage. Seeing as they are only 16 miles from me I offered £3 and collection, they countered at £4.50 and collection which I accepted and paid. Checked my Paypal receipt and they still took postage!
I messaged them about a refund of postage, if they wont, then I'll ask for a couple 2N2905A collected at the same time, otherwise they can cancel the order and I'll somewhere else.

Thanks for your very kind efforts.

[bd139]

They're pretty good fighters in that department. I spent a week arguing with them over a single FET

Good luck