My home vintage electronics lab consists mainly of Hewlett-Packard test equipment from the 1960s -1980s, although I may occasionally highlight other stuff. I collect these magnificent old instruments, restore them to their former glory, and put them to use in my lab.
Nice, I also have a 3325A and just did this test myself. I used a HP 70K series SA and discovered using the same parameters of CF 1MHz, Span 1MHz, RBW 300Hz and 1kHz VBW my 3325A @ 1MHz, 0dB out measures -89 to -88.42dB @ 400kHz from the carrier. I have also measured the phase noise performance of the 3325A that I have and have discovered that it is well within the specifications listed in the service manual. These instruments will last seemingly forever if properly cared for. *edit* Guess I'm going to have to get a "lab coat" now. :)
Thanks for your presentation. If you look at the circuit closely you will see that the two heater/thermistor pairs are connected back to back in the feedback loop of the op-amp - this is not for improving sensitivity as you state. This connection accomplishes the AC to DC conversion and nulls out any common mode errors due to ambient temperature changes etc. Jim Williams was a big fan of this instrument, its elegant operating principle was used in an IC he had a hand in designing.
great video, usefull and simple idea. Just a note: to have the same polarity on both channels use the low side of the transformer(0V) as common ground and the black probe/1k union as Y out.
That certainly does make sense! I don’t really know why I didn’t do that from the beginning. There isn’t any need for the black probe to be at ground potential. Thanks for the tip!
Nicely built! I really like it. Got a curve tracer in my old Hameg 204 scope. I also gutted an old single-channel scope and repurposed it as a curve tracer among other things, using schematics from Paul Carlson.
I used to perform electronics repairs 35 years ago; when a customer would bring in a stereo with one dead channel, my curve tracer was the fastest way to identify bad or weakend components in the dead channel, simply by comparing the traces of the two channels. A very handy tool.
Those things are awesome, Hameg scopes had them built in, they made me a fortune as a repair tech, once you have a little understanding of what they show you and you're used to the shapes you see it's so fast and easy to sprint through analogue components in power supplies, amps, monitors, TVs, all sorts of stuff, you can spot leaky caps, failed diodes, transistors, FETs and all sorts of semiconductors almost instantly
I miss the old scopes. Even though the new digital scopes are far more capable, there's something about looking at a nice blue/green trace that's directly connected to the signal instead of what is effectively just another image on a pc screen. Oh, and you didn't have to listen to a constantly whining fan. Plus they've done away with the Z input.
I added a Leo Bodnar Electronics GPSDO to my bench about a month ago. The Bodnar unit is made in the UK and customer support exists, though I have not needed support. There are a few models to pick from, and I opted for the dual BNC output which is about the size of a US cigarette pack. I chose this brand over the no-name Chinese units which are common on eBay and Amazon which I assume are throw-away items if problems arise. Bodnar refers to their GPSDO as a “clock” but they mean that in the sense of strictly a system frequently reference. The two outputs are independent and programmable from 450 Hz to 880 MHz using a USB set-up program. Once set. The unit no longer needs to be plugged into a computer. The power input is 3.3 to 16 Volts DC and a small USB supply is a good power source. I found the unit locked on in about 15 seconds with the antenna puck resting on my bench, which is not near a window. The stated accuracy is 1/10^11 and a pretty low phase noise error. My service monitor and spectrum analyzer have OCXOs, but with the Bodnar unit, I was able to improve the measurement accuracy by about two orders of magnitude. The GPSDO reference unit is an accuracy game changer, and now super precise frequency measurements are within reach of most hobbyists.
Gotta love these fairly simple repairs. Many of these sub-assemblies contain custom HP unobtanium parts that can only be acquired from a doner unit. I have several units that I have simply replaced the entire sub-assembly because I didn't have the test gear to properly troubleshoot the issue. I'd love to get my hands on a 3585A eventually.
those little tones at -70dbc should not exist. they seem to be about 20khz apart and not harmonically related to the 1MHz input. Either the 3325a or 3585a still has a problem.
Yes, you are correct. When I use an analog oscillator as my signal source, those spurious signals disappear, so the problem isn’t in the 3585A. I am aware of this situation, it is a problem in the fractional-N PLL circuit on my 3325A. That repair may be coming up in a future video.
These are great old instruments, I also have one of these meters, it is mostly used to measure power supply rail noise and receiver sensitivity testing. Mine also uses a Nuvisor.
Nice little useful project. I have an old EICO 688 that I have used on and off for many years. However, I was very fortunate to find a mint condition Tektronix 571 curve tracer that has been quite the bees knees. But I still use the EICO in some special use cases. Great instructional video.
This is a rip off. There is a cheaper 1/3 the price with LCD with a deviation value from the crystal oscillator in ppb and count. The most relible peice of gadgets that I own.
There are a plethora of Chinese GPSDOs on offer of Amazon and eBay. I suspect some of those units are using recycled Trimble boards which were originally in some other equipment. Despite the verbose listings, it is difficult to get a true sense of what you’re buying because the listings are clones of each other. The generic GPSDO cost isn’t huge, but it isn’t zero either. I don’t feel comfortable with the no-name eBay GPSDOs. I got a Leo Bodnar Electronics GPSDO for my bench last month. The Bodnar unit is simply a programmable frequency standard with 450 Hz - 880 MHz settable output. I feel better buying the made in UK Bodnar unit because the cost is about the same and I believe Bodnar will be around if support should be needed.
Interesting bit of kit. I see they have been extensively discussed on EEVblog pages and have been around in various versions since about at least 2015. Interesting that yours has no year-month-day text marking after the BG7TBL on the front panel. Guess there are many copies of the original product with it being difficult to see if its by the original designer/producer.
What did you replace them with? They are still for sale and used in many devices. The next best thing, a TCXO, is not nearly as stable, but much smaller and needs less current.
The ALM light is supposed to mean "Don't rely on the output frequency". The light is supposed to go off after the GPS signal has been stable for long enough that the PLL has locked and the oscillator has been tuned on frequency. The GPS Lock light means it can decode the GPS satellite signal. The blinking RUN light indicates that the software is working. I've found the PLL in these do not lock exactly on frequency and they wander by millihertz. They're close enough to use as an OK frequency standard, but they're not nearly as good as they should be. That might not be a problem if you're just using it to put your VHF ham radio on frequency or using it in place of a quartz oscillator on your 1 Hz resolution frequency counter, but if you use these as a timing source the errors compound and you'll find that your time of day clock drifts significantly and randomly where using a better unit like a Trimble Thunderbolt will stay within nanoseconds day after day. I've seen mine drift over 100ns in 30 minutes. This drift is not what one expects from an atomic referenced standard. I've found the smaller black versions of these ebay GPSDOs can still be significantly off frequency by many hertz when the ALM light first goes off. If you're wanting this for anything serious, get a used Trimble Thunderbolt instead.
I’ve been wondering what is inside the generic Chinese GPSDOs for sale on eBay and Amazon. I guessed they are using recycled Trimble modules out of other equipment. I’ve been using a Leo Bodnar unit as the frequency standard for my service monitor and spectrum analyzer. Bodnar states the frequency accuracy is 1/10^11. The phase noise is acceptable, but I’d have to look it up on Bodnar’s web site. I’m not generating SMPTE time codes or standardizing an NTP server, so I’ve found the Bodnar unit acceptable for a very reasonable cost.
Nice. I also just got a GPSDO, but I went with one of Leo Bodnar's lbe-1420. It works great, and is super easy to use, but I do wish it had multiple outputs. My antenna is just sitting outside of a window on the windowsill right against the house and I am picking up 23 satellites. Currently set to output 24Mhz to be used with my SDRplay RSPDx.
The Leo Bodnar unit is the way to go if you need something other than 10 MHz. My antenna is just sitting inside the window sill and works great in that position
I got the dual BNC Bodnar unit. Part of the PLL divisor chain is shared between the two outputs, but the set-up program figures out how to shuffle the numbers to come up with the two frequencies you want - if you want two different frequencies, that is. I’ve been very pleased with the results so far. If all you need is a frequency standard, and don’t care about 1 PPS, SMPTE time codes and NTP serving, then the Bodnar units are the way to go for a very affordable cost. Better yet, they’re made in the UK by a company which will be there if you need to ask for support.
I'm glad you showed us this. Oscillators these days show no resect for authority or others. It's honesty a huge problem. It's nice to know that there are still disciplined Oscillators out there. 😊
Very nice. The GPS antenna doesn't have to be on the roof. It only needs to see the sky, so that can be done through almost any window facing outside. Most of the time if you just sit the antenna on the windowsill it will work just fine. The alarm LED is likely tied in to the crystal oven and will be lit until it reaches temperature.
Many windows today have low-e coatings, which is a layer of metal on the glass. They're not transparent to microwaves. So next to a window might not be the best place. The GPS receivers in these are sensitive enough to work through the (non-metal) roof of a house. Mine works in my basement workshop.
@@stargazer7644 Good grief. There is always something new to put a spanner in the works. My windows don't have any special coatings on them and a GPS antenna sitting on the sill works perfectly. In fact I have a LED digital clock with GPS receiver and it sits above the fireplace against a solid wall and has no problems locking on to a dozen satellites from there - and it's nowhere near a window. I can only suspect it's receiving through the wall.
@@sw6188 The windows affected are advertised as energy saving "Low-E" or "Low Emissivity" windows. If they're tinted, you may have the same problem as some window tint (like used in office buildings) is very thin metallic film. The walls may also be a problem if you have foil backed insulation board (celotex) in the walls, or plaster or stucco walls with metal mesh in them, depending on the size of the mesh. Usually the roof is fine if you have wood construction and asphalt shingles. I have low-e windows and celotex in the walls of my house.
Prices vary. You can check the prices on eBay, I’m sure they’re also available from other sites too. It’s a surprisingly sophisticated piece of tech for not very much money.
If you don't mind ordering direct from China the lowest price on eBay this morning is about $70. The lowest price I see from a U.S. address is about $90. There are quite a few options in the $200 range so take your pick.
I don’t have the actual paper copy of the manual for this instrument, which would have the schematic. The block diagram came from the Keysight website, but their scanned image of the schematic was of very poor quality, and I didn’t think it would be helpful.
Nice explanation. Just a note, later versions of the HP 3400A did use a JFET in the Impedance Converter stage and not the Nuvistor. Nuvistor A2 assembly P/N 03400-66502 and later JFET A2 Assembly P/N 03400-66514.