HP 3314A Function Generator Repair

I found a great deal on a Hewlett Packard 3314A 20 MHz func­tion gen­er­a­tor that was non-func­tion­al, but over­all looked to be in good shape. It arrived in great con­di­tion and was well packed for a nice change. The sell­er clear­ly stat­ed the issues with this unit such as errors when start­ed up cold, which wors­ened over time until it final­ly lost the dis­play and LED’s.

The HP 3314A is a 0.001 Hz to 19.99 MHz func­tion gen­er­a­tor with Sine, Triangle, and Squarewave func­tions along with pro­gram­ma­ble Arbitrary wave­forms. It also is capa­ble of lin­ear and log­a­rith­mic sweeps, AM & FM mod­u­la­tion, inter­nal and exter­nal trig­ger­ing.
This unit is a lat­er mod­el with a 2836A ser­i­al num­ber pre­fix and some com­po­nents dat­ed ear­ly 1991. Making it 33 years old as of the writ­ing of this blog. Hewlett Packard made quite a few changes in this revi­sion, some of them good and some not so good.

There are sev­er­al good YouTube videos out there that show repairs of some com­mon and uncom­mon issues, espe­cial­ly with the ear­ly ver­sions of this func­tion gen­er­a­tor. While wait­ing for my unit to arrive I spent a good amount of time review­ing the videos and down­loaded man­u­als.
As usu­al my first steps were a thor­ough inspec­tion and a gen­er­al clean­ing mak­ing it eas­i­er to work on.

The only vis­i­ble issue that I saw was a dis­col­ored −24 volt cir­cuit fil­ter capac­i­tor which prompt­ed me to check all the elec­trolyt­ic capac­i­tors which are main­ly on the A3 board. The dis­col­ored capac­i­tor checked out OK, but the 5 volt out­put capac­i­tor on the switch­ing reg­u­la­tor test­ed very bad.
After replac­ing the 5 volt capac­i­tor and dis­con­nect­ing the cables to the oth­er boards and front pan­el, I was almost ready to pow­er it up. But first a quick check of the AC volt­age selec­tor switch and fit­ment of a prop­er­ly rat­ed fuse was prob­a­bly pru­dent.
With sev­er­al mul­ti­me­ters attached to the three DC volt­age out­puts I flipped the pow­er switch and all volt­ages were well with­in fac­to­ry spec­i­fi­ca­tions.
Next I con­nect­ed the front pan­el cable and then re-applied pow­er and was pleas­ant­ly sur­prised with the boot-up count­down dis­play, and then a myr­i­ad of error codes which I expect­ed due to the oth­er two main boards being dis­con­nect­ed.
The next step was to start con­nect­ing the A2 and A1 boards while con­tin­u­ing to mon­i­tor the pow­er sup­plies on startup.

After con­nect­ing the A2 board, I no longer had prop­er volt­ages on any of the pow­er sup­plies and of course no front pan­el dis­play. A quick check of the A2 board­’s pow­er rails showed a near zero Ohm short on the +15 and −15 rails and good read­ings on the 5 volt rail.
Of course the first sus­pects were the three mold­ed tan­ta­lum capac­i­tors. After un-sol­der­ing one leg on each of the capac­i­tors, two of the three were short­ed as I sus­pect­ed.
I just hap­pened to have some her­met­i­cal­ly sealed ver­sions of the tan­ta­lum capac­i­tors in my spares sup­ply, so I replaced all three.

It seems that in this lat­er revi­sion of the 3314A, Hewlett Packard under­took some cost sav­ing mea­sures specif­i­cal­ly replac­ing the expen­sive her­met­i­cal­ly sealed Tantalum capac­i­tors with a low­er cost mold­ed Tantalum ver­sion.
So far in all my decades of work­ing on elec­tron­ics, I have nev­er seen a her­met­i­cal­ly sealed ver­sion of a Tantalum capac­i­tor fail. But I have replaced dozens upon dozens of the mold­ed Tantalum vari­ants. Then again her­met­i­cal­ly sealed Tantalum capac­i­tors are less com­mon, so I should see less fail­ures.
You can prob­a­bly see where this is going. I must replace all the mold­ed Tantalum capac­i­tors in this unit if I want it to remain oper­at­ing reli­ably.
Did I men­tion that her­met­i­cal­ly sealed Tantalum’s are real­ly expensive!

After replac­ing the Tantalum capac­i­tors on the board, I recon­nect­ed the cable to the A2 board and once again pow­ered up the unit. The boot-up count­down was now work­ing again, and all volt­ages were now in their prop­er ranges.
After pow­er­ing down and con­nect­ing the A1 board, I was pleas­ant­ly sur­prised to see the boot-up count­down and no errors dis­played on the front pan­el.
A quick con­nec­tion to my oscil­lo­scope showed prop­er wave­forms and ampli­tudes for all func­tions and ranges of this unit.

After a parts order had arrived, I replaced all the elec­trolyt­ic and mold­ed Tantalum capac­i­tors on the A3 board. I had also ordered and replaced the thru-board con­nec­tors for the pow­er sup­ply frame mount­ed tran­sis­tors, as these have been not­ed as a point of fail­ure by many peo­ple repair­ing these units. The back­up bat­tery test­ed good at 3.1 volts and showed no signs of leak­age so it was left in place.

The A1 board need­ed 7 Tantalum capac­i­tors replaced and I also took the time to clean-up the twist­ed pair wire jumper on the back of the board. The adhe­sive had dete­ri­o­rat­ed over time and left a crusty residue. I used some elec­tron­ics grade sil­i­cone adhe­sive to reat­tach the heat shrink tub­ing which seems to encap­su­late some fer­rites back to the board.

After replac­ing all the Tantalum and elec­trolyt­ic capac­i­tors on the main boards, It was time to tack­le the front pan­el dis­play board.
I had already used up all the her­met­i­cal­ly sealed Tantalum capac­i­tors that I had ordered and all of my spares.

A deci­sion had to be made. I could order anoth­er $70 plus dol­lars of capac­i­tors or use some new mold­ed tan­ta­lums that I had in spares for the A11 dis­play board.
My lim­it­ed bud­get won out and all 7 capac­i­tors on the 5 volt rail were replaced with new 39 uF at 10 volt mold­ed Tantalum com­po­nents. Time will tell if this was a good deci­sion.
This dis­play board revi­sion is an inter­est­ing design with all 7 of the capac­i­tors con­nect­ed in par­al­lel. My guess is that due to height restric­tions from the prox­im­i­ty of the A12 board above, it was just eas­i­er to use mul­ti­ple thin­ner diam­e­ter capac­i­tors to achieve the need­ed 230 uF capac­i­tance with the ben­e­fit of a low­er total ESR val­ue.
Earlier ver­sions of the (A4) dis­play / keyswitch board used a sin­gle board and one large 220 uF at 10V Tantalum capac­i­tor for the 5 volt rail.

There is also a very large pink ZEBRA elas­tomer­ic strip con­nec­tor J3 that con­nects this board to the A12 LED and key switch board, which was­n’t nec­es­sary for the sin­gle board (A4).

The last board to work on was the A8 out­put ampli­fi­er board which need­ed to have it’s two elec­trolyt­ic capac­i­tors replaced, but also had been worked on some­time in the past.

The U4 relay dri­ver IC had been replaced and sock­et­ed, but it looks like who­ev­er per­formed the repair did not have a 16 pin DIP sock­et so they cut down a cheep larg­er sock­et to fit the board with some work­ing but mar­gin­al sol­der­ing of the sock­et to the board.
The sock­et was removed and the board cleaned up before replac­ing the sock­et with a qual­i­ty Mill-Max Beryllium Copper — Gold plat­ed 16 pin sock­et. The next best thing to direct­ly sol­der­ing the IC to the board in my opinion.

After every­thing was reassem­bled, I let the unit warm up for 30 min­utes then re-adjust­ed the + & − 15 volt sup­plies with­in a few mil­li­volts of 15 volts. The 5 volt rail which uses the +15 sup­ply as a ref­er­ence was well with­in spec­i­fi­ca­tions. ROM, RAM, and front pan­el diag­nos­tics were run with no errors, and all but­tons worked per­fect­ly.
Most of the per­for­mance tests were com­plet­ed with the excep­tion of tests requir­ing a spec­trum ana­lyz­er, due to lack of said ana­lyz­er.
All in all this is a great piece of test equip­ment, and I was again impressed with the qual­i­ty of engi­neer­ing and craft­man­ship that went into design­ing and fab­ri­cat­ing this unit with a few small excep­tions (mold­ed Tantalums). Although in the ear­ly 1980’s the price of Tantalum increased dra­mat­i­cal­ly which may have prompt­ed the cost savings.