HP 8568B 1.5 GHz Spectrum Analyzer Repair
ฝัง
- เผยแพร่เมื่อ 27 ก.ย. 2024
- Our HP 8568B 1.5GHz Spectrum Analyzer is broken again (we first repaired it 8 years ago!). The CRT display has gone bonkers. And its unusual Lithium Thionyl battery needs to be changed, followed by a recalibration. Let's fix all this.
Previous "YTO Unlock" repair series from 8 years ago: • HP8568B Spectrum Analy...
Our sponsors
- PCBWay: fast turn PCBs, www.pcbway.com
- Keysight: test instruments: www.keysight.com
- Samtec: connectors: www.samtec.com
Support the team on Patreon: / curiousmarc
Buy our merch on Teespring: teespring.com/...
Get more info on the companion site: www.curiousmar...
Contact info: www.youtube.co...
![[🔴LIVE] งานประกาศรางวัล MAYA TV AWARDS 2024 : มายามหาชน](http://i.ytimg.com/vi/nwmWYXXVwBY/mqdefault.jpg)
Recapping is often seen on TH-cam these days (as the subtitle ironically suggests...) but this is the only place where the dead cap gets spliced on a lathe for further analysis... Amazing, informative and entertaining as always. Bonne année 2022 !
Very nice to see what's wrong inside a dead cap.
If I had gone into the video at the cutting open of the cap I'd first consider Big Clive to be involved.
Amen to that!
Nicely done Marc! Thanks for taking the time to share this with us.
The fight against entropy continues
That's the drill. Repair them faster than they break. Not always easy to stay ahead!
Happy New Year! So it turned out that wasn't a mistake in the schematic, HP have indicated the dry capacitance, 4700pF = 4.7nF, you got 6nF. Close enough. :)
My thought too!
They really do think of everything!
@@zyeborm*Indeed. Dr. Siegfried Linkwitz and his engineering design team certainly did.*
One way to avoid problems when re-lacing memory-backup batteries is to:
1) measure the voltage of the battery that’s in there now, and set that voltage on a bench power supply. Then hook the power supply to the circuit in parallel with the battery.
2) remove the old battery, and measure the voltage on the new, replacement, battery.
3) gently crank up the voltage on the bench power supply until it matches the new battery voltage, then solder in the new battery.
4) you can now remove the power supply.
Happy times. 🤗 (Unless the old battery drained so much the memory was erased anyway. Then sad times. 🥺)
I didn't know batteries could even get stretched laces.
William I know that you are extremely clever but please check your posts for spelling mistakes.
I suspect that Marc and his viewers will have worked out how to preserve the calibration if it was not lost.
Sounds like a normal day when I am in the lab, fix two things to fix one thing, except I usually break one other thingy in the process 😊
HP/Agilent/Keysight should underwrite your work, Marc. I marvel at your inquisitiveness and perseverance.
And they do! They are one of the channel new sponsors. If you see a (very expensive!) 1 GHz Keysight 4-channel scope, it's one of their donations to the channel, and I loove it! We are currently rounding up more even help for more Apollo RF testing equipment, in part through their equipment rental partner Electro-Rent, also a new sponsor.
@@CuriousMarc Cool, and congratulations! I get the impression that you are HP's #1 salesman.
Another fine episode. I used one of these for over a decade at my work repairing Data Link Test Sets. It was my favorite Spectrum Analyzer in the shop. Loved the green screen and doing the Recall 8 cal. We had to repair it once as well, including the battery. I did learn something on the external monitor, we never had to use it so I never knew it was there. Thanks for sharing. Best Wishes & Blessings. Keith Noneya
Failure analysis of the defective capacitor was fantastic! I've never seen a cap cut open with a metal lathe before. Gripping and fun.
Remembers me of my job 30 years ago, fixing all these beautiful hp instruments.
*Been there and done that too. Respect!*
Great for bringing a HP8568B back to life. Used one one these for nearly 2 decades to design, test and build telemetry systems for rotating components in gas turbine R&D programs.
I don't remember how many of these I've repaired and/or calibrated over 25 years, but I nearly memorized the schematics and service manual!
I used Hewlett Packard RF equipment every day of my whole working life. The instruments improved dramatically up to this series. I was using an 8566B alongside the later replacement models right up to last year when I finished work. I believe this range to be the pinnacle of useable RF kit. They came with a gigantic pile of manuals, they were wonderfully reliable and formed the industry standard. However something peculiar seemed to happen at HP and they started a long descent which doesn’t seem ever to end. I bought in wonderful expensive new equipment replacements which just seemed to be the same machine but with a Windows front end doing what Windows does best and being obstructive, obtuse and telling you what you want rather than doing what you want. Running later vector network analysers is frustrating and annoying. Such a shame. I expect RF houses across the world are still thrilled to use these instruments and get fantastic productivity for their dollar/pound/rupee/etc.
Well put! It is exactly as you say it.
I've used the HP8566B and HP8568B at work and here at home and (for me at least) they were the industry standard up until about 2001 when the Agilent 26GHz E4440A PSA was released. The PSA is a fantastic spectrum analyser (even today) and it provided a significant improvement in performance in all areas. The 8568B and 8566B are also fabulous spectrum analysers but they can't really compete with the E4440A PSA. The PSA OS is still HP (not Windows) so I think you missed out this analyser in your analysis. Check out the PSA analyser. It picked up where the 8568B and 8566B left off and the performance is incredible even today.
I worked on RF modules for HP printers a few years back.
One of the engineers was meticulously evaluating each bypass capacitor as to whether it could be removed or not. For cost savings...and I'm thinking well, it was put in the circuit for some good reason at the time, suddenly it's no longer needed?
When you make millions of printers annually, saving 1/10's of a cent per capacitor adds up quickly.
Marc, I enjoyed your comments , restro., in a restro., in a restro. and you didn't really sound surprised. Thank for a great video
I had to smile on this video - as a collector of vintage electronics myself I know the sometimes painful cascade of repairs that is required when you pull out a "known good" instrument of your pile, just to learn that it has developed a fault - and to see that the other instrument you drag in to fix it also has developed a fault... But if you end up with aöl of them back working its an extremely rewarding experience 🙂.
Happy new year Marc!
Nested restorations - my favourite kind!
a rotary copper pipe cutter works pretty good to open up capacitor cans too. I cut a bunch open once when I got bored (bad ones) to see what actually dries out, and to see if I could rejuvenate one of them with some distilled water. it didn't work, but I guess cutting them open that way would be decent for restuffing old ones.
It's why we come here.... to see restorations within restorations within restorations..... ;)
It's like machining channels where they make tools to make tools to make tools. :)
Thank you for this video! I got an 8568B from a retired AT&T guy who said he was using it up until 5 years ago!
Since it went into storage it seems that a lot has gone wrong. The bus interconnect cable got damaged, YTO is unlocked, 249 is unlocked, and I get the battery warning.
I have ordered a new bus cable, but in the meantime I have made a temporary bodge to repair the cable. I have also printed out the 858 page service manual!
I am grateful that this unit can be calibrated by just using itself, I do not have an extensive lab or a infinite number of dollars to send it out. Today I will go to my lab and leave this guy on and see if it is willing to reform it's own capacitors before I start the debug process. Maybe the lithium will be willing to charge too.
I also got a 3478A DMM from the same guy, it seems like it is still calibrated, I will have to order the battery for that too 😵💫
Good luck with that unit. I have another video for the YTO unlock on that same unit, one if my very first videos when the channel switched from my R2-D2 build to what it is now.
*Self-cal? Well, sorta. Before we wised up and put the separated decks in the equipment carts, idiot expediters would unload and carry them by the interconnect cables. Fired a couple for that. Tadrian lithium cell is not rechargeable. If it reads
@@blackrifle6736 Oof, I cannot imagine the negligence required to move equipment like that. The 8568B retailed for an equivalent of $144k in todays dollars! Thanks for the tip, I have since realized that the battery is non-rechargeable and am looking for a replacement. I have also realized that everyone online is lying and providing the service manual for the 8566B. It took my an embarrassing amount of time staring at the manual and the inside of my unit to realize that they have very little in common. Luckily someone on eBay sells a CD-ROM with scans of the correct manuals.
Hopefully the repairs are simple and the references are still good. I would hate to have to buy a frequency counter and a function generator at this moment. My sweet deal wouldn’t be so sweet anymore! :)
Excellent repair, nice to see old HP gear restored to working condition. When checking power supplies with suspected problems I switch the multimeter over to AC as a quick check.
Wonderful to have an instrument with full schematics.
Restoraception/Repairoception is something everyone with vintage equipment usually knows very well ;-)
We apply proper voltage to the battery contacts then slip the old battery out and new one in,,,,,, this way the data is not lost !!! I guess you know that.... Great Channel !!!!! Picked up a lot of good tips !!!
Nice spectrum analyzers. I have been looking for these on and off over the years but were either priced a bit higher than I was willing to pay for, or they were all beat up and look like someone used a paint scraper to remove old cal / asset stickers. So in the interim I eventually settled on a HP 70000 that also needed battery changes. The 70000 series has two batteries, one in the local oscillator and one for the graphics display. As far as the calibration procedure is concerned it completely automatic, just start the procedure and enjoy the show.
Happy New Year Marc! Wonderful video. :) Your basement has almost as much history as any of the NASA facilities had back in the 60's. Huge thanks to you and your entire crew. Well done!
Still one of the best electronics channels on TH-cam!
the amount of engineering that goes into these tools is mind blowing
love these series, not exactly the equipment i used, but so similar to the stuff i used as a trainee engineer back in the early 80's. I find it so fascinating to watch !
Excellent fault analysis and excellent repair !!!
What an absolute beauty of an engineering piece this machine is! I miss that sometimes in modern stuff, though very expensive machines probably still have that... yet missing the possibility to service yourself I guess.
Little brother of the 8566b doomsday machine. Best devices to work with, I love them.
*Ha!, ha! Not many remember Dr Linkwitz's comment. Cheers!*
I'm astonished by the volume of the fan on the h8568b. I wouldn't have wanted to be in a lab with several of them.
Yes, this is the only limitation of this SA (true also for the 22GHz version): the fan noise level. Anyway, back in the days they were a kind of world record instruments,: UNBEATABLE! I used the 22GHz version for my first job, the development of an X band DRO....
*Ha!, ha! So right. Walk into any RF/MW or Cal Lab back in the day and that is all one would hear. After awhile just became easily ignored background white noise. Fun fact: Raytheon still has about 100 B's in daily use. Cheers!*
Love the videos Marc. It's the only place you'd find an impact driver used to remove a bad cap 🤪 but even then somehow I can't stay mad at you. 🤪 Keep up the good work, I for one have learnt heaps following you for a while. 👍
I remember us buying one of these analysers in the late 70's. I seem to recall that the cost was astronomical.
No kidding. The most expensive instrument of the whole catalog, to be exact. About $70k in 1980 dollars for the HP 8566 if I remember correctly, the 8568 was maybe 40k? That's part of why it's so extraordinary inside.
@@CuriousMarc that's what I remember. It did come in some nice cases to take on board our submarines. HP was quality gear superbly engineered.
At the time I was using a network of HP2100 mini computers and FFT boxes to do analysis. We had written our own executive and using HP Assembly language to shoe horn the code into those machines. It was a fascinating time.
We later moved on to use HP21MX machines with FPS AP120B attached PROCESSORS programming in APAL and PPAL for our PIOP processors used to read the data from our ADC's. Men were men in those days none of your C++ nonsence.
@@CuriousMarc *IIRC, 8567 was ~$40k at introduction. 8568B was more. Should not have left my 1979 h/p catalog behind when I retired. Cheers!*
Happy new year Marc. Beautiful repair. Have you noted how IF section of this SA looks like Apollo stuff ? HP’s instruments are unbeatable.
HNY. I love how you dig down to see why that cap is no good. Curious indeed.
That HP is a spaceship in itself, looks amazing with all those shiny internal connectors!
Always great to see the innards of those great HP instruments of the day! Happy New Year, and I'm looking forward to your next video already :)
Re 24:43 in the video where you see the wobble on spans above 1MHz. On spans below 1MHz the 8568B PLL operates all the time and on a typical test signal on a 500kHz span you should see a -112dBc/Hz (typical) phase noise pedestal out to about an 80kHz offset. At wider spans than 1MHz the analyser goes into 'lock and roll' mode where it relocks every sweep but isn't locked during the sweep. To see the difference, put the HP8568B on a 1MHz span and look at the noise close to carrier. Then key in SPAN 1.0001MHz. The span is still effectively just 1MHz but you should seen the close to carrier phase noise is reduced. The analyser is then in lock and roll mode. The sweep won't be as accurate but you can see the close to carrier PLL noise has gone. The swapping between these modes at a 1MHz span probably explains the wobbly cal signal that appeared for a while. It looks like something wasn't right with lock and roll mode until the fault cleared.
Marc, I love watching your videos, for both the electronic and the mechanical repairs. I was astonished, however, when you put your Fluke DVM on that test point, saw the low DC voltage, and then didn't simply switch the Fluke to read ACV. I ALWAYS check both DCV and ACV when checking out PS outputs. It's such a quick and easy test, when your probe is already there.
Regarding the fill of the capacitor, I think what you found was completely normal. The fill is usually pitch (tar) and it's an expense, so manufacturers use no more than necessary. It's only there to keep the wound core from banging around under g-force or vibration. More of it does not make the cap any more reliable. I've cut apart many Al electrolytics, some on a lathe and some with a hacksaw, and yours was typical. In general, you'll find more pitch in older caps, when oil was cheap.
Yes, the tab corroded away, that happens and I don't know the chemistry behind that. It would have been interesting to see if you could make contact with the remnants of that tab and see what the capacitance was. My GUESS is that it would have been low, due to the electrolyte drying out.
Thanks for all your extremely interesting videos, and good luck on your future repairs. I look forward to them.
Very satisfying using those HP spectrum analyzer's. Loved the layout and button feel not to mention the excellent performance. I used the higher frequency version possibly HP 8566B Spectrum Analyzer, 100 Hz to 22 GHz. Interesting to see inside those crusty old electrolytic caps and to see how you installed the new axial lead ones in the place of the old screw terminal ones. Guess it would be difficult to find screw terminal caps now. I've changed batteries in some Agilent gear with much trepidation, hoping to maintain all the calibration data.
Really interesting, thanks, it makes me think that I'll have to start troubleshooting the power supply of my HP8561E. Good luck, Mr. CuriousMarc.
This is helpful. I need to change the battery on my 8568B.
What an ease to replace those screw terminal capacitors! It’s a new year gift Marc!
It was another great episode! Best of success in the new year, CuriousMarc! Someday I hope to see you get the HP 21MX minicomputer running. I took my BASIC class on what we affectionately called "Hewie". Then I worked in IT at the same college, and took care of the Lear Siegler ADM-3a and DEC LA-36 DECWriter printing terminals connected to Hewie. Thanks so much.
The 21MX has been in the queue for so long! It’s getting pushed out by the Apollo stuff…
The 21MX was the first computer I ever used. My high school had one that was not for use by students. A math teacher let me into a room with an ASR 33 that I could use. Eventually I got access to an HP terminal (I don’t remember the model). I look forward to seeing the @CuriousMarc 21MX resurrected!
The best damn SA ever built! I'm glad to see it still in action.
I can't remember the last time that I saw one with such a clear screen!
*Agreed. Most of ours have solid state replacement displays from various vendors. Problem is the H/P 85685 Preselector not always happy with the change. Cheers!*
@@blackrifle6736 Oh, yes! We couldn't have done our job without the 85685. Amazing hardware. And, they were more sturdy than they looked! One of our team members dropped one at the top of the stairs to a 777 once. It rolled end over end all the way down to the tarmac. After replacing a couple of bent connectors and a re-cal it was as good as new 😎
Another awesome repair job. Happy new year from Australia 🥂
Thank you Marc for all of the great content this past year. Happy New Year to you , Ken and Mike!
I found three of those exact same (yellow BR 2/3A) batteries in my Tektronix DSA602A. It seems those were a favorite of some designers while Dallas chips were used by others. That said, I’ve seen how the Tek 2440 had a revision to remove the battery and replace the sram with a Dallas chip.
Gotta be smart to use these fixing them is next level....(I used a 8565b for several years for microwave radio testing)
Always love your repair videos. Thanks for keeping us entertained!
Some capacitor manufacturers also sell brackets. They are designed to mount the condenser with the screws facing up. Now if the body of the condenser is a tank, why do we install them upside down? Could it be that they were originally designed to be face up? It might also be good practice to store some equipment that way and prevent damage to circuit boards from corrosion.
You can tell you love doing the calibration process, because removing the capacitor with an impact wrench will make the whole analyzer shake. :) Best Regards!!
What a beautiful piece of machinery! Thanks for the video!
Restoration within a restoration within a restoration, it's restorations all the way down.
I recaped a old 141t Spectrum Analyzer the cap improved the focus by a lot the cap do a lot with the display.
Happy new year 🥳 love your Chanel
Fantastic job =D Love watching everything you do!
Cool projects as always! Some viewer insight - recaps are always appreciated. Though maybe I'm the outlier because I'll often take hiatus from channels. Also, it's mostly tech babble that goes right over my head, so thanks for presenting it in an accessible way. Cheers
8566B FTW! 22GHz, baby!
All this is way above my education but such a great job explaining for just about anyone to understand and very interesting.
Thinking about the RF gain of this antenna. A few kW amplifier will give a tremendous ERP (several Mega watts)
May be enough to disturb plane communication in the direction of the beam... Or to cook pigeons 😂
Man, I can't find lots of common electrolytics for my builds. It's amazing. Basic power supply caps and mouser is down to 12 or 50, when they would normally stock 3000 or more.
since watching this i realised i need a LCR meter to make my testing and repairs ever so slightly easier.....slightly
It's restoration Inception! That's the downside of vintage test equipment unfortunately.
Happy New Year.
I have the 8566B and so far has been really reliable since the YTO unlock problem like yours. Tend to power it up on a regular basis which seems to help. I did replace the main electrolytic's as a precaution although they tested like new. I love these machines, I don't understand half of how they work, but manage to keep mine going. Happy New Year all.
Why replace caps that test good?
@@ToTheGAMES Because they fail short when 40 years old.
@@ToTheGAMES - Caps are rated for maximum operating hours as well. Given the rarity of this gear, its best to err on the side of caution and recap the power supplies. Especially when the gear has the dreaded Rifa caps in them.
I don't think of this action as being the typical "foaming at the mouth" replace all da caps! knee-jerk reaction... it's just preventative maintenance on a piece of precision test gear.
Cheers,
@@EngineeringVignettes Absolutely. I am not a recaper for the sake of it person. As you say these units are so complex I would want to replace known unreliable parts inside them. The caps despite being cooled by the fan, have a hard life with high ripple currents and long periods without use.
@@SoddingaboutSi They almost never fail shorted. It happens, granted, but most often they go high-ESR like this one did. Also, because no one uses large "computer grade" electrolytics in production these days, any new ones you buy have probably been on the shelf for years (like his battery was.) It could be worthwhile to reform them before installation, when you do replace them. IMO given the high prices it's not worthwhile to replace these large electros proactively.
A restoration within a restoration, recursion rules the day...
nicely done Marc!
I had *exactly* this issue with mine. Same filter cap
Wow, never replaced a cap so big it used screws... Nice bonus machining content too!
*Oh, you should take a look inside an h/p linear power supply. Also computer-grade, we call them Beer Can caps. Cheers!*
Amazing machines thanks a lot!
Happy new year btw :D
Watching DiodeGoneWild fix a vintage TV recently... he was "concealing" modern small caps inside the busted cases of the old ones.... the finished result is "almost authentic".
Its restorations all the way down!
Vote up, thanks for sharing the video :)
Ahh HP don't make them like they used to. Used one of these for many a year and it's higher frequency brother. I've still not seen an 8902A measuring receiver anywhere in the backgroudn yet, those were my favourite bits of kit, haha!
*After he does, he should find the h/p laser interferometer and waveguide below cutoff attenuator calibration set for the 8902A. Not unicorns, they do exist. Cheers!*
This series is like game. Fighting to get the box running with the occasional side quest :)
Restoration Inception
Happy New Year Marc 🎉🎊🥂
Hi Marc, Happy New Year! I have an 8566B too, a couple years ago I took it apart and measured the Tadiran battery that had a date of 1996 and it was still in spec. I checked Digikey and they are still available to buy! I didn't bother replacing it.
Ah! Another repair video that becomes a repair tangent (I watch because of tangents btw)...
restoration within a restoration within a restoratSTACK OVERFLOW
Super expert!
Is there any reason you are SO afraid of that HV power supply? It's nothing to be scared of, it's current regulated so it's pretty safe even in the case of a direct discharge through your body. I tried it just to see what it feels like, I had an old CRT TV, right foot on a grounded piece of metal and right hand touching the wire. Yeah it makes you jump a bit but nothing more than that. Mains voltage is a thousand times nastier.
And one thing I'm very curious about: what would you choose between these old ones and the modern brand new ones, in terms of reliability, precision and ease to repair? I am mesmerized by how the internals of these old machines look like.
@@AngelaTheSephira I'm sorry that I don't have any CRT TV around any more, I'd show you right now.
They most definitely have some kind of protection, because the moment the discharge reaches a high enough current it just shuts off and starts pulsing like 1 time a second.
And it's not the first HV supply I "tested" on myself. I built a few ozone generators for all this virus thing to sterilize stuff like groceries, keys, money etc. and got zapped by those as well, one of them from left hand to right hand. Funny how I felt my tongue being electrocuted.
Still, the times I got electrocuted at mains voltage were the absolute worst.
The thumbnail shows the good and the bad. But, where's the ugly?
I have a bottle of HP sauce in the fridge somewhere.
When you have to fix the test equipment used to fix the test equipment for fixing the... OMG! :)
how good was the past with complete diagrams in every big electronic device?? so u can repair it easy....good for you and for the planet! Today they thrown everything in the junkyard with a tiny small bad component!
sadly!
Amazingly clever, you are!
It is good idea to clean the CRTs on that ols HP by removing the front glass. They getting very dusty over the yesrs. The screan really come a life.
Why using the 8568B and not simply staying with the 8566B ? I was looking at the specs and it look like the 8566B is not that bad at low frequency.
A cap in my PC power supply also failed in exact same way. It just went open and upon opening it, there was also one contact eaten away
Vidéo passionnante encore une fois 😁👏👏👏
15:39 Clickspring!
On an electrolytic Cap, the second contact goes to the Aluminium enclosure which acts as the second "plate" of the capacitor together with the electrolyte. are you sure this is Epoxy at the bottom ?
Bonne année 2022
At 20:19 what if you connect another battery while changing the current one in order to avoid loosing the calibration information? Thank you for the video Marc!
So they used the European date format for the date of the firmware.
I’d contend that if you screw up repairing an 18kv power supply, you may end up being less hairy very quickly.
A 1.5 GHz restoration inside trying to rebuild a 22 GHz system? That's like 23.5 billion restorations in one!
At the start of this video you show pictures of the s-band maintenance bench. Where did you find this. Would love to get a better show of those benches with equipment.