HP / Agilent 53131 / 53132 / 53181 GPSDO timebase

[snickers]

Hi everyone! This is my custom-designed GPSDO timebase for HP 53131/132/181 counters. It fits perfectly in place of the original option. Based on ublox GPS module and high-quality OCXO.
Totally reversible modification. Uses existing rear-panel knockouts, so no drilling is required. Ideal if you want to keep your unit in factory condition.

Price is $140(+shipment) via PayPal for forum members (cheaper than my eBay listing).
If anyone is interested, feel free to PM me with your shipping details, and I'll send over the PayPal invoice.

It is available here:
https://www.ebay.com/itm/126384459263
https://www.ebay.com/itm/127611556956 (GPS antenna included)

[Simon]

Ands you never noticed that there is a buy/sell section.

[Zucca]

Interesting, can you post a picture of an assembled module in the 5313x? I mean where do you mount the antenna BNC female connector?

[the Chris]

It replaced the OXCO in my 53181A, it should be no different in the 5313xA. The module is delivered with a SMA to BNC jack that fits into one of the free inputs on the backside, but I messed to take a photo of it when finished. The module works great, btw.

[snickers]

@the Chris, thank you for the kind words and the photo! I'm really glad the module is serving you well in your 53181A.

@Zucca, exactly as Chris mentioned. The kit comes with an SMA-to-BNC pigtail. You just mount the BNC connector into one of the knockout holes on the rear panel — no drilling required. See the attached photo for a closer look. Let me know if you have any other questions!

[pizzigri]

Just for info, what is shipping in EU? I’m in Italy, but more or less shipping intraEU should be very similar.
ETA this morning’s fixing should translate the US$ price in just below 125 euro including spread and stuff.

[snickers]

Yes, absolutely. I ship to Italy regularly.
Since we are both in the EU, there are no customs fees or import taxes involved.
I’ve just double-checked the current rates via InPost (to Parcel Locker or Parcel Point).
Cost: ~€15.
Time: Typically 3–5 working days.
Please PM me your shipping details if you'd like to proceed, and I'll get the invoice ready for you.

[Zucca]

Nice,

probably I am not the only one asking if this is compatible with Lady Heather?

http://www.ke5fx.com/heather/readme.htm

Also how do I know the board is locked to a GPS (satellite)?

I am sorry if they are stupid questions, I consider myself a noob in the time nuts field, have mercy on me.

[snickers]

Hi Zucca,
There are no stupid questions here! These are actually very important details.

1. Visual Indication (Daily Use)
You don't strictly need external software to check the status. I designed the board with two onboard LEDs positioned to shine through the existing translucent window on the counter's front panel. You will see them without opening the case:
Yellow LED (flashing 1 sec): Indicates the 1 PPS signal. If it is blinking, it means GPS is Locked (satellites are acquired).
Green LED (solid): Indicates that the internal OCXO is fully synchronized and stable.

2. Lady Heather & u-center Support (Engineering Mode)
To answer your first question: Yes, it is fully compatible.
However, accessing the data stream is intended as an "Engineering Mode". Since the counter doesn't pass this data to its own rear ports, you would need to:
Open the cover.
Close the "NMEA" jumper on GPSDO board.
Connect a USB-to-TTL(3.3V) adapter to the pin header UART.
In this configuration, your PC recognizes it as a standard generic GPS module.
I have attached a few screenshots below showing the module connected in this mode, running successfully with both Lady Heather and u-blox u-center.

Best regards,
snickers

[Zucca]

 

Man, that whole band started acting up again at my place — after calibration, it turned out that the 2.4 GHz modulation was picking up some phantoms from clock harmonics, so I had to manually boost the offset on the synthesizer. And you know what? It's still floating like an Arduino on a Chinese stabilizer.

On top of that, their “smart” PLL loop got so out of whack that the phase looks like someone drew it with a mouse on spilled beer.

Finally, I disconnected everything, reset the bias, and connected the good old x10 probe — and suddenly, as if by magic, a clean peak appeared, as if the interference had gotten embarrassed and gone home on its own.

If this continues, I think I'll solder a second filter myself, because this factory RF section is more capricious than an ESP8266 on a cold start.


Translated with DeepL.com (free version)

[snickers]

Option feature: Smart Fan Controller & Thermal Stabilization
Note: This step is optional. You can still use the GPSDO as "plug & play" without touching the fan wiring if you prefer.
Many of you know that the stock HP 5313x fans can be quite noisy and only run at 2 fixed speeds. This not only creates annoying noise in the lab but also causes "thermal cycling" inside the case, which can affect the oscillator stability.
I’ve integrated a custom solution directly into the GPSDO board to fix this!
The Problem: The stock fan runs only 2 speeds(low/high) fast/loud and causes temperature fluctuations.
The Solution: An onboard PID controller that smoothly drives the fan based on the PSU power transistor temperature.

How it works:
1. You move the fan wires from the PSU board to GPSDO header.
2. You attach 10k thermistor to the PSU heatsink and connect to header NTC on GPSDO board.

The Result:
Silence: In standby or light load, the fan is barely audible. No more "jet engine" sound on powering on counter.
Precision: The PID loop keeps the internal temperature constant(65C - PSU power transistor temp), eliminating thermal drift and allowing the OCXO to perform at its best.
Note: This step is optional. You can still use the GPSDO as "plug & play" without touching the fan wiring if you prefer.

[Zucca]

Excellent!

Quick question.
Is the goal to keep the OCXO or the PSU temperature constant?

I would have moved the NTC close to the OCXO for better regulation.

[snickers]

Fan control & Sensor placement
While placing the sensor on the OCXO makes sense for a pure thermostat, here we have to prioritize the hardware safety.
The fan’s primary function is to cool the Power Supply Unit (the Delta board). The primary (hot) side power transistor is the most thermally stressed component in the entire device. Its temperature rises based on the total load across all three power rails (+5V, -12V, +12V).
If we regulated the fan speed based only on the OCXO temperature, we could run into a situation where the PSU is under heavy load but the fan spins too slowly, causing that main transistor to overheat and burn out.
Therefore, the NTC must be mounted specifically on that transistor's heatsink. By keeping that critical point at a constant temperature, we ensure the PSU survives, and as a side effect, the airflow stabilizes the entire internal chassis temperature, which is exactly what the OCXO needs.

[neverendingstudent]

While this may not be the optimal solution for stable OCXO temperatures, it is very likely the necessary solution to keep the unit alive.  That doesn't mean there isn't some better concept for 'why not have both' but I'm not thinking of it off the top of my head while keeping a single-fan design.  I have seen PLENTY of no-power 531xxA units on eBay with dead power supplies, and bought some of them myself.  The power supply is definitely at risk of overheat as stated.

EDIT: My personal solution has become feeding a GPSDO 10MHz reference into the unit and routing the power through a 1U rackmount power delivery switch.  I leave power disconnected from the unit when I'm not using it and have very fast lock to my reference signal when I turn it on.  Not the best solution for the care and feeding of the ovenized oscillator, but much better for my noise-sensitive ears and mind.

[Zucca]

Thanks for clarifying I was just not sure! Also less unnecessary air moving means less dust accumulation, which is definitely one of the common problems.

Here the PSU schematics, just in case:


   

[neverendingstudent]

Oh right, while I'm remembering it, it seems relevant to bring it up here.  I know at least one of those 5 watt resistors in the power supply is also a known failure point - it overheats and burns out, going open which then prevents the power supply from running.

[frantisek]

The fan runs at ONE speed! I wonder which HP 53131A or 132A runs at two speeds...  I wonder why it says in the offer: https://www.ebay.com/itm/127313866944?itmmeta=01KCE84CMSYJGZ6N3QXP9E51RB&hash=item1da47f70c0:g:DbYAAeSwk6JoocwN  is the price "slightly" different?!

[the Chris]

The fan runs at ONE speed! I wonder which HP 53131A or 132A runs at two speeds...  I wonder why it says in the offer: https://www.ebay.com/itm/127313866944?itmmeta=01KCE84CMSYJGZ6N3QXP9E51RB&hash=item1da47f70c0:g:DbYAAeSwk6JoocwN  is the price "slightly" different?!

The linked offer is for snickers' 3GHz prescaler. This thread is about snickers' GPSDO. Two different things although you can buy both from snickers (like I did) and they will work just fine together.

[frantisek]

It does, because it sells both. I also bought a divider in July, but only just installed it (because I had a warranty seal from the seller on the counter, which I kept), and it doesn't work. The seller doesn't want to deal with the problem because it's been more than six months, and I gave him a negative rating because he didn't communicate and I paid $25 for shipping. According to him, the shipping costs are determined by eBay, but pizzigri got a slightly different answer...  So that's about it.

[snickers]

@frantisek,
You stated clearly that you purchased the unit in July. You are now claiming it "doesn't work" 6 months later.
1. Timeline: No seller on eBay or in the industry provides a return window of 6 months for open-box electronics components. Standard practice involves inspecting items upon receipt, not half a year later. Storage conditions or installation errors during this long period are entirely out of my control.
2. Feedback: You admit leaving negative feedback immediately after purchase solely because of shipping costs (which are calculated by the carrier, not me), before you even tested the unit. This confirms the bias.
All my devices are individually tested before shipping. Addressing a claim after such a significant delay is impossible.
You have already used the eBay feedback system to express your dissatisfaction regarding the shipping costs.
Please stop cluttering this technical thread with disputes that belong in private messages or eBay resolution center (which has time limits for exactly this reason).
Let's keep this topic strictly for technical questions regarding the GPSDO/Timebase functionality.

Please respect the community rules and stop derailing this thread with unrelated individual grievances.

This is my final public reply to you on this matter.

[Zucca]

In few days this 10K NTC will be delivered.

Keep you posted.

PS: I did not fire up my timebase yet...

[frantisek]

Just to clarify! You refused to communicate on eBay, so I am writing here. In our conversation, you emphasized about five times that you check EVERY piece before shipping. That is a LIE!!! How could the divider you sent me, which had a +5V MIC2920A stabilizer in an SOT 223-3 case, pass this test? Everything is documented. After replacing it with SOT223, the divider started measuring. BUT! The divider oscillates on its own, and even without a connected signal, the display shows approximately 2.214 GHz. It measures with a signal. Is this a feature? You did not mention this anywhere, unlike some reputable Chinese sellers. I will not comment on other lies, such as that the shipping price is determined by eBay. But what I will do is verify the input sensitivity and share the results.
To err is human, and everyone makes mistakes. But few people can admit their mistakes, let alone apologize.

[Zucca]

Very happy with the thermal PID regulation that this board offers.

The NTCASCWE3103F is made for this.

Tapped a 4mm .7mm thread in Q1 heatsink



drilled by hand with a 4mm hole about .5mm deep into the thread to make sure the NTC surface will rest on the heatsink (in other words, the NTC screw thread stops at about .5mm... and do not cover the entire screw)

Very gently with a litte thermal paste I put everything together with a 4mm bolt on the other side



I used a FAN embpapst 412/2, the FAN PIN OUT of the connector is the following

- Red 12V
- Blue GND
- White Tacho


 
above my final result.

Worth to mention for future buyers:

- unit works 100%, tested against a locked HP Z3805A
- the M3 mount screws were not included
- included documentation could be better: FAN PID not described (pinout FAN???), red led error code unknown

My personal suggestions

- molex asymmetric connectors for the NTC and, especially, for the FAN one.
- LED mounted at 90 degrees for better visibility

- contact dimmog to put the 3 leds on his board.

[snickers]

Hi Zucca,
Thank you for the detailed feedback and the photos! Comparing it against a HP Z3805A is a serious test, and I am glad to hear the results are spot on.
Your method of installing the NTC sensor (tapping a thread in the heatsink) is top-tier engineering! That definitely provides the best thermal coupling possible. (Though for others reading this – usually, just clipping it or using thermal glue/tape is enough, no need to drill if you don't want to).
Regarding the documentation and the PID Fan Controller:
To be honest, the thermal management feature was added to the board as an "experimental bonus".
The product is sold primarily as a Timebase, but since I personally hate the fan noise, I decided to include the fan control circuitry and firmware for free as an extra feature for enthusiasts. Because it is an "extra" outside of the main spec, the documentation for it is indeed a bit lighter than for the core GPSDO functions. I appreciate you sharing the pinout you used — I will update the manual to make it clearer for future users!

The lack of documentation is also because currently 110% of my R&D time is focused on the new CMU-B11/B12 GPSDO replacement for Rohde & Schwarz CMU200 units (prototypes are already in testing!), so I haven't had the chance to polish the HP user guide yet.

Point taken regarding the mounting screws and the LED placement. I will look into adding polarized connectors in the next hardware revision to make connections fool-proof.
Thanks again for the review!

[neverendingstudent]

The lack of documentation is also because currently 110% of my R&D time is focused on the new CMU-B11/B12 GPSDO replacement for Rohde & Schwarz CMU200 units (prototypes are already in testing!), so I haven't had the chance to polish the HP user guide yet.

I've seen those all the time for cheap on eBay (the CMU200) and when I have done research on them the consensus I came up with was along the lines of 'yep, the hardware inside those can give you some pretty useful capabilities, but the firmware was pretty strictly designed / locked in for early-gen cell phone testing'.

There are people who have made efforts towards repurposing this device towards general use utility (and the *potential* there is very attractive), but as of the last time I did research on the unit the ability to use it for broad general utility (to basically unshackle it from the narrow-purpose cellular testing capacity) was still at high-complexity / high-effort / low convenience.

One of the people who have been working on this (that I know of):

Has more progress been made on this since... roughly a year / year and a half ago?  I think that's when I last looked into it.  Has anyone developed (working / reliable) drop-in modded firmware?