HP 3325B Function Generator [FAIL 023]

[Bakafish]

Hi guys. I got unlucky with a unit I bought at auction, which arrived today with FAIL errors on startup. 

Who I am:
I'm a newbie to this stuff, I got lucky with a pair of Agilent 663XB Power Supplies and rolled the dice on an auction thinking this would complete my basic electronics set. I paid more than I should of, because I like the look and feel of the classic gear and thought this would be a better choice than a cheap import box.

What I have:
The janky service manual PDF from the Keyspan website, a 200 MHz 2 channel scope, several good multimeters and the desire to mitigate the loss of $400, but I'm going to need a good deal of help from the community.

What I know so far:
Seems like a common error with these units based on a quick google, but I didn't find a clearly documented fix. I'm assuming it will be an adventure to isolate.
I saw another link on this forum with that unit showing clear damage to lots of resistors in the output section, but my visual inspection shows nothing out of sorts with the same board, and no obvious signs of damage anywhere else. The sine, square and triangle wave modes all show " - CAL -" followed by "FAIL 023" then display the frequency.
The Saw tooth modes show "FAIL 023" followed with a "FAIL 041" that stays present.

Requesting a 1000 Hz sine wave at 1 volt of output shows the waveform clipped at the top and offset ~5 volts positive as near as I can tell. The other waveforms appear to be clipped and offset similarly. Adjusting the DC Offset to its maximum of -4.5 improves, but does not resolve the clipping.

That's what I know. Patience and assistance greatly desired! I should be able to follow instruction, but I'm somewhat new to this world and don't have formal training, so I hope I'm worth the effort.

[vindoline]

Hi Backfish, that's a shame about your 3325. My 3325B threw the same fail 023 as yours. I actually documented my repairs pretty well here: https://www.eevblog.com/forum/repair/hp-3325b-repair/
I would highly recommend getting the scanned service manual from Artek manuals. Very reasonably priced and the scans are excellent. It has a very detailed troubleshooter Nguyen guide.  Is there any output at the BNC? Do you hear the attenuators relays clicking on startup and changing ranges? I was able to identify all the problems in my unit by careful examination of the circuit boards. I would start by checking the attenuators board. Also, please provide some pictures!

Good luck.

[Bakafish]

Yes, I read your post. The service manual you recommended has already been ordered as the factory scan is pretty awful.

There is output, the frequency is accurate, but the waveform is DC offset by ~+5v and the top of all waves are soft clipped.

The relays all clickity-click and I can change the aptitude and pull the DC offset down -4.5v which helps but doesn't resolve the clipping. Disabling the auto-calibration with the internal jumper seemed to correct the large positive offset, but the waveform was still clipped on top. All wave-forms are similarly affected.

All the boards are really really clean, there is no indication of any thermal issues, prior service or failed caps. The service manual seemed to point the finger at 3 different IC's all of which are out of production. One of the chips is listed at a salvage shop in Akihabara (I'm in Tokyo) and I ordered a few as a cheap trail by error component. The same part commonly failed on the 3325A, so I'm hopeful. The isolation steps outlined in the service manual will take some reading, they are a bit inscrutable, but I'm hoping a clean copy will help me better identify the test points and such.

If it turns out to be one of the other IC's I am going to have to make them out of discrete components or do more digging. Really hoping my luck changes.

I'll upload photo's tomorrow, I was a bit depressed about this today. It is really unusual to be taken for a ride in Japan, and I let my guard down.

The machine is a beauty otherwise. Great button feel and functional design. If I can get her fully working again I will feel a real accomplishment, my attitude is starting to recover and I'm ready to enter the breech. 

[wn1fju]

For what it's worth, my 3325B would periodically throw those errors.  Simply opening it up and manipulating the three ribbon cables (the ones
approximately in the middle) seemed to immediately cure the problem.  I think I isolated the culprit to one of the three and eventually touched up
the solder joints on the connectors on the board in desperation as I really couldn't see anything wrong with the cable itself.  I'm pretty sure they
changed the design of these cables when they went from the 3325A to the 3325B (at least on my 3325A, they are different than on my 3325B),
and they are a weak point in my opinion.

[Bakafish]

Thanks for your input!

Did you happen to experience similar DC offset issues or waveform clipping as I described? I'll check all the cables in the morning, it can't hurt. I've seen reports of the error being intermittent for some people. I'm looking for more correlation with the symptoms I'm seeing though.

[Bakafish]

Okay, I couldn't resist some late night cable re-seating, and sure as shite! I got a new fail code on power up, but it went away as soon as I switched the waveform, and they are all now working nice as can be!

I will do a deeper dive about possible contact oxidation and clean everything up, but I want to say thanks for the wonderful lead!

I feel cautiously optimistic and am looking forward to learning about my new device. I love this community!

[dacman]

I would consider using dielectric grease on those pins.

[vindoline]

Great news Bakafish! I love it when it's the simple things.

[Bakafish]

A more thorough cleaning and careful reseating of the 3 ribbon cables resulted in an error free power up and clean waveforms of all creeds and nations! 

I'm now just cleaning up old sticker residue and the patina of 29 years of loving operation and will order a modern fan to try and turn her melodious whine down a notch.

I've added the promised shots of the two boards (JPEG2000 encoding), someone Googling with more serious issues may get some value from seeing a happy board.

I would consider using dielectric grease on those pins.

These are all high end gold contacts, dielectric grease is great for keeping environmental contaminants away from a connection, but isn't something I would put directly on my pins as it is an insulator and strikes me as something I don't want involved in proper mating. I used high purity alcohol and reversed the direction of the short ribbon in order to refresh the contact surfaces. There is some subtle metallurgical thing happening here to make gold contacts degrade, but 29 years is a long time and it is voodoo beyond my ken.

[Bakafish]

Holy High Voltage Batman! 

It's a 115v 80mm case fan! The first time I've seen that before

We are 100v 50Hz here in Tokyo, so it is already spinning slow I suspect, but the thing is it must be running off it's own transformer winding or power supply as the whole unit can be switched from a range of 100-240v

I think I will leave it alone for now.

[TerraHertz]

Glad you got it going. FYI, the very first thing you should do with any old equipment experiencing errors, is 'wiggle all the connectors' and see if anything changes. Bad connectors, especially ribbon cables and most definitely IDC connector ribbon cables, are probably the number one cause of old equipment faults. Maybe the machine was working fine when the seller checked it, so didn't mean to cheat you.

I have a question about your photos. It turns out they are .JP2 format, not .JPG. See http://fileformats.archiveteam.org/wiki/JPEG_2000  I'd never even heard of .JP2 before.

Is that the output of your camera? Or an image editing utility? What type(s)?

With the fan, it's the usual practice to run that from a transformer tap. Other than using a LV DC fan, how else are they going to make the machine work with 110 or 240V mains?

[Bakafish]

Yes, jp2 is the ideal extension, but since most modern browsers and image viewers ignore the extension and this blog software doesn't accept that extension I chose to upload it this way. As for why I used jpeg2000, this blog also has some understandable image size limitations and so after cropping the images I wanted to preserve as much resolution as possible. The compression algorithms in jp2k are better and will give some desperate soul trying to decode the resister value or component orientation a fighting chance.

I think there is excessive glare, I need to work on a more optimal lighting setup. The camera is an iPad Pro 9.7” and I used Acorn on a Mac, to crop, convert and compress the images. They show up inline on my iPad and Mac just fine, and can be downloaded to be viewed locally by unsupported browsers if needed. There are even better, more modern file formats that offer better compression, but jpeg2000 should be far better supported at this time.

As for the fan, I’ve just never seen a fan of that style in anything other than low voltage DC. I was just taken aback. It’s not that noisy, but that’s a low hanging fruit mod that can provide instant benefits.

As for the connectors, the unit was so clean, and the connectors were all gold plated, it just never occurred to me that they could be the issue. But recognizing a good lessen when you see one is the heart of learning 

[merox]

It's a 115v 80mm case fan! The first time I've seen that before

That's not completely uncommon but unusual, i know at least another intrument using a 110/115V fan: The Keithley 237 SMU. It's using a seperate winding of the transformer to power it, no matter if it's a 220V device (yes, it's old, dating back to the 80ies) or a 110 one.

[texaspyro]

I've fixed several 3325's over the years and the most common failures were some shorted tantalum caps and bad connections with those coax cable that terminate in what looks a lot like an RCA jack.

[jcline01]

I know this topic is two years old, but for the record, the ribbon cables on the 3325A and 3325B are problematic. In almost every case, the gold plated pins and PCB are fine - no deoxit required!!! The problem is that the sockets within the ribbon are mechanically crimped to the ribbon wire. Each time the cable is removed (usually incorrectly), the wire to socket connection becomes slightly loser, until the connection is intermittent. This can cause errors of several types. Most commonly, the Ground, or +/- 15V rails have high resistance. this affects the output amplitude, as the  amplitude is highly dependent upon the +/- 15V being constant and calibrated. In DC Output only mode, slightly wriggling the ribbon cables from A14, A3 or A26 boards can cause the DC Output level to jump by several 10's of millivolts  - no way that can stay in CAL after pressing the CAL button. This is a tale-tell sign that your ribbon cable is bad. I have resorted to making my own cables to the units that I refurbish. There are other issues that may cause Fail 023, such as the A3 board adjustments / defective parts, so do your due diligence to make sure the correct fix is applied.

Joe, KN5U

[wn1fju]

I heartily concur with the previous post.  Although I never had a problem with my HP 3325A, I sure did have intermittent fail messages with my 3325B.  At one point, I put a few rubber bands around the problematic ribbon cable to try and force the cable to connect properly.  It never lasted very long.  I finally got so fed up that I took a bunch of individual jumper wires with sockets on each end and replaced the ribbon cable with a set of them.  Problem solved.

[SilverSolder]

Are the parts used to make the ribbon cables fairly standard, or did HP venture off the reservation for these?

[HalFoster]

+1 on the answer from jcline01 - On any new 3325 I just go ahead and replace the ribbon cables.  Cheap and they generally *will* fail sooner or later - usually from vibrations during shipping to the person you just sold it to...  What happened to you, I guess.  Anyway, the other connectors can, good as mentioned with any equipment, benefit from a disconnect/connect cycle.  These are really solid basic pieces of equipment and are great to have.

Hal

[JJ]

I had the same problem crop up today - Fail 023 and Fail 041 when the triangular wave is selected. I just let the unit warm up for 15 minutes and the problem disappeared. If you don't want to mess with your unit right away, try letting it warm up. The expansion of temp increase probably affects the ribbon cable connections.

[texaspyro]

A big source of '3325 problems is those "RCA jack" coax cable connectors.

[ARF]

From time to time, when I turn it on, I can get "FAIL 026" on my 3325B.

[SilverSolder]

Mine have been stable ever since I cleaned all connectors with DeOxit, followed by an application of Stabilant 22A.  So many magic potions needed to keep the magic smoke contained in these!

[lobstah]

Another FAIL 026 here on both start-up and self-cal (sine wave only).   Wiggling the ribbon cables didn't seem to help.  I ordered some F-F breadboard jumper cables and replaced the 3 21-pin ribbon cables with the jumper cables.   Turned it on.  NO ERROR!

Such a welcome surprise and easy fix.  I was dreading having to replace the multiplier, which was going to be the next step in the process.

[HalFoster]

Another FAIL 026 here on both start-up and self-cal (sine wave only).   Wiggling the ribbon cables didn't seem to help.  I ordered some F-F breadboard jumper cables and replaced the 3 21-pin ribbon cables with the jumper cables.   Turned it on.  NO ERROR!

That is what I have done - as you said, an easy, inexpensive and lasting fix.

Hal