Workshop GPS 10MHz Timebase Wrap Up....

 Gee, now that I've got the HP Z3801A "SmartClock" up and running nicely, why don't I go ahead and use it as the 10MHz Master Oscillator for all my test equipment?

So I spent this afternoon doing that. Had to move some things and string a cable from the receiver over to the antenna sitting in the window, and then plug it all together.

Here's the antenna, doing a very nice rendition of "Poor, Poor, Pitiful Me" as it's at least 20" below ground, in an East facing window, with a metal window well. I'm picking up satellites from the other side of the house when their elevation is high enough. I'm guessing it's reflections it's picking up. "It Works", but will work as-intended once I get the antenna outside and up in the air.

This is the software I'm using. It's "Lady Heather's Disciplined Oscillator Control Program", emphasis mine. The program icon is a Dominatrix with a whip, as seen in the extreme lower left corner. It does everything I wanted to do, and more, and it's a free download compared to the other paid (and crappy) software I'd been using.

 So here it all sits, an "obsolete", surplus GPS Clock, with a cast-off, rescued old laptop I was given, providing an extremely accurate, stable, 10MHz reference signal to all my test gear.

"Old" Technology can be very useful when properly applied.....

And speaking of Old Technology, the photocells for the Heathkit THD Analyzer have finally arrived, so that turkey can go back on the bench.....

HP Z3801A GPS "SmartClock" Repair

 So there I was.....

With apologies to juvat over at The Chant....

As part of the electronics shop "upgrade", I dug out my old HP "SmartClock", which is a GPS-based time and frequency standard. I've had this little box for around 20 years, and it always worked flawlessly. I originally bought it to use as high-accuracy portable clock to synchronize my laptop clock for portable satellite operation, as well as give me an accurate LAT/LON position to enter into my satellite tracking program. Kinda killed two birds with one stone. It also has a VERY good 10MHz timebase in it, which I had plans to use, but never did. These were originally made for use at cellphone towers, where accurate time was used for billing and hand-off timing between towers.

The neat thing about the 10MHz timebase is that it's not only an HP "Double Oven" oscillator, but it's also "GPS Disciplined", meaning it gets locked to the extraordinarily accurate clocks in the GPS satellites, and provides a 10MHz signal that accurate to something like one part in 10e12. Crystal oscillators are about "as good as it gets" for short-term stability, but can, and usually do, drift over time. Part of the GPS signal is a one-pulse-per-second (aka "1pps") timing hack that's ONE SECOND, by God. Well, about as good as He lets us make for now.  How the GPSDO works is explained in the link above, so I won't go into it here.

Anywhoo.....When I plugged it in some years back, none of the lights came on, so I shelved it until I could a round tuit. Opened it up the other night, and started poking around to figure out why it wasn't powering up. The supply was good, but no blinky lights. Then I noticed this:

Well there's your problem....One of the Molex contacts came loose in the housing, and slid out of position when the housing was plugged into the board.

Easy-peasy fix. Pry the tiny locking tab out a bit, snap the contact back in the housing, and reconnect.

Fixed and ready to plug back in.

The next step was finding the old software I had, and then getting it to run on Windows 7, which is still a "Work In Progress". In the meantime, I was able to connect to the clock using a terminal program, and an RS232-to-USB adapter. Two issues with older GPS units is that if the clock has been turned off for ~90 days, and/or moved a few hundred miles, it won't have a proper list of satellites to look for when it's turned back on.

This one's been off for at least 4 years, and has moved 1100 miles since it was last used. In that case, you have to tell the device to run a "Sky Survey", and collect a current almanac from a satellite ASAP. So, connecting to the clock, I sent ":gpsystem:survey once", so it would grab the first satellite in view, and get an almanac. With a GPS antenna sitting in the basement window, it was able to do this, barely, and at least get itself locked to two satellites.

So it's just sitting there now collecting data, and begging me for a better antenna, which will have to wait for Spring when I can drill another hole through the foundation (poured concrete) and snake a new cable into the basement.

And I'm still waiting for the photocells I need to finish the THD Analyzer. USPS says they cleared Denver this morning, so hopefully I'll get them tomorrow.

Have a happy and blessed Sunday!

Minor Upgrade Project Completed!

 Finally got off my duff and put all of my test equipment on the same 10MHz timebase. Years ago (at least 10) one of my friends gave me a little Video Distribution Amplifier, and I toyed with this idea, but never had enough test equipment to make it worthwhile. Then I bought an HP Z3801A GPS Receiver, which was used at cellphone sites to keep things synchronized.

And again, the little VDA sat unused. Well, I got it out the other night to see if it even worked, and it did, but two of the output connectors had low signals on them. I cleaned the connectors and nothing changed, so I'm guessing that's why my friend gave it to me. Fortunately, since I have three instruments to lock together, I was able to use the "Loop Through" port to pass the reference signal through to one instrument, and had two good ports for my other two instruments.

Not much to see....a black box on a black (dusty!) file cabinet with black cables running to and fro...

 Since all the instruments have an internal timebase, the question becomes which one to pick as a Master Timebase? The RF generator has the optional HP "High Stability" timebase, so I picked that one. I connected it to my counter, and it read between 10,000,000Hz and 9,999,999Hz, so they agreed to one part in ten million, which is pretty good. I used the Loop Through connector to drive the 8904A Multifunction Waveform Synthesizer, and the other two outputs to drive the Spectrum Analyzer and the counter.

The internal timebase in the 8904A read 9,999,685, which meets specs, but it's several hundred Hertz low. The timebase in the spectrum analyzer measured 10,000,450Hz, which also meets specs, but is a bit high. Still, 450 parts out of  TEN MILLION amounts to fly specks, but it gives me the warm and fuzzies to have all the equipment locked to the best timebase in the house.

RF generator set to 10MHz and 0.0 dBm output.

 

Spectrum Analyzer displaying the RF generator output.

Pretty close, and to get them to "exactly" agree would require sending ALL the equipment in use to a reputable Calibration Lab, and paying a whole bunch of $$$.

I consider this to be acceptable accuracy for a home shop, and I'm aware of the ambiguities in my system. The signal generator was calibrated before it shipped, as was the waveform synthesizer. Basically the differences I'm seeing is the difference between "Meets Specs", and "Dead Nuts On".

One more little thing scratched off the list....

10" Of Snow? HAH! I Laugh At You!

 Got about 10" here, as seen on my "snow gauge".

And my son and I had it all cleared in about 30 minutes. Very dry, powdery snow like the skiers love, and the Ariens could almost blow it all the way across the street.

I'm quite pleased with how well the snowblower works, and it's pretty easy on gas, too. The last time we used it I started with a full tank, and when I refueled it after we were done, it took maybe a quart of gas. It steers very easily (YAY!) and I'm getting used to the controls and how to drive it.

The Calm Before The Storm

 Been having flurries since noon, but not much is sticking....for now. The shot below was taken at 1700 local, and at approx 1800 local we had about an inch. Expected overnight snowfall is anywhere from 6~10", with another 2~4" expected tomorrow.

So for now, we're just patiently waiting.....

Tonight's project in the workshop is to make four three-foot BNC jumper cables so I can connect all my test equipment to a single timebase generator. Three of my instruments just have the standard timebase oscillator in them, but my RF signal generator has the "High Stability" option. I'll take the output of that, run it to a little 10MHz distribution amplifier, and then to each of the other instruments. This way all the readings I make are referenced to a single timebase, ensuring their accuracy is improved.

SLW is feeling much better. Minor cough today, no fever, and no other symptoms of The Plague.

Uh-Oh.....SLW has "The Coof"....

 Well, the test kits she got at the Safeway indicates she has it, had it, or has been exposed to it. Now considering she's "Double Vaxxed" (no boosters), and has been around The Public much more than I have, she probably has a greater chance of having been exposed to it than I do.

Or will being vaxxed show you as "exposed" to the virus?

I tested negative with the other kit she brought home, even though two weeks ago I  had a nasty cold, which I thought may have been The Plague, but apparently wasn't.

Or at least the "Made In China" test kit says it wasn't.

Thought I'd read something about these test kits being about as accurate as a coin-toss, but seeing as there are several different kits, I'm sure YMMV. We just received one of the free kits from the State of Colorado, and she says she can get three more kits for free.

Be very interesting to re-test myself with one of these kits, and see what it says....

We're supposed to get 2~4" of snow Tuesday night, and another 2~4" on Wednesday, so I get to warm up the Ariens again and drive it around. Seeing as it laughed at the 6" we had earlier, it'll be interesting to see how it handles 8".......

RATS!! Now I'll Never Know.....

 Well, things are the workbench were going well with the Heathkit THD Analyzer, until the other night. I'd soldered two leads onto the circuit board pads that the photocell solders to, and this allowed me keep my Ohmmeter attached while I played with the position of the cell. This let me get a feel for how the cell picks up the light from the bulb, and how to position the cell. Having the shield cap off while playing likes this makes the resistance value of the cell go all over the place from room lighting, but at least I can see how moving the cell  up/down, right/left, in/out affects the reading, and could help me find the sweet spot where the cell should be.

The reading on the meter indicates 56,000 Ohms for this cell. The target is 28,000 to 32,000 Ohms, so it's wee bit off.

I moved the bulb as much as possible, and positioned the cell every which way to get a feel for it, and no matter how much I try, I cannot get it below 42,000 Ohms. The little foil reflectors I was paying with were required to even get it down that far. Now granted, I'm assuming this thing did work at one time, so what's changed?

I pulled the bulb out to examine it more closely, which revealed nothing. And then when I was putting the bulb back in, the "padded" tweezers I was using slipped, and TWANG!, the bulb goes sailing off to Where The Wild 10mm's Live......Buh-BYE!

The offending tool: 

Yes, some colorful language was used....

The replacement bulbs I'd ordered last week came in the day after, so all was not lost. These are small bulbs:

One thing I immediately noticed after I put the "replacement" bulb in place, is that the filament is about one-third the size of the OEM bulb:

And of course, it made no difference. I've checked all the voltages in the lamp driver circuit, and they're very close to what the schematic says they should be, indicating the the current flowing through the bulb is in the correct range, meaning the brightness of the bulb is probably correct. The other identical lamp driver circuit also has voltages that match the schematic, so it looks like things are working properly.

The other side with the four photocells adjusted just fine. The target resistance was not-too-difficult to get set, and adjusting them went about how I expected it should go. This leaves the one photocell as the last thing that's suspect. Looks like I'll have to go photocell shopping.....

Had about 6" of snow on Tuesday, and the Ariens took about 20 minutes from start to finish. It fired right up, and 6" of snow just disappeared. All I had to do was walk behind it and guide it, and I had to use a higher speed to keep the maw full, as it seemed to throw the snow better the more you stuffed into it. Too bad I didn't have it the last two winters!

A Bit Of Rearranging In The Electronics Shop

 Since I've been cranking away on projects, I decided to rearrange things on the bench a bit to make it easier to use my test equipment.

I went from this mess:

To this mess:

I've been using my oscilloscope a lot on these audio projects, and this new stack also lets me integrate a new acquisition into the mix.

The newest instrument is in the second level, on the right side. It's an HP 8904A Multifunction Synthesizer. This will be used as an external modulation source for my HP 8657B RF Signal Generator, specifically for generating the FM Stereo Multiplex Composite Waveform (whew! whatta mouthful!) so I can do a complete alignment on an FM Broadcast receiver.

The old Heathkit IG-37 that I have that basically got me back to the bench Big Time is meant to do this, but I'm not really happy with how well it works. It's a 1950's design, is really touchy to get properly aligned, and has a reputation for having a high distortion level. Thew new HP will generate a very clean, very stable waveform that will modulate the RF generator. This instrument can also produce Sine, Square, Ramp, and Triangle waveforms, and will allow you to mix them, sum them, or modulate each other. It's very close to being an "Arbitrary Waveform Generator", but not as complex.

The shop/lab/dungeon is still cluttered and messy, but I'm slowly making progress on getting it organized into a proper work area. That cardboard honeycomb core door has to go, as it doesn't give me enough workspace for equipment, tools, and Unit Under Repair. I'd like to get some countertops, and make U-Shaped workbench, but I still have to come up with a frame to hold the new work surface. 

Oh, well.....It Keeps Me Out Of Bars At Night!

Kludge or Clever?.....You Decide!

 SOOO....after I unsoldered the leads connected to the photocells, as directed to in the manual, I was able to get the photoresistors for the "Reactive" side of the bridge all set to the values listed, but the "Resistive" side of the bridge was still cranky. I went through and measured all the voltages in that section of the circuit, and they're all within tolerance per the schematic. I checked the resistors in the circuit, and they all measure within tolerance of their marked value. Everything looks "AOK", and I still can't get the photoresistor close enough to the bulb for the brightness of the bulb to lower it's value to within specs. I don't know if/how much the photresistors can drift with age, but "Bulb Aging" is a definite possibility. There's a series resistor (360 Ohms for each bulb), but  I'm hesitant to lower the value, which would make the bulb brighter, and most likely shorten it's life. Hmmmm...what to do?

I was able to track down the industry part number for the bulbs used, and they're an "1869D" bulb, still available. The range in price from Nine Dollars per bulb, to Ninety Cents per bulb, depending on vendor. I bought a couple of "10 packs" from the ninety cent guy, as it turns out they use these bulbs quite a bit in various audio oscillator circuits as part of the amplitude stabilization loop, and sure enough, the same bulbs are used in one of my Heathkit Sinewave Generators. Probably a good idea to have a few spares around here...

But in the meantime, what could I do, if anything, to make what's on-the-bench get back to where it's supposed to be?

The first thing I tried was a small piece of paper behind the bulb to position it as close as I could to the photocell (yeah, I got tired of typing the whole name), and saw some improvement.

But was it the position of the bulb, or the white paper? How's about if I try and make the most of the light the bulb produces?

I cut and bent a small piece of Aluminum foil, and wedged it in, and the value of the photocell dropped to almost in range:

So for now, it looks like by changing the size and shape of the new "reflector" in there, I should be able to get this last one adjusted to be in range. And yes, I'll tack the parts down so they don't move around and affect the adjustment.

SO, dear readers, it's up to YOU! Is this a "Kludge", a "Kluge", or a "Clever"?

As long as it works, I'll leave it "AS-IS".....

Olde Skool Tech.....In Action!

 AKA "I'll take Electronics I learned in Grade School for $1000, Alex!"....

Even though the copyright on the manual for this Heathkit THD Analyzer I'm restoring is dated 1978, I found a notation in ball-point on the underside of the chassis stating it was built in 1979. The actual circuit has been around "forever", and I've read in places this instrument is a virtual clone of the Hewlett-Packard 3xx series of Analyzers, as mentioned here on trona.com :

HP 331A thru 334A series (and Heath IM-5258)
These are solid-state, discrete-transistor designs; no opamps. They are Wien Bridge units with very wide range; in the HP units, up to 600kHz fundamentals(!!) and wide measuring bandwidth, above 3MHz except on the most-sensitive range. They offer low noise, and, in the 333 and 334 (and Heath unit), auto tuning. They employ average-responding metering. They have either a 400Hz passive LC high pass filter that is very good, or a passive 30kHz low pass LC filter, also very good, but not both filters. They desperately need expanded low pass filtering for better performance in the audio mid-band. Max full scale sensitivity is 0.1%, fully usable to 0.01% and below, except that their residual THD floor is around 0.01% (I've been told that the Heath has a higher floor but I've not used one). In the HP units I've used, this floor is dominated by 2nd harmonic distortion, which comes from the bridge amplifier/notch feedback system. I've never checked the depth of the fundamental null. I really should but I have good reason to believe that it is better than -90dB and may approach -100dB. Except for their high residual THD floor, and somewhat insensitive metering (300uV/0.1% full scale), the low noise and nulling of these analyzers may make them useful for measurements to 0.003% or even below.

So, looking up the HP Analyzers, we see that these are 1960's designs, all analog (or "Linear" BITD), and use very sound design principles. One of the things mentioned is that they have an "Auto Null" feature (called Auto Tuning above) that uses a servo loop to make much finer adjustments to the Wein Bridge than a human operator could. Imagine another set of tuning controls, being driven by a mechanism that can detect smaller differences than humanly possible, and tunes the other set of controls with much greater resolution than a human could.

Well, these "controls" are made of Photoresistors. Good, old Cadmium Sulphide Light Dependent Resistors, going back to at least the 1930's. They have a "Dark Resistance" of many Megohms, and a "Light Resistance" of anywhere from a few hundred Ohms to a few hundred thousand Megohms.

Think of them as a "Light Controlled Volume Control", and you're very close to having the concept.

And we have five of them in the Heatkit Analyzer:

The lonely little one on the left controls the "Resistive Component" of the bridge, and the four on the right control the "Reactive Component" of the bridge. You can see the lamps glowing dimly. These normally have black plastic covers over them, but to adjust the "idling" value of them, you have to move them closer or farther away from the lamp.

This is with the cover over the Reactive Component sensor in place:

The procedure is very tedious. Measure LDR with cover in place. If adjustment needed, remove cover and reposition the LDR. Install cover and measure LDR. If adjustment still needed, rinse, lather and repeat procedure until LDR measures the required value.

I measured them, and all were WAY off. I experimented with a couple of them, and they don't seem to want to adjust within range. Rather than freak out and order all new ones, I'm going to examine the circuit driving the lamp. They all seem to need more brightness to get down to the correct value, and that seems a little suspicious to me, so I'll do all the voltage checks of the circuits involved, and see if something is holding the lamp brightness down. I think it's a safe guess to assume these moved around a bit during the jolt the chassis received that bent up a few things.

Heathkit IM-5258 Total Harmonic Distortion Analyzer Update

 As promised, here's the wrap-up on this section of the progress.....

The original problem was that the negative voltage regulator integrated circuit had failed, producing -30VDC instead of the -24VDC it was supposed to produce. The regulator IC is an obsolete part, and while they are available, I wasn't willing to pay the going price. Since this isn't a "Pebble Beach" restoration, I decided to use modern parts and build two individual voltage regulators, one for the positive voltage, and one for the negative. Between waiting for parts and getting distracted by other duties, it took me two weeks to get here.

The two "daughter boards" ready to be installed:

The sharp eyed will notice I reduced the size of the heatsinks compared to what I had mocked-up, showed propped up on the screwdriver handle.. The original regulators were rated at 500mA, and these are rated at 1500mA, meaning they can safely regulate three times the current the OEM parts could handle. From some looking at the specs of this thing, I figured out the circuit only draws a few hundred milliamps of current, and these should do fine with a smaller heatsink. 

So here they are all moved into their new home:

 The positive regulator has an output of 24.35VDC, and the negative regulator has an output of -24.30 volts, so they're within 50 millivolts of each other, and well within the original spec of "24 Volts +/- 1.5 Volts", which is an absurdly loose spec for an electronic  voltage "regulator".

 I left it powered up with the cover on for about 45 minutes, and my IR thermometer shows the heatsinks get up to about 120*F, while the TO-220 packages read about 103*F, so the "small" heatsinks are adequate.

So this part of the project has been completed, and I can now pick up the adjustment/calibration procedure at Step 4.  I stopped when it failed Step 3, which was to measure the regulator outputs.

Now on to exorcising the next gremlin......Yes, this is a rabbit hole I've fallen into, and I'm NOT expecting the adjustment/calibration procedure to go smoothly.

Getting this thing functional again has almost turned into a matter of pride......