HP 3335A Level Generator Repair

[Tony_G]

Hi All,

I thought I'd spin up a thread with my videos of this repair in it because several people have asked me privately for help and I'd like to make sure that all the discussion is publicly searchable and not locked in an email/pm silo.

Anyway here we go....

In the first video I looked at the unit and started poking around and checking the power rails on the motherboard

https://youtu.be/RaC2mzWbaFc

For Part 2 I wanted to see if the issue was in the power supply or in the motherboard. I'd never taken on of these units apart before so it has some cool construction for working on the device (this is covered in the manual but hey who reads that). I pretty much ruled out the motherboard as it was not placing a load on the power supply. I thought it might be the cards but it wasn't.

https://youtu.be/EDAzsNKGAMY

With Part 3 I grabbed the working motherboard assembly from my other 3335A (I got 2 as part of the great auction score of 2017) and verified that both motherboards works in the working unit so the issue has to be in the power supply. The supply is made up of 2 boards, the A15 pass transistor board and the A14 power supply board. You'll notice that I check out the current limiting function of power supply at 12:20 

https://youtu.be/-eItJ8GID3I

Now I didn't get the 15V rail back to 15V - I think the issue is that the previous repair swapped out some components. This unit has a funky power supply design with a capacitor/diode voltage doubler being used to bias the pass transistors and various feedback into those transistors. I'm going to order the right parts per the HP manual and replace them but for now the unit "works" (for a given value of work) so I'll get it off the bench and wait for the parts to arrive. This will let me get back to working on some of my other stuff (alternatively I could expand the lab either into the bedroom or garage for more bench space but in either case I suspect that my wife won't be happy  ).

https://youtu.be/vo0oV0zLgQs

So overall an interesting instrument and repair - It really needs to be restored but as I already have a working unit I'll be able to put it aside until my next vacation.

TonyG

[wn1fju]

I've had some problems with my HP 3335A over the years.  Besides replacing most of the electrolytics on the P/S boards, and a few on the motherboard
itself, mine has managed to blow 3 out of the 4 pass transistors individually at one point or another.  I would suggest that the P/S is inadequately cooled
on these units.  Also, mine would occasionally not startup, and that was traced to the processor getting no clock.  That in turn was traced to the 40 MHz
master oscillator that needed a little tweak to the adjustment pot, as I guess components did age over the decades. 

But I really like the 3335A and its accuracy is pretty wonderful.

[Tony_G]

Thanks - Did you ever end up in a situation where you couldn't adjust the 15V rail into spec? I suspect this is due to the caps & diodes but I haven't checked yet to see if the previous owner modified the 15V reference voltage divider.

[wn1fju]

I don't recall having any adjustment problems on the P/S.  It was obvious when the voltages were sagging that something critical was wrong.
After all the capacitor and pass transistor replacements, things adjusted well if I remember (it was a while ago).

It is a strange beast, however, with the whole inner carcass lifting out of the unit.  I don't recall ever seeing another HP piece of that era
quite like it.

[Tony_G]

It is a strange beast, however, with the whole inner carcass lifting out of the unit.  I don't recall ever seeing another HP piece of that era
quite like it.

Thanks - I completely agree - I was 100% with HP doing the gull wing model for serviceability but this just seemed strange - If I had to cast unfair aspersions, I'd say this was a new PHD/Masters engineer showing how cool they could be - Which is strange given how grounded HP was with respect to design. I should check to see if there was a HP Journal article on it so I can apologize for that (To step back, I'm generally have great respect for how the HP guys designed and built gear. It was clear they really knew what they were doing but this one just seems weird for no reason).

[Ordinaryman1971]

Well, apart from the whole chassis lifting out it’s still pretty standard HP design. Everything enclosed in aluminum cans covered with more aluminum and screwed together with a 101 screws.
I like it. Maybe there was a reason for this approach. Faster field serviceability... possible military application. We would probably never know unless there is somebody that was a part of this project that would tell the story.
I would like you to continue the project into the fronts panel and all those stubborn buttons. Great piece of gear.

[Tony_G]

Agreed - In general the build quality is fantastic and this is why I was saying "unfair aspersions" - The circuit design just seems a bit weird for no apparent reason.
I'm planning on getting back to it but I need to get some other stuff on the bench in the meantime.

TonyG

[basilio]

dear Tony
a few months ago I wrote you a note in the video you published on YOUTUBE
concerning the repair of the HP 335A generator. I found
of my interest the four videos published as they are grappling
with the repair of this generator. Also following your directions are
come to the conclusion that the main failure lies in the power supply. In particular
the first problem that I find is:
the voltage of + 15 V drops when it supplies the downstream circuits.
I adjusted the vacuum voltage to +15.00 using the potentiometer R10, but under load
it goes down to about 13.5 V. I noticed that this happens when the current exceeds 1 A.
I tried to change the threshold using the R25 potentiometer, but nothing happens.
I do not understand how the U1A operatives, for the regulation of the current threshold, and U1B for the regulation
of the voltage work so linear without any feedback. In particular in the case of U1a when the tensions on the two vary
inputs (2-3) the output (1) either goes to + 20 V or to -8 + 5.11 = -2.89 V ie we do not have an output
linear when the inputs change. The same thing happens for U1B. Is this so?
I ask you, kindly you can explain to me the operation of the relative circuit of +15 v.
I thank you and I salute you.
Basilio

[Tony_G]

Hey Basilio - I'm, more than happy to help but I'm on vacation in Aus right now - Can you reply to this thread in a week or so when I'm back in Seattle and have access to my gear? I'd really like to help so please get back in touch.

TonyG

[basilio]

I thank you for your reply, thanks again and happy holidays.
basilio

[basilio]

Dear Tony, in July I wrote to you asking for clarification for the power. I solved all the power problems by replacing all the electrolytic capacitors of the A15 board. But still HP3335A does not work. Now the situation is as follows: on the display appears only the word UNLOCKED and the LEDs DS5 and DS6 are lit, which are next to the sweep buttons (start singol 10 sec and start sing 50 sec) and the LED next to it is on to the MANUAL TUNE knob.
After about 5 minutes the word UNLOCKED disappears and the aforementioned LEDs remain on.
Please give me some advice on how to move forward.
I thank you and I greet you cordially.
Basilio

[Tony_G]

Unlocked is expected - That is just the oscillator warming up and locking.

From the sound of it I think your problems are on the A13 controller card. I'd check the peripheral bus signals before and after U20 to see if the micro is working and the interface adapter/driver IC is working as well. From there you're just going to need to go over the schematic and test things like RAM, ROM, Clock etc to try and find what isn't working on the controller.

TonyG

[basilio]

Thank you, I will keep you informed of developments
Yours sincerely

[basilio]

.....the story continues....
dear Tony
casually I discovered that the micro was blocked I think because the start-up did not happen correctly and then to the
micro did not get the reset. There will be some problems in the U12A, U12B and U21A circuit of the A13 board (controller).
Waiting to resolve the anomaly I gave a manual reset putting for a moment on the ground the pin 3 of U12A (RST).
As a miracle, the display appeared correctly indicating 1000000 Hz and a level of -86.98 dbm.
I tried to change the level and everything was ok. On the output connector, with 0 dbm on the oscilloscope we could see one
beautiful sinusoid. Changing the level the sine wave changed size, perfect everything ok. I thought I had solved everything, except
of the small problem of reset ..... but no !!!!
As soon as I tried to change the frequency, even a few hz, both from the keyboard and manually with the tune knob,
while the display indicated the new frequency on the output connector the marked disappeared. This even if I set
the frequency again at 1 MHz. Nothing went out anymore !!!
The only way to review an output signal was to reset the micro. In doing so, outgoing one had to
new the signal of 1 MHz and the level of -86.98 dbm, but as soon as we tried to change the output frequency
the signal disappeared and instead the display indicated the new frequency and the level that had been set,
but outgoing nothing !!!!
Strange thing by scrolling all frequencies with the tune knob (but also with the keyboard) at some point, when
the display indicated 14 MHz (approx.) in the output a signal of about 22MHz appeared and continuing to change frequency
at 44 MHz (approx.) a lower level signal was again (maybe a 2nd harmonic?).
I hope I have clearly illustrated the anomaly and ask for your opinion.
Excuse me if I disturb you continuously but given your experience I'm sure you can give me good advice.
Best wishes

[Tony_G]

Happy to try and help. I would fix the issue with the start-up circuit first then step through the flow chart that is part of service group O. I suspect that something is going wrong with how A13 is working and that is causing the issue. It should be providing a BCD value that controls the VCO so you should check that from service group N.

Let me know how it goes.

TonyG

[basilio]

ok, grazie
ciao Basilio

[basilio]

Dear Tony,
can you tell me if the signals that are on the J2 connector of the A10:

- SER pin 18
- LSH pin 17
- FS3 pin 16
- FS1 pin 15
- FS2 pin 14
- NF1 pin 10
- FIV pin 9
- FDC pin 8
- NF2 pin 7
- NF4 pin 6
- NF8 pin 5
- NLCH pin 3
- NSH pin 2

are signals that can be displayed with the oscilloscope, ie square waves or logical levels (0, 1)
or are impulsive signals of short duration and not viewable?
thanks bye

[Tony_G]

Interesting question - I've never actually looked directly at them but looking at the schematic they're coming from 74 series logic chips so they should be digital (I.e square pulses). You should be able to trigger on them and see states.

Are you getting nothing? Might be worth attaching anLA to them and see what happens over time.

TonyG

[basilio]

explain to me better what you mean by "anLA", thank you
Basilio

[Tony_G]

Sorry - Typo. "anLA" was supposed to be "a LA" or "a Logic Analyzer"

TonyG

[basilio]

I thank you for the answer.
The problem is that I do not own an LA !!

[basilio]

Tony, I'm sorry if I disturb you continually, but I think your directions are essential.
I ask you: A-B-C pins test of A13 controller assembly are normally connected to ground?
ciao

[Tony_G]

No they should be disconnected and left to be pulled up to 5V by the pull-up resistors.

TonyG

[basilio]

OK thanks
Basilio

[basilio]

dear, every now and then I put my hands back on the hp3335a generator that is still broken.
I could not go on due to lack of free time.
Since I think the ROMs of the A13 could be broken, from the KO4BB site
(Http://ftb.ko4bb.com/getsimple/index.php?id=manuals&dir=HP_Agilent/HP_3335A_Synthesizer)
I downloaded the ROM images that have no extension.
Now I ask, since the eprom programmers I know accept files of the type .bin or .hex
the files that I downloaded as I have to treat them?
I always thank you for your patience and I greet you cordially