EEVblog
Вставка
- Опубліковано 5 жов 2024
- Dave repairs a classic HP 35660A Dynamic Signal Analyser (DSA), and does a bit of a teardown with random commentary in the process.
Service Manual: cp.literature.a...
Forum: www.eevblog.com...
EEVblog Main Web Site:
www.eevblog.com
EEVblog Amazon Store:
astore.amazon.c...
Donations:
www.eevblog.com...
Projects:
www.eevblog.com...
Electronics Info Wiki:
www.eevblog.com...
I'll cut to the chase and tell the problem straight . Apologies if sounds like arrogant.
The deflection yoke came loose and pulled back just push it forward and center the position of the picture. The dark areas on the edges caused by the pulled back yoke too.
The magnets are just for geometry correction on the edges/corners.
Push the field scan coils forward, thats the corner problem.
Its been dropped on its end and the momentum has pulled the scan coils away from the crt, same with the pcb.
Smashing device mate, great score :-))
Yep I place my Bet on it Dropped on it Bum..Well Most Tech have Wimpy Arms, One of the Reasons My old Du Mont 224A is on one Spot Only it 50lbs, the old 585 Tektronix are also a Killer on the Back..
I agree. It's likely that it could have been placed on a cushioned surface, on it's back and moved from pinot a to point b, I.e. a cart, or a car, and was josseled repeatedly. Hence no damage to the housing? ;)
I agree
Watching these videos really brings me back to my childhood. My parents practically had to hide any technology they didn't want violated by my screwdrivers. I'd taken everything apart from TV's, VCR's, old computers, radios, anything with a PCB in it. Perhaps unfortunately, I changed course to software rather than hardware as I got into my later teens, a decision I sometimes regret as I pine for the good old days of my youth, ripping things apart and fixing them. Thanks for these videos.
Dave. The magnets are for linearity adjustments of the image provided by the yolk. At assembly the calibration tech will call up a grid pattern and then slap calibration template on the screen. He will then add and subtract magnets or reposition them in order to make the projected image conform to the template. It's a very fine adjustment so you will not see anything with the naked eye without the help of the grid and template.
You'd be surprised. I've had these DSA's on mobile trolleys that gets wheeled around factories all day long for years. The amount of vibration they get in that environment can be incredible. Free-standing TO-220's don't last a month before they shear off.
in order to get the full image on the screen push the coils on the neck, the same coils you touched to correct the rotation, forward a bit more, it looks like they slipped back.
I think the instrument was dropped inside a shipping box and got shocked but not dented.. good luck.
ciao Dave!
I'm amazed that you almost recovered this old beaty beast, on which I worked for years. The only drawback of theis machine is that the pushbutton keys or "software keys" have labels not so easily "intuitive"... I mean the "friendliness of a TEK scope is beyond any question...
Hi Dave this unit has been dropped whilst in packing, the ring with the magnets should be nearest to the front around the circumference of the scan coil these are pin cushion correction magnets. The picture being cut off in the corners is due to the scan coil having moved away from the bowl of the crt so the beam is hitting the neck at the point of gratest deflection i.e the corners.
I worked in a monitor manufacturing company in the UK for 13 years and set up more screens than I can remember
When you got the lid off of that CRT, it was clear that it was'nt good old HP quality - and it was clear that the CRT innards was made for bench use only..
Thanks for all the educational videos mate!
Cheers from Denmark!
The corners are because the yoke on the CRT is too far back on the neck.
I thought it might have something to do with the shadow mask but then I remembered these monochromes don't use them :/
Yea, monochrome CRT's have no shadow mask. However, oscilloscope tubes have some kind of screen inside, but I forget what it is called.
I agree, the yoke could do with pushing forward, it probably turned when it slid back explaining the rotation you observed.
Exactly what I thought having spent the 1980's servicing TV's.
Hi Dave. In my college years I trained on Black and White TV's. Thos. e permanent magnets were used to prevent 'Pin Cushion' effect
My trade. A long time ago. Scan coils need to be pushed forward "snug" to the tube.
Agree with hjrmartins.
you need to slide the yoke in or out to fix the corner issue
The dark corners are caused by the yolk coil being moved back. The dark areas are caused by a shadow created where the neck flares out to the bell. Slide the yolk forward toward the CRT face. Whatever knocked the CRT PCB off also dislodged the yolk coil. Again super easy fix.
My guess the unit either had the CRT unit subbed out in order to make 1 working unit out of 2 malfunctioning units. Or it was in some kind of hard case that protected the chassis and it was dropped to land on the back end. You know the type of case, fiberglass outer shell with custom cut foam interior and heavy duty locking latches. Not inconceivable for something that would be more useful out in the field while looking at physical structures like buildings.
The 4 magnets on the gray ring are for linearity adjustments. Just make sure the ring is back exactly where it was from the factory.
As far as my experience with CRT-based TV sets goes, these caps hold their charge only for a few hours, but it depends on the circuitr around them. When I was a teenager and I disassembled my first TV set, the first thing I did after switching it off was to pull the HV plug from the CRT and I touched it. Boy, my hair stood vertical!
I love these kinds of non smd repair and tear down videos! From a different golden age of electronics.
Great vid, nice to see you working with one of my favorite ranges of old HP equipment. I work in a sonar/acoustics lab and we have the 3562a, a whole pile of 35665a's, and a 35670a. They get more use than almost all of our other test equipment. The swept sine feature is great- one can measure impedance and frequency response easily and with zero hassle. I don't know of any PC based system that is as simple and quick as these boxes. I was sad to see that Keysight just disco'd the 35670a.
There are pc ones that are just as easy to use, and of course better specs and data export/reporting. data physics quattro is probably the best one, just need that and a tablet or laptop
Dave the reason I bought one of these was for loop analysis. Since most control loops like in power supplies have bandwidths within the range of these instruments they are great for determining loop stability. All the other stuff like frequency analysis, correlation, coherence etc. can be done on a PC with long record lengths for detailed analysis. But, for loop analysis these are the bomb.
The 35660a was my first project as a new firmware engineer at HP. It's cool to see these are still in use. I built in an easter egg. As I remember it, you have to set the X offset marker to -35660, then set the X marker to 35660.
I've always wondered how difficult would be to design an LCD upgrade kit that would be able to plug straight in instead of the bulky CRT. It would probably require unreasonable amount of computing power etc but wouldn't that be great
not really. Most probably the signals going in are VGA-ish. So all you need is a VGA-to-Digital converter.
i love these repair videos. its like going on a small adventure
I don't know about the actual randomness in terms of cryptography uses, but IIRC it is flat across the band to within a few dB, as you saw in the video.
Great repair video Dave! Yeah that thing will make your shelves sag for sure :D The collar of permanent magents might be set for your particular geo-magnetic environment. A Uni lecturer suggested that these need to be different for different areas of the world, but I'M NO EXPERT. Glad you got it working well.
The magnets on the side of the deflection coil are used to adjust the geometry of the screen and the ring shaped ones on the neck of the tube adjust the horizontal/vertical beam centering. Also, pushing the deflection coil assembly towards the front of the screen should help get rid of those dead spots.
The "filter" noted covering the screen is actually an EMI blocking mesh. It is a grounded mesh to shield the CRT from noise just as the rest of the metal case is used to shield the components. It is basically a screen for a window, but in this case it is keeping out EMI instead of bugs.
The yoke does not only rotate the image but by moving the yoke either forward or back will restore the full image.
And it's important to note that power supply and display circuits themselves (for both CRT and LCD) will typically emit relative large fields that will easily corrupt the low-level analog signals. Without shield, the instrument would essentially jam itself and be completely inoperable.
Repairs are my favorite kind of video as I seem to learn the most. Just my two cents.
Another great video! Nothing better than peaking inside older test equipment!
It depends on how the video is being fed to the CRT. Some units will use a TTL signal to drive the CRT, which would require extensive hardware decoding to drive an LCD display.
Just wanted to say repair videos are my favorite so thanks for making this video!!
It's fun to see the reflection of you waving your hand around during your explanation in the into :)
Great video - as usual. It would be interesting to combine the 35660A with a tutorial on active filters at frequencies less than 100kHz. Using it to display a bode plot of the filter response. It would be a good demo for those of us still learning. ;-)
I thought it was tight on initial check, but yes, perhaps there is more to it, I need to investigate further.
HP used to build absolutely beautiful equipment both internally and externally. I miss those days. Today, everything seems to be just a few chips and a PC. I have a hardcover HP catalog from 1987 with over 700 pages of wonderful equipment from that era. I used to drool over the gear back then. Even HP's catalog is top quality. The closest DSA I could find to yours is the Model 3562A which was probably a few versions before yours but looks vaguely similar.
The TMS320 DSP actually works at 40MHz in this thing.
I agree the scan coils have moved backwards due to a drop and the electron beam is hitting the neck in the corners instead of the screen. I'd get the screening off covering the coils. There are usually a couple of rubber wedges stuck to the CRT which the scan coils are pushed up against. There should be some witness marks to help you out.
thing was operated in vertical position and with vibrations , the deflection yoke as moved and is to close to rear of tube , push back and you will reduce opening and hence have good full screen view , magnets where there to compensate some magnetic field due to shielding ( like boat compass compensation ) , anyways it shows are reliable and sturdy is HP old equipment , great video ...
Hi Dave, to fix the horizontal issue with the CRT, you need to push the deflection coil IN (away from the neck board). You'll probably need to ajust to focus after that, but that should cure the issue no problem at all.
Hi Dave, I think the problem of the shadows in the corners is because the deflection coil has moved a little to the back. Thanks for your videos, are very interesting!
Pretty neat find. I think these instruments costed up to 70.000 $ when they originally came out.
I think somebody tried to pulled out the CRT just as Dave "WAS" going to do by taking out the 4 CRT corner screws and in the process dislodged the magnets, yoke and the neck board. That is why you don't see any external damage. When they gave up they pushed the CRT back into place leaving it in the state Dave found.
they're called "pincushioning" magnets...used to square up the grid trace lines.
The fringing is the beam hitting the neck or sides of the tube, try moving the deflection coils back and forward and see if it fixes the dead spots
@M0OFC, That is my guess too. Just as Dave 'WAS" going to take the CRT corner screws out and pull out the CRT. I think somebody did just that and in the process pulled out those magnets and the neck board. When they couldn't get it out that way they pushed it all back in leaving it in the state Dave found it.
Dave, the deflection yolk has almost certainly slid back towards the neck board on the shaft of the CRT. Loosen the screw you did for the rotation and push the yolk back towards the front of the screen and your roused corners will be fixed.
I've seen several comments wondering why you'd want one of these, or asking for practical examples. You can still find ap notes for various DSA measurements on the Agilent website. AN243, AN243-1, and the user manuals for the HP35660A, HP35665A, HP35670A, HP3561A and HP3562A all have practical applications info, I believe.
That was genius. But it is always brilliant to to see you work your engineering magic.
Hi Angus... DSAs have very large dynamic range, typically 80 to 100 dB. Meaning that they work with very small signals up to normal volt range signals. Because the small signals can be in the range of micro-volts, the input analog interface / signal processing must be extremely clean and EMI shielded to avoid corrupting micro-volt level signals.
Exactly the same childhood scenario here. I constantly got into trouble! Fantastic!
Thought you'd appreciate knowing, Dave, that your UA-cam video has spawned a lot of great recollections and story-swapping among the people who were associated with the HP35660 and spin-off projects. Thanks for tickling our memories. -- Though the digitization has gotten a lot easier than it was back then, the care and thoroughness put into the software/firmware as well as the effort to insure that spurious signals don't pollute the measurements, is still very significant value-added.
Yes, of course, it's the logical conclusion. I need to take it apart further and give it a better shove.
yes please Dave, would be nice to see it in use..
Maybe get ur Analog mate in.. from Rhode
Doug Ford?
This is my new favorite eevblog episode. Fixing CRTs for the win.
I don't recall the advantages of the expanded memory. Probably not worth the effort. Could likely still get the IC's as old stock somewhere.
Latest vid from Andy was only a week or two ago. He's working on the most enormous light source I've ever sen. If he gets it lit, it should be visible across the world!
Try moving the deflection yoke on the neck of the tube. A gentle nudge forward should fix the dark corners. ROG
(Continuing...) Something you missed about the plastic piece over the front of the CRT: it's metalized, for EMI shielding. Realized that the CRT scan frequencies, both V and H, are right in the bandwidth of the instrument. -- Also although there are much faster processors that easily could do the whole job for an audio+ DSA, the ADCs have moved on too, so now you'd have to deal with all those complicated measurements (zoom and decimation, modulation analysis, on and on) in real-time at 100MHz
Repair videos are my favorite :D thankyou for uploading!!
Sometimes the copper rails to the L/R of the logic and memory boards are power rails. One would be ground and the other 5V most likely. Depends on if one is insulated from the chassis and has a power cable going to it from the supply. In this case it appears power comes from the ribbon cable (multiple pins per voltage to carry the current).
desoldering are for beam bending similar to convergence, just using magnetism for one gun, deflection shaping and bending.
Might check to see if the yoke (defelection coils) are displaced towards the rear of the tube causing the beam to deflect into the 'neck' of the tube (bending the beam too close to the gun). Being how everything else seems to have moved towards the rear of the tube in that module (magnets and gun board), it may have slid the yoke back as well.
I didn't read all the comments, so pardon me if this has been mentioned, but the deflection coil assy could have shifted as well as rotated. Your edge drop-off problem may be solved by moving the deflection coil assy forward or backward. (sorry, it's been a while since my days in the TV shop working with CRTs)
I've actually taken a look at Agilent 35670A DSA's, both new and used and the price difference can be astronomical (new@ 22,749 USD vs. used@ ~5,000 USD), yet the equipment appears to be the same spec. I can see why used equipment like this can be a nice pickup (if it's not BER).
Tom Bruhns wrote :" The whole image (grid plus data trace) is generated on the processor board,"
In this particular case that is true. But if you rewatch the you will take note that Dave mentions the approximate vintage of the unit is 1988. The unit is a hybrid and can generate its grid pattern electronic. At that time all CRT analog CRT test equipment had grids built into their screens. Including my trusty 2465B Oscilloscope. For that reason image bowing had to be reduced to a minimum
Skew of the image on the CRT won't affect accuracy of the displayed signal. The whole image (grid plus data trace) is generated on the processor board, so there's no way the CRT will cause data to be mis-aligned with the grid.
I'm really happy you got it working mate. Fantastic,
Hi Dave, it looks like you need to check that the deflection yoke is all the way forward against the bell of the crt, should get rid of the shadowing and get a full screen display, cheers, Raff.
I'd love it if your tear downs,if you talk about what your guess is how many people it took to design it, and what a product like it would take to get made.
I was given 3 'broken' air conditioners that were originally used on a yacht, and turned out, on all three of them, the main capacitor(dual capacitor) had come unplugged. It was due to a bad design, the negative wire to mains connected directly to the capacitor, so tugging on the wire would cause it to come unplugged.
The using this with an accelerometer sounds really cool, would like to see a video of that!!
Dave, maybe you could do an episode on what safety measures to take when trying to repair an CRT-display?
A practical example of where you'd use one of these guys would be great!
that mc68000p is running at 10 mhz not 8 as you thought :)
Also Fun chip you missed is the MC68440 - Dual-Channel Direct Memory Access Controller. I saw that and had to look it up. Im designing my own mc68K system from scratch so finding this chip will help me in my design
I don't know if anyone has mentioned this but the vignetting is caused by the deflection yoke being too far from the CRT bell or you need to adjust the static centering with the tabs on the back of the yoke.
He did a whole episode on vignetting after this episode.
Dave, use the force, push the scan coil assembly forward until it meets the back of the CRT, the faults indicate the unit has fallen hard on it's back !
That is one huge unit, bigger than my Anritsu 37225A, but maybe lighter in weight ?, still looking for a 50 pin SCSI 2.5" hard drive for my unit !,
There is TMS320 DSP chip working as the math co-processor, working at 40MHz.
Ah, that explains it, thanks for the explanation. So more gilding the lilly?
A Tone of TMS320 Goodness!! LOL Luv it and cannot argue. On YaY Dave!!!
Some days you get lucky! Wish I could find an RF Spectrum Analyzer with that sort of problem.
Lily Gilding ? That all depends on what your requirements are and what the spec calls for. A -69db signal might read -69db at the left and right edges of the screen but read -70db in the middle. For most applications that might be fine. But then others might demand the RCH be Split clear across the screen. It's all about what you requirements. The magnets provide an effective, although not perfect, way of cleaning up the image.
I second the motion of zx8401ztv. If it were dropped on its top or bottom it's possible that the CRT neck could've been broken. That would surely been an end game for the display unit. I believe the magnet assy is for geometry distortion adjustments. They are trying to make the grid square looking.
Dave you've never worked on CRT TVs. You'd have recognized the yoke had slid back on the neck of the CRT.
I know it's been a long time since this video was aired. If you haven't fixed the display yet on this fine find then just like I said above, loosen the yoke, slide it forward on the neck then make adjustments like width, horz. lin., V size V lin etc.
Hope this helps after all this time.
Paul
24:09 when it came on I had a big grin my face. Awesome find.
Push the deflection coils up against the bell of the CRT; that'll cure the shadows. The magnets are to adjust the shape of the picture; those coils aren't perfect and a little permanent magnet help gets the image square. The neck board fell off with the tube socket; pop it out of the socket on that board and glue it back onto the CRT. All of this should get that old CRT monitor working like it's supposed to.
Yes - scan coils look like they're not sitting all the way forward - looks like this was dropped on its arse.
I think that the dark corners are because the deflection array has moved back on the tube. Move it towards the front of the CRT ant it should fix it.
Interesting to note all the right angle traces on the main board.
"quite some G's"
Yeap. I've seen that phenomena in rack drawers. Some guys with too much energy push the equip hard into the rack. Sometimes the 'plastics' of the yoke are close to the PCB socket and push it away. The yoke is relatively heavy. I have also seen flyback transformers knocked out of the PCB!
Matters are more serious in color CRTs. Specially in triniton tubes, the 'stripes' glue in bunches and we have to knock the CRT until they 'unglue'.
This video is no yoke!
am i a yoke to you?
Yeah, that's the thing. No sign of physical trauma on the unit at all, and to do such a thing would require some quite some G's which would usually damage the case in some way. But in any case, Shock's Your Uncle.
you could push the tracing coils bit more further, and that should get those dark spots fixed
You should be able to get the screen corners back in by pulling the coils back a bit on the neck of the tube.
Ta
Tony
Those magnets are supposed to correct edge focus if the crt does not use dynamic focus.
Check if the deflection yoke is tight against CRT. If it is a little bit away (towards the CRT pins) the beam will be curved too son and hit the CRT glass. Then you get 'corners'.
As is, it's working well enough it'd keep me happy. Still, another mod that probably wouldn't be _too_ hard would be to add an optional external monitor jack. I'd assume that ribbon cable just has the simple analogue signals coming over it, yes?
That's my guess also.- I played with magnets on a CRT colour tv when i was a kid,- it gives a nice rainbow effect.. The magnets bend the electrons path.
dont worry, with money and time you can still do hardware, i did. i'm studying CS but also learn electronics on my own thanks to the internet, i even bought an oscilloscope just for learning (it's a must). then it's just a matter of reading, and trying, just like when you program, but instead of a bytes, you need to buy the components and equipment (scope, power suply, multimeter at least and then tranasistors, resistors, caps, etc, a few of each would do)
How somebody can be this lucky guy :)))) Wooohoooo!!! Have nice usage Dave :)
And I repaired old full VALVE TV sets aswell. And don't remember that something like "Focus coil" existed.
The blanking in the corners - I will bet that the source of this problem is related to why the magnets and neck board were displaced. The magnets help with aligning the beam onto the front phosphorous (sp?) of the CRT. If the magnets are displaced, then the beam is too wide and part of it impacting the side of the CRT rather than the front. My $0.02.