HP 105B frequency standard (00105-6013) oscillator teardown

[tkamiya]

These days, an only output frequency that will be useful to you will be 10MHz.  For that, you'll have to tap off of the output of 11801 and bring it outside through buffer amplifier of some kind.

As to other frequencies, first thing I do is to unplug cards one at a time and closely inspect, then plug it back in.  That it self fixes a lot of problems....  Even though they are all gold plated, contacts can be iffy after a decade or two. 

Mine had a failure in voltage regulator transistor and filter capacitors. So it makes sense to check the power supply voltages as well. 

BE CAREFUL!  HP105B has mains voltage on front panel near battery charging switches!  I learned it the painful way.

[uncle_bob]

Hello forum members,

Here I am with HP 105 unit. I picked up as I love to collect such engineering marvels. it was last calibrated in 1998 by hp, the calibration seal was still intact. I powered up, as people reported, the meter goes above on OVEN, 18.7V and 5MHz switch position, the meter shows ~30 on both 1MHz and 100KHz.

Only 5 MHz output is working. the frequency is 5MHz (I do not have precision device to measure it to full accuracy of the 105 device) but 1MHz and 100KHz output is not there.

I opened up the unit, it has ovenized oscillator 10811A (which is newer than the pictures above posted, i think). I am wondering what would be first step to look into to find out the issue on 1MHz and 100KHz output. Any suggestion at this point is helpful.

Regards,
Xtentec

Hi

The first thing you will need is a manual for the later version of the 105. Most of what's out there is for the earlier version. With the manual, working out what's what is do-able. Without it, you will spend a lot of time tracing out circuits and guessing what they do.

Bob

[Xtremexp]

Thank you for valuable feedback and alert on mains voltage.

For sure I will check the power supplies first. I will keep updating as I find about more about this unit.

Thanks!!

[5065AGuru]

Xtentec,

Hi, You have a nice unit.

The 1Mhz and 100Khz outputs are derived from self starting regenerative dividers.

They are in series so if the 1Mhz divider is inoperative the 100Khz will not get an input.

Best to look in the manual and troubleshoot the 1Hhz divider.

Cheers,

Corby

[Xtremexp]

Xtentec,

Hi, You have a nice unit.

The 1Mhz and 100Khz outputs are derived from self starting regenerative dividers.

They are in series so if the 1Mhz divider is inoperative the 100Khz will not get an input.

Best to look in the manual and troubleshoot the 1Hhz divider.

Cheers,

Corby

Thank you for the clue. 1MHz divider is also not working. I probed the power supplies, the 24V rail is 29V and same for 18.7V rail. I suspect that 29V on 18.7V rail brought down all modules? It is scary. I found possible suspect (Q8 and zenor diode) on the A5 board. lets see how the fix goes. Technically I am still wondering even if the 18.7V rail is 29V, the 1MHz module should work as I see the 1MHz module has it's own internal voltage generation.

For guys who is interested on the unit. I attached picures of my unit. It's really nice unit. I love it.

[5065AGuru]

Take a look at the correct manual.

The version you want is at:

http://ftb.ko4bb.com/manuals/132.1.207.129/HP_105B_Late_Model_Quartz_Oscillator_Manual.pdf

The power supply should be +18.7V.

Cheers,

Corby

[5065AGuru]

Bob,

Per your guess:

"Just to keep things straight ... The crystal is most likely a BT cut. That was HP's "secret weapon" back before the SC."

The crystal cut in the older version oscillator is AT!

The BT cut was used in the HP 10544A

Cheers,

Corby

[tkamiya]

I just received a second HP-105B.  Transformer inside is dated 5/31/89, so it's probably close to the last version.  If you need me to measure or take picture of something, let me know.  I'll be glad to help.

This thread is somewhat confusing to me as the ORIGINAL poster had a unit with older oven at 5MHz and latest discussion is for HP-10811.  I think they key will be if the oscillator is working or not.  Everything else, you can repair.... 

http://hparchive.com/Manuals/HP-10811AB-Manual.pdf

As you can see on page 8, it takes two input.  24V for pin 14, 12V for pin 3.  With just that, you should get a signal at 10MHz on pin 1.  Please see respective ground connection on that chart.  You CAN buy 10811 variant fairly cheaply but I *think* ones used on HP105B was hand selected version.  Bob can probably attest to it. 

**update**
I just re-read the second OP's post.  He has 5MHz out, so at least OCXO is working and divide by two circuit is working.  Phew.
\**update**

Likely, the rough control on the front panel won't work.  On mine, fiberglass shaft burned out and metal blade fell out.  I had to remake mine.  I have a dimension for that somewhere on this forum. 

I'll be working on fixing few thing on my new one.  Good luck with yours.

[Uky]

BTW:

The last version of the HP105B (that have a yellow LED for AC interruption) featured a redesigned PSU board (A5) using "modern" IC's such as 3-terminal regulators. I uploaded an addendum to the manual describing the Oscillator (10811-version) as well as the A5-board. It should have ended up in KO4BB's "HP105"-directory but the addendum ( a .ZIP-file) with two .PDF's inside instead is in the root (?) folder on his server. This .ZIP-file seems to be the only one that also has the schematic drawing for the A5 board in it.

[Xtremexp]

An update on my unit. I fixed the 18.7V rail by replacing the transistor Q8 on the power supply board. Now I have outputs in 5MHz, 1MHz and 100KHz.

My next goal is to add 10MHz output on the system as the oscillator has 10MHz output. Any idea of buffering the output to get 10MHz signal to drive 50Ohm load is welcome.

Also if anyone needs help on probing signals then please feel free to message me.

Thanks all.

[tkamiya]

I just did that to my second HP105B.

I did it by tapping off 10MHz directly from a socket where 10811 (OCXO itself) socket.  It then goes into a buffer amplifier with relatively high input impedance so it won't affect the original signal level to other circuit match.  An output connector (SMA) was placed 1 1/4 down from center of the coarse tuning hole.

The amplifier is THIS circuit from John Miles:
http://www.ke5fx.com/norton.htm

Please read second from the last paragraph under "Source Material and..."  He has words on circuit loading and how to deal with connection to HP10811.

I have a PCB for it if you'd like.

[Xtremexp]

I just did that to my second HP105B.

I did it by tapping off 10MHz directly from a socket where 10811 (OCXO itself) socket.  It then goes into a buffer amplifier with relatively high input impedance so it won't affect the original signal level to other circuit match.  An output connector (SMA) was placed 1 1/4 down from center of the coarse tuning hole.

The amplifier is THIS circuit from John Miles:
http://www.ke5fx.com/norton.htm

I have a PCB for it if you'd like.

Sure I am interested in the PCB and curious to see how your unit looks like. I sent PM to you.

[tkamiya]

I don't have inside picture.  I don't want to open it as it is going through an aging cycle.  But I can give you an outside image.  The modification for output port is right below the coarse adjust hole.  I mounted a bulkhead SMA 1 3/16" away.  Center to center.  Looks pretty stock doesn't it?  Location of this is kind a tricky because there is a "lip" inside.

What I have done is, on a PC board  directly behind the OCXO, tap off 18.7V and feed it into 7812 with 0.1uF in the usual way.  Mount the 7812 onto an aluminum bracket (right next to the existing one).  Tap off 10MHz from pin 1 of the OCXO through 100pF capacitor.  Ground goes to the nearest ground.  That goes into a thin coax.  That coax connects to the input side of the buffer amp.  (J1)  J2, the output goes to a short coax harness to a connector on the front panel.

I removed the Nicad pack completely.  I have everything connected to a big UPS.  I don't need per-unit backup.  I mounted the buffer amp in that space using studs.

[Xtremexp]

I don't have inside picture.  I don't want to open it as it is going through an aging cycle.  But I can give you an outside image.  The modification is right below the coarse adjust hole.  Looks pretty stock doesn't it?  Location of this is kind a tricky because there is a "lip" inside.

What I have done is, on a PC board  directly behind the OCXO, tap off 18.7V and feed it into 7812 with 0.1uF in the usual way.  Mount the 7812 onto an aluminum bracket (right next to the existing one).  Tap off 10MHz from pin 1 of the OCXO through 100pF capacitor.  Ground goes to the nearest ground.  That goes into a thin coax.  That coax connects to the input side of the buffer amp.  (J1)  J2, the output goes to a short coax harness to a connector on the front panel.

I removed the Nicad pack completely.  I have everything connected to a big UPS.  I don't need per-unit backup.  I mounted the buffer amp in that space using studs.

No need to open it now. It looks amazing!! such a beauty

[uncle_bob]

Hi

Corby has done a few 105's, 5065's, and others with the Clifton Labs buffer. I have one example and it seems to work fine.

https://www.dxengineering.com/parts/dxe-z10000u-kit

Power it off the 18.7V rail, attach the input to the 10811's output. Bring the output to a BNC somewhere ( either put in a new one or re-purpose one that's there).

Bob

[tkamiya]

I have a few Clifton Lab buffer amp here, too.  They work well.  They are especially nice on low level signal.

An only limitation I thought undesirable for my need was its output level.  At around 5 to 6 dBm it starts to saturate.  If it could do just a couple db more, it would be a perfect ready made solution.

[Uky]

I once made (or rather: copied) a low noise Frequency Doubler, 5 -> 10 MHz design referenced by NIST. I think the original design was found on Bruce Griffiths home page.
See attached picture. I have schematic, layout (Cadence OrCAD/Allegro ".BRD") and gerbers should anyone be interested.

 

[Xtremexp]

Thanks tkamiya for the amplifier buffer. I modified same way that tkamiya  added BNC on the front panel. I added the PCB between the power supply and the 1 MHz divider space with existing PCB support. Now I have 10MHz source.

Thanks all for the support!!

[tkamiya]

Looks almost stock doesn't it? 

As you already know, I have two.  First one drifted quite a bit when I bought it and it took good 6 months to get to the ultimate stability.  Second one has been running for few months but it is still drifting.  (frequency goes lower and lower....)

I hope you don't have the old battery elements themselves in there.  Those are NiCd.  They can leak and can cause severe corrosion.

[trobbins]

I'm part way through repairing a 00105-6013 crystal oven assembly used in a HP5065A. 

It's first identifiable fault was leakage to ground on the 115Vac fast warm-up heater which is detected by IR measurement of mains input to earth, and also tripping a safety ELCB.  But that heater is not pedantically needed, so the next fault was magic smoke from the connector end, which led to a few other faults being identified.  The main problem appears to have been decades of long term high temp operation, perhaps including overheating, such that the internal foam insulation has hardened and charred in places, and insulation of wires in the ribbon-cable that passes between ends had degraded.  Repairing the poor interconnect wires, as well as two shorted diodes used in the voltage doubler that provides the heater temp feedback, and a charred resistor that may have seen a bad interconnect, showed that the circuitry was still working, but ad-hoc conduction from the normal heater to ground meant there was still a fault situation - and the internal can had to be excavated from the hardened foam coffin to explore further.  Extracting the can showed the heaters were resistance wire wound on the outer surface of the can, and that insulation tape had degraded and the charred foam was provided a leakage path to the can.  Removing charred leakage paths has allowed the 100 ohm normal heater to return to a high resistance to earth (still some degraded insulation but no overt conduction paths) even when powered - enough to warrant a return to service, so I'm just about to reconstruct the ribbon wiring and somehow centre the internal can back in the outer assembly and retest to see if all functions have recovered.  There is still a risk that the heater could form a low resistance path to ground again, in which case the heater could plausibly be stripped off and an array of fixed resistors affixed to the can as a psuedo heater.