Products > Test Equipment
HP 5316A Teardown and Frequency Standard
<< < (15/22) > >>
amc184:
Hi Mike,

That's interesting, it seems your unit is different to the others.  Looking at Luthor's unit, it looks like the large DIP packaged crystal oscillator might be socketed using individual pin receptacles.  This would also explain why it is zip tied down.  It should be easy to remove (no desoldering needed), which is good because I think you'll need to.

I've attached this section of the circuit below.  J2 at the left is the connector that you'll plug the OCXO board into.  Pin 2 is the output of that board.  Y1 is the crystal oscillator installed in your mainboard, and after the buffer this joins into the same net used by the OCXO output (which I've highlighted yellow).  You wouldn't want both clock sources driving the same net, so Y1 must go.

Besides that, I think you need to use the hole right underneath Y1 to mount the OCXO board, right?
pinpassion:
Hi Luthor2K,

Thanks for removing your board and taking a picture. Very interesting! You have the circuitry for the on-board crystal (Y1) but not the crystal! It's place on the mainboard can be seen, so the pads are there, and the same holes that are on my board where the zip tie is holding it down. You even have the oscillator adjust trimpot. What I think may explain this is when they introduced the "B" counter model, HP wanted to exhaust the supply of the external boards that were used on the "A" model, and when that was done, they soldered the crystal to the main board to replace the daughter board entirely. This must have been the plan when the PC board was manufactured and that explains the pads and circuitry.

I downloaded a service manual scan online for the 5316B, and had a look at the schematic. The on-board crystal is shown. Examining the schematic, it looks like I will have to desolder my crystal to take advantage of the OCXO board before I install it in J2 on the mainboard. If I didn't, both 10MHZ signals would be injected. There are ways around doing it, like lifting pin 2 from the OXCO board and running that into the external input and keeping the rear panel switch in the external position, but I think I will just bite the bullet and remove the crystal. If I did wire the OCXO into the external input, the one interesting benefit is i would be able to switch back and forth to the on-board crystal and the OCXO board easily to compare both references..Nah...I know the OCXO board is going to be a lot better.

Thanks for taking the picture. It helped me to understand the whole external oscillator issue.

Mike
pinpassion:
Hi AMC184,

Wow..Our posts crossed!! You came to the same conclusions as me! I was hoping the crystal was socketed on the mainboard, but as you can see in the board bottom photo, it is soldered...OR, is that the pins to the socket that are soldered...? I don't know!! I am hoping it is a socket!! We will see after I remove the zip tie. Unfortunately, there is an explanation as to why HP may have used the zip tie outside of securing it to a socket. From what I understand, crystals are mechanically vulnerable, and can even be influenced by their relative position. The zip tie could have been used to stabilize it on the cheap. I hope it is your assumption that prevails! Then I could just pop it out.

Any yes, the crystal does get in the way of one of the daughter board mounting holes. Removing it would solve the mounting problem as well!

Thanks again for your help!!! I have a board and some components on the way from Luthor2K. The MV85 I have coming from Ebay. Once I have the board ready, I will cut the zip tie, and the soldered or socketed question will be answered. I will let you know!

Mike
amc184:
Mike has asked me for a better bill of materials for the PCB, and I thought I'd post it here so anyone building this PCB in the future can see it.

C101 and C105: 330µF 35V Ø10mm THT low impedance electrolytic capacitor
Nippon Chemicon EKZE350ELL331MJ16S
Element14, Mouser, Digikey

C102, C104 and C106: 100nF 100V THT 5mm pitch metalised PET capacitor
Wima MKS2D031001A00KSSD or MKS2D031001A00MSSD
Element14, Mouser, Digikey

C103: 10µF 63V Ø5mm THT general purpose electrolytic capacitor
Nippon Chemicon EKMG630ELL100ME11D
Element14, Mouser, Digikey

C107: 1nF 400V THT 5mm pitch metalised PET capacitor
Wima FKS2G011001A00KSSD or FKS2D011001A00MSSD
Element14, Mouser, Digikey

J101: 2.54mm pitch THT pin header, 10.8mm total length, 6 way
Amphenol 68000-204HLF
Element14, Mouser, Digikey

R101 and R105: 100R DIN0204 size THT metal film resistor
Welwyn MFR3-100RFC or Vishay MBA02040C1000FCT00
Element14, Mouser, Digikey

R102: 680R DIN0204 size THT metal film resistor
Welwyn MFR3-680RFC or Vishay MBA02040C6800FCT00
Element14, Mouser, Digikey

R103: 560R DIN0204 size THT metal film resistor
Welwyn MFR3-560RFC or Vishay MBA02040C5600FCT00
Element14, Mouser, Digikey

R104: 20kR 25 turn THT potentiometer with staggered leads and side adjustment screw
Bourns 3296Z-1-203LF
Element14, Mouser, Digikey

U101: LM1084, LM1085 or LM1086 type TO220 adjustable linear voltage regulator
ST Micro LD1086BV
Element14, Mouser, Digikey

X101: Morion MV85A C20F-5V-SIN
I bought mine on eBay, at the time these were a compact, cheap and power efficient option with the right output type and level.

Other things you'll need:

* Two UNC6-32 threaded spacers, male to female type with a 7/16" body length to mount the board.  You'll also need a couple of 6-32UNC machine screws.
* An M3x10mm machine screw and two M3 nuts.
* A TO-220 heatsink, such as a Wakefield 274-2AB Element14, Mouser, Digikey.
When mounting the heatsink the M3 machine screw should go through from the underside of the board, then have the following stack; PCB, M3 nut, heatsink, TO-220 voltage regulator, M3 nut.  The first nut is important, it spaces the heatsink off the PCB.  Make sure the head of the screw isn't so large that it's bigger than the diameter of the area clear of groundplane on the PCB.  You don't need to use insulating hardware to isolate the voltage regulator from the heatsink unless it's touching something conductive, which it shouldn't be.
amc184:
In the bill of material I previously posted I used an adjustable voltage regulator, but it's possible to use a fixed one as well.

In the original design an adjustable LM1086 is used, which needs a resistive divider to set its output voltage.  R101 and the parallel combination of R102 and R103 form this divider - I used 560 and 680R in parallel to make a divider for 5V without having to use anything other than standard E12 series resistors.  C103 is optional, improving ripple rejection.  The ability to add C103 is one of the advantages of an adjustable regulator - most fixed regulators do not give access to the feedback node to add this capacitor (though some do).

When using a fixed voltage regulator R101, R102, R103 and C103 are not needed.  Pin 1 of U101 should also be connected to ground, the neatest way to do that is to fit a zero ohm jumper to the R102 pattern.  So the changes to the bill of materials would be:

C103: do not fit.

R101 and R103: do not fit.

R102: DIN0204 size jumper
Vishay MBA02040Z0000ZCT00
Element14, Mouser, Digikey
Or use a piece of wire, but the jumper will look neater.

U101: LM1084, LM1085 or LM1086 type TO220 fixed 5V linear voltage regulator
Texas Instruments LM1086CT-5.0
Element14, Mouser, Digikey


A couple of things to note about voltage regulator choice in this circuit:

* Both the fixed and adjustable versions of LM1084 / LM1085 / LM1086 regulators use the same pinout; 1: adjust or ground, 2: output, 3: input.  This is the same pinout as many adjustable resistors (such as the LM317), but is not the same pinout as most fixed regulators.  For example, the LM7805 uses 1: input, 2: ground, 3: output.
* Most linear voltage regulators are damaged if the voltage at the output becomes higher than the input voltage.  In this circuit this may happen if something was shorted on the input rail, taking it down to 0V while C105 is still charged.  An LM1084 / LM1085 / LM1086 regulator builds a protection diode in to avoid this situation, but many regulators do not.  When using an LM317 or similar these need to be provided externally.  This board does not include reverse bias diodes, so for this reason I would avoid using the LM317 or similar here.
Navigation
Message Index
Next page
Previous page
There was an error while thanking
Thanking...

Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod