Hi Kerry I really have enjoyed your videos and appreciate you sharing not only how things work, but also the elusive art of troubleshooting. It’s really hard to teach and I use your videos as practical examples for those I mentor. Regarding the HP856X SAs, these boat anchors are still tough to beat from a noise floor and selectivity perspective, especially for a street price that’s often well under $1000. I have both an 8566B and a 8668B, plus the preselector, quasi-peak adapter and tracking generator - all that I bought used “as is” and were easily repaired since factory documentation is available for free. One thing HP did was make their products repairable which is why these old beasts are still popular. The newer products are difficult to repair and the documentation sparse. They are meant to be disposable or like R&S SAs, repaired at great expense. While you have the unit on the bench, buy a replacement lithium battery from DigiKey for the memory and replace it proactively. It’s an easy service operation (two solder joints) on the pull-out CPU board. Also, you can marvel that the whole unit runs on an old 68K CPU running at single digit MHz clock speeds.
Thanks! Yeah, one main issue of today's test gear is that most of the logic is implemented in FPGA or software, making them almost impossible to service hardware-wise.
Kerry, was looking forward to this but somehow missed it when you posted it. Thanks for the debugging experience. Very cool & I'm really happy you found the problem. I tore down my 8566B a year or so ago and located the 1990 Vintage Tadiran battery used on the logic based. I measured it and was amazed to find that it was exactly spot on 27 years after manufacture. You can still buy the batteries on DigiKey for a very reasonable price, however, I decided to leave mine alone.
Thanks for sharing Kerry, glad you found the issue...I've been procrastinating digging into my IFR1500 service monitor that is having a RF generator output issue and you've nudged me closer to tearing into it!
Potentiometer from Vishay they do special TC Pots, ask for samples ;) you could clean up in a ultrasonic bath but only use very small ammount of cleaner in a bag, then the rest fill up with water, it should work, but Free Samples are great, tell them your wanting to check stability of a design using this component,
Hello Kerry, thank you for this video! I used section 28:47 of your video to fix the YTO UNLOCK error on my newly acquired 8566B. May I ask where you acquired the schematics? The ones I have found online and through Keysight have been difficult to use.
The service manuals were from Keysights' website. If I remembered correctly I had to print them onto multiple pages as otherwise the schematics would be too small to read.
Great video, nice work! It would have been interesting to put the meter side by side with the display to see if the voltage fluxuation and the YTO Unlock message occurred at the same time.
BTW the Plastic Top Hat stand - offs are for Air Circulation, as you can see on the WW Resistor, some HP's use ceramic tubes, The Resistors would burn the PCB IF the Resistors are in contact with the PCB hence why they take the Resistors off up in the air, Retired Aerospace Hardware Qualification Engineer, - UKAS Test Equipment Manager, PS OLD HP is very well built when HP was HP & Not Agilent - Keysite :(
Interesting video. My first suspect before even viewing was some amplifier and that seemed to be confirmed at first when that -30dBm signal you were getting popped up. Caps were the second thing on my mind. However, pots have been known to cause problems too. Especially ones that have stayed in a specific spot/turn for years on end. I've heard about this type of failure a few other places before but this is definitely added to my list of things on what to watch out for since its becoming more common in my view. I wonder what the true failure of that pot was. We can always claim aging but I believe it goes well beyond that. Like at the physical/chemical/microscopic level. Heat in particular is great at introducing either thermal stress or pushing a chemical reaction. I'm thinking something synomynous to what happens to an incandescent light bulb eventually plus added pressure of the contact to winding/carbon film. Still, it did its job considering this is equipment from the late 70s/mid 85s. Glad you've fixed it. Looking forward to a continuation of the tracking gen vids.
I think the true failure mode (at least from what I could see) is that the resistive coating material had started peeling off at various spots due to the age.
That can happen too. Synonymous to say iron oxide peeling off the plastic backing of some older HP magnetic tapes. A.K.A "sticky-shed syndrome". In that case, a combination between the binding agent chemically degrading and thermal stress. Instead of water vapor attacking maybe oxygen itself plus years of heat? Depends on how well sealed the pot is. Could even be the lubercant for the worm gear inside or other introduced "materials/impurities" during manufacturing. Interesting stuff. All things considering, as I said 30+ years or so. It did its job. These pieces of kit are awesome. 8566B isn't the only one with that design. I can confirm that the 8340 syn. sweep models look exactly like that layout while the 8673B/D is just a bit slightly different. The only difference being that the layout is topside access rather than bottom. Had there been more to the YTO problem, you've would have possibly done something interesting. That is the entire YTO loop assembly is designed to "stand up" for either diagnosing or calibrating more trimmers.
Nice fix! My approach would be to replace this trim pot with 3 or 4 high accuracy fixed resistors. Once correctly adjusted you will never go back there to re adjust in your lifetime.
Hi Kerry, have you tried to use some DeoxIt on this potentiometer? It might not be enough (especially it miht not be easy to make the fluid penetrate the case of the trimmer), but it worth a try (IMHO deoxit work pretty well).
If you have the manual, why didn't you use the manufacturer's suggested troubleshooting plan for specific error conditions? The HP manuals for that era will tell you what tolerance each value should be (especially for critical tuning voltages). BTW, you should have tried spraying some control cleaner onto that potientiometer's shaft (it probably isn't a hermetic seal) and then adjusting the control from end to end several times. It's an easy trick, and it often will revive a bad potentiometer.
The suggested troubleshooting is for those boards (A19-A21). I didn't find any tolerance figures on the power supply rails. But again, the service manual is 900+ pages long and the PDF version I have is not search-friendly so I couldn've missed that somewhere.
The HP-8566B was intended to absolutely dominate the spectrum analyzer market, both in performance and reliability (and possibly in weight too). I bought an 8571A Receiver system (that used the 8566B plus pre-amps, custom IF's & filters and a tracking pre-selector) in 1995, and after 17 years of daily use, it had not exhibited even one failure. While new analyzers (being more of a computer than a SA) offer tremendous math, waveform functionality and speed of measurements, an 8566B on your bench is still a great piece of gear and offers an analog sanity check on today's digital magic. BTW, there are some companies which offer an LCD replacement for the CRT displays in these classic HP analyzers (you delete the CRT, HV power supply and magnetic shields). See ua-cam.com/video/OT8pfr4a2K8/v-deo.html
Agreed, even today there are not many SAs can do 22GHz. I saw the LCD replacement a while ago, my thoughts are that if the original one works, why changing the CRT? it doesn't look like to an easy replacement and it cost quite a bit as well...
My company had a number of HP-8560 series SA's (and some HP-4192 Network Analyzers which used the same CRT) that had experienced CRT failures (typically production test instruments that often displayed the same waveform all day long), and CRT's were the component that HP ran out of in their support chain. I believe HP placed a re-order for CRT's with a new vendor, but the new batch of CRT's were just a bit different, so HP had to supply additional circuitry to make them "drop-in" replacements. This was more hassle to HP than it was worth, and besides, every fixed old SA was a new SA that never got bought. We had found a USA company, near Carlsbad, California, that sold display replacement kits for about $700 (circa 2010). Installation was fairly simple, maybe 3 hours labor. I was in favor of refurbishing about a dozen SA's, but changing business requirements led us to buy several cheaper, new SA's and those old CRT SA's were scrapped. BTW, a good path for capability above 20 GHz is to use external harmonic mixers. True, harmonic mixers are delicate and wide-open, but you can often find these orphans at a very good price. I used an HP-8562A SA with a set of HP-11970 Mixers that gave me coverage out to 75 GHz. These also work with the 8566 SA, but I never used that combination. HP-11970K 18 to 26.5 GHz HP-11970A 26.5 to 40 GHz HP-11970Q 33 to 50 GHz (I didn't buy this one) HP-11970U 40 to 60 GHz HP-11970V 50 to 75 GHz HP-11970W 75 to 110 GHz (I didn't buy this one)
There is only one aluminium electrolytic on each of those boards so I would had gone ahead and replaced both of them. Those Kemet glass-sealed tantalum capacitors rarely fail.
The standoffs on those big carbon comp resistors on the supply boards look to actually be TO package screw insulating washers (basically plastic shoulder washers). That's really amusing.
Those were pretty stable, I think it might be me not making steady contacts. At the end though, it doesn't seem that these voltages caused the YTO unlock problem.
I have SA N9320A like video: ua-cam.com/video/td9SOOgdzls/v-deo.html Sometime it arlam LO UNLOCK then I can not receive any signal I will open the packet but It looks difficult to open Does anyones has experience open this device? thanks
Hi Kerry
I really have enjoyed your videos and appreciate you sharing not only how things work, but also the elusive art of troubleshooting. It’s really hard to teach and I use your videos as practical examples for those I mentor. Regarding the HP856X SAs, these boat anchors are still tough to beat from a noise floor and selectivity perspective, especially for a street price that’s often well under $1000. I have both an 8566B and a 8668B, plus the preselector, quasi-peak adapter and tracking generator - all that I bought used “as is” and were easily repaired since factory documentation is available for free. One thing HP did was make their products repairable which is why these old beasts are still popular. The newer products are difficult to repair and the documentation sparse. They are meant to be disposable or like R&S SAs, repaired at great expense. While you have the unit on the bench, buy a replacement lithium battery from DigiKey for the memory and replace it proactively. It’s an easy service operation (two solder joints) on the pull-out CPU board. Also, you can marvel that the whole unit runs on an old 68K CPU running at single digit MHz clock speeds.
Thanks! Yeah, one main issue of today's test gear is that most of the logic is implemented in FPGA or software, making them almost impossible to service hardware-wise.
Kerry, was looking forward to this but somehow missed it when you posted it. Thanks for the debugging experience. Very cool & I'm really happy you found the problem. I tore down my 8566B a year or so ago and located the 1990 Vintage Tadiran battery used on the logic based. I measured it and was amazed to find that it was exactly spot on 27 years after manufacture. You can still buy the batteries on DigiKey for a very reasonable price, however, I decided to leave mine alone.
Yeah, I wanted to replace the battery in mine as well but it appeared to be in pretty good shape so I also left it alone.
Kerry, thank you for your time and sharing with us. Hope you find the correct replacement for the cap and pot. Your friend Joel.
Thanks Joel!
Thanks for sharing Kerry, glad you found the issue...I've been procrastinating digging into my IFR1500 service monitor that is having a RF generator output issue and you've nudged me closer to tearing into it!
IFR1500 is such a nice compact service monitor I almost got one before I got my 8566B.
I love the quality of these old Hewlett Packard bit of equipment. It's a shame they don't make things this well today.
Potentiometer from Vishay they do special TC Pots, ask for samples ;) you could clean up in a ultrasonic bath but only use very small ammount of cleaner in a bag, then the rest fill up with water, it should work, but Free Samples are great, tell them your wanting to check stability of a design using this component,
Hello Kerry, thank you for this video! I used section 28:47 of your video to fix the YTO UNLOCK error on my newly acquired 8566B. May I ask where you acquired the schematics? The ones I have found online and through Keysight have been difficult to use.
The service manuals were from Keysights' website. If I remembered correctly I had to print them onto multiple pages as otherwise the schematics would be too small to read.
Great video, nice work! It would have been interesting to put the meter side by side with the display to see if the voltage fluxuation and the YTO Unlock message occurred at the same time.
BTW the Plastic Top Hat stand - offs are for Air Circulation, as you can see on the WW Resistor, some HP's use ceramic tubes, The Resistors would burn the PCB IF the Resistors are in contact with the PCB hence why they take the Resistors off up in the air, Retired Aerospace Hardware Qualification Engineer, - UKAS Test Equipment Manager, PS OLD HP is very well built when HP was HP & Not Agilent - Keysite :(
Interesting video. My first suspect before even viewing was some amplifier and that seemed to be confirmed at first when that -30dBm signal you were getting popped up. Caps were the second thing on my mind. However, pots have been known to cause problems too. Especially ones that have stayed in a specific spot/turn for years on end. I've heard about this type of failure a few other places before but this is definitely added to my list of things on what to watch out for since its becoming more common in my view.
I wonder what the true failure of that pot was. We can always claim aging but I believe it goes well beyond that. Like at the physical/chemical/microscopic level. Heat in particular is great at introducing either thermal stress or pushing a chemical reaction. I'm thinking something synomynous to what happens to an incandescent light bulb eventually plus added pressure of the contact to winding/carbon film.
Still, it did its job considering this is equipment from the late 70s/mid 85s. Glad you've fixed it. Looking forward to a continuation of the tracking gen vids.
I think the true failure mode (at least from what I could see) is that the resistive coating material had started peeling off at various spots due to the age.
That can happen too. Synonymous to say iron oxide peeling off the plastic backing of some older HP magnetic tapes. A.K.A "sticky-shed syndrome". In that case, a combination between the binding agent chemically degrading and thermal stress. Instead of water vapor attacking maybe oxygen itself plus years of heat? Depends on how well sealed the pot is. Could even be the lubercant for the worm gear inside or other introduced "materials/impurities" during manufacturing. Interesting stuff.
All things considering, as I said 30+ years or so. It did its job. These pieces of kit are awesome. 8566B isn't the only one with that design. I can confirm that the 8340 syn. sweep models look exactly like that layout while the 8673B/D is just a bit slightly different. The only difference being that the layout is topside access rather than bottom. Had there been more to the YTO problem, you've would have possibly done something interesting. That is the entire YTO loop assembly is designed to "stand up" for either diagnosing or calibrating more trimmers.
Nice fix! My approach would be to replace this trim pot with 3 or 4 high accuracy fixed resistors. Once correctly adjusted you will never go back there to re adjust in your lifetime.
Thanks. If I ever need to replace the pot again, I will probably do that.
Great repair, Kerry!
I think those caps are wet tantalum and are usually better as you say than a modern budget capacitor.
Yeah, it's amazing those original Kemet caps are still in excellent shape.
Hi Kerry, have you tried to use some DeoxIt on this potentiometer? It might not be enough (especially it miht not be easy to make the fluid penetrate the case of the trimmer), but it worth a try (IMHO deoxit work pretty well).
I tried with some QD electronic cleaner but it didn't quite work so I didn't pursue cleaning the original pot further.
Excellent video. Thank you.
If you have the manual, why didn't you use the manufacturer's suggested troubleshooting plan for specific error conditions? The HP manuals for that era will tell you what tolerance each value should be (especially for critical tuning voltages). BTW, you should have tried spraying some control cleaner onto that potientiometer's shaft (it probably isn't a hermetic seal) and then adjusting the control from end to end several times. It's an easy trick, and it often will revive a bad potentiometer.
The suggested troubleshooting is for those boards (A19-A21). I didn't find any tolerance figures on the power supply rails. But again, the service manual is 900+ pages long and the PDF version I have is not search-friendly so I couldn've missed that somewhere.
The HP-8566B was intended to absolutely dominate the spectrum analyzer market, both in performance and reliability (and possibly in weight too). I bought an 8571A Receiver system (that used the 8566B plus pre-amps, custom IF's & filters and a tracking pre-selector) in 1995, and after 17 years of daily use, it had not exhibited even one failure. While new analyzers (being more of a computer than a SA) offer tremendous math, waveform functionality and speed of measurements, an 8566B on your bench is still a great piece of gear and offers an analog sanity check on today's digital magic.
BTW, there are some companies which offer an LCD replacement for the CRT displays in these classic HP analyzers (you delete the CRT, HV power supply and magnetic shields).
See ua-cam.com/video/OT8pfr4a2K8/v-deo.html
Agreed, even today there are not many SAs can do 22GHz. I saw the LCD replacement a while ago, my thoughts are that if the original one works, why changing the CRT? it doesn't look like to an easy replacement and it cost quite a bit as well...
My company had a number of HP-8560 series SA's (and some HP-4192 Network Analyzers which used the same CRT) that had experienced CRT failures (typically production test instruments that often displayed the same waveform all day long), and CRT's were the component that HP ran out of in their support chain. I believe HP placed a re-order for CRT's with a new vendor, but the new batch of CRT's were just a bit different, so HP had to supply additional circuitry to make them "drop-in" replacements. This was more hassle to HP than it was worth, and besides, every fixed old SA was a new SA that never got bought.
We had found a USA company, near Carlsbad, California, that sold display replacement kits for about $700 (circa 2010). Installation was fairly simple, maybe 3 hours labor. I was in favor of refurbishing about a dozen SA's, but changing business requirements led us to buy several cheaper, new SA's and those old CRT SA's were scrapped.
BTW, a good path for capability above 20 GHz is to use external harmonic mixers. True, harmonic mixers are delicate and wide-open, but you can often find these orphans at a very good price. I used an HP-8562A SA with a set of HP-11970 Mixers that gave me coverage out to 75 GHz. These also work with the 8566 SA, but I never used that combination.
HP-11970K 18 to 26.5 GHz
HP-11970A 26.5 to 40 GHz
HP-11970Q 33 to 50 GHz (I didn't buy this one)
HP-11970U 40 to 60 GHz
HP-11970V 50 to 75 GHz
HP-11970W 75 to 110 GHz (I didn't buy this one)
There is only one aluminium electrolytic on each of those boards so I would had gone ahead and replaced both of them. Those Kemet glass-sealed tantalum capacitors rarely fail.
The standoffs on those big carbon comp resistors on the supply boards look to actually be TO package screw insulating washers (basically plastic shoulder washers).
That's really amusing.
Hey Connor! It's nice to hear you. Haven't seen your video for a long while! Hopefully you will make one soon!
Thanks for sharing your knowledge! :)
You are welcome!
Are you sure that this is Electrolitic cap ? i think it is non polarity ones.... ( polyester)
Wet slug tantalum have very low ESR, and unless they have been subjected to some weird event, they last more or less forever.
The 11v and 54v were unstable, so could be poor regulation.
Those were pretty stable, I think it might be me not making steady contacts. At the end though, it doesn't seem that these voltages caused the YTO unlock problem.
Опять твой HP сломался ?
I have SA N9320A
like video: ua-cam.com/video/td9SOOgdzls/v-deo.html
Sometime it arlam LO UNLOCK then I can not receive any signal
I will open the packet but It looks difficult to open
Does anyones has experience open this device?
thanks