We used to make stuff like that. ( I have that same Wavetek and have used it since the 1970's. Bulletproof. ) It's like sausage... you really don't want to know how it's made!
Every nano Henry and pico Farad counts with a complex design like this. It probably was hand built a few at a time and took at least a day factory tech time to calibrate. 60+yo technology! ONE hundred dollar FPGA with some extras (step attenuator etc) would do the job today...
It is odd that in electronics the expectation of low volume production products be the same as high volume. That is to say, it seems to me to be unlike any other subject I'm familiar with. For example, in mechanics a low volume product would be mostly machined, while a high volume product might include stampings and maybe castings which are only viable with large product counts. In electronics it seems the lack of these high volume necessities is seen as a detriment, though I don't really see how.
A fella from Indy told me that the modules were flying all over the place in the repair center. So having all the same rev and matching diagrams was pretty much impossible in a given generator. To boot, some modules don't like others. So, simultaneously, the engineers were working out problems, the purchasing agents were shopping and changing parts, and the repair centers were trying to get something out the door. All, with Uncle Sam screaming for contract deliverables. Well, anyway, it makes a GREAT repair video saga! 73's DE W3IHM
Nice job of troubleshooting. I admire your persistence. (I have a nice Tek 2754 that probably needs 1000 tantalums replaced and it looks like the Space Shuttle inside. 21 GHz lab spectrum analyzer with ungodly dynamic range. I've been postponing doing what you are doing for 15 years. When it works, it's amazing. Was over $75,000 new. A work of art. )
There is about a week of 'wire girl' time and a day of tech time to manufacture something like this back in the late 70s. $6.50/hr and $12/hr was premium wage back then for the best... BOM parts was about the same cost = 1/3 each plus brick & mortar expenses to wholesale. Add another 40% on top for initial engineering, profits and sales/advertising etc. Even the Chinese couldn't do it today...
Another very busy can. So many levels and angles. Get a sponsored GHz scope? Vintage Tektronix are great at 1Ghz but are there any non-Windoze Tek scopes at >2G? Maybe the tds794d? You don't have the space😂
@@jrkorman AD797 or OPA1612 are 10dB quieter, lower bias, lower offsets etc. than 4558's (just have to watch the supply doesn't exceed +-17v). ACmos for the 40MHz comb generator would get 6dB more output at 120MHz. Lower power dissipation (current loops, too) would cool the machine down significantly - probably running at 50c inside the case - and decrease the spurious freqs ( -30dBc ain't so great). Probably could reduce phase noise by half on the UHF varactors. Gotta know what your doin... and check everything before/after. 👍 edit: They are using upper/lower sideband mixes for various freq bands - snapshot FFT at extremes with 350MHz Rigol scope to verify spurs. He has the tools! You could probably replace most of the discrete RF amplifiers with Minicircuit's microdot X packages and drop supply currents quite a bit... VHF/UHF analog artistry and a little experience 👌
@@jrkorman That is a rather unique machine, it can do 500KHz peak deviation - well worth modifying and upgrading ! VHF goldmine for hams... or older data streaming repairs etc.
Hi, I saw your video and I fell in love with it. On Sunday I bought a trunk with thousands of components at a flea market, I understood that there was something good in the middle. in fact I managed to find different versions of the 8041, 8085, 8088, 8086, ram, rom, eeprom, plus hundreds of TTLs as well as thousands of transistors, capacitors, .... etc. I wanted to know if you could help me choose the best versions of components and build a computer from scratch. It has always been a passion of mine. I hope you can help me, thanks...
The "no-see-um" caps look like tiny early polystyrene caps but I didn't have 4K turned on to see detail. These module were not built for service ability.They have the air of "git-r-done'. Don't beat yourself up about the mixing. TBH who would think that a device of that vintage would be using the low side of a mixer to get to 500mhz? I'd be in that boat with you. See if you can find a 7000 series scope from Tek with a sampling head.
I was thinking about how to do modulation: AM can be done with an unbalanced diode mixer ring. FM over that range you would have to have a fixed LO that you could FM and mix down to the desired frequency. Just my thought.
Lot's of companies used "high side" mixdown to reduce in band spurs from a double balanced mixer. 3X higher than max desired freq = NO inband spurs (walking down in FFT or spectrum). Plus it was much easier to build a 1.3 : 1 ratio than octave wide oscillators. The limit was UHF bipolar cheap transistors and varactors. 5GHz Ft's and FR4 board. A generation earlier was much more complicated mix divide or double mix divide to make fine freq resolution synthesizers with YIG oscillators (hard to pull those).
@@jim9930 The point was a device prior to the late 80s using GHZ generators to get to a lower frequency would have rare and so not worthy of beating yourself up for.
@@mnoxman I briefly looked at the owners manual; the guys who designed it struggled to make that work. Anybody outside the narrow field of broadband RF test equipment would be lost... He is a physics major with specialty in optics - fantastic work to catch on to old tricks and dig in to repair that beast ! Back in the 80s, there were only a few hundred engineers capable of building it. And the manual is hard to read... my hats off. Fixing stuff is the best way to learn.
... a little story for ya? We hired a technician to calibrate and test broadband RF test equipment in the 80's. John was the lead tech responsible for the telemetry link from 2ncd stage on the Saturn V's. He told stories of launch day - "Lock , baby, lock" was their montra because of microphonics in the PLL synthesizers. (not only caps, board flexing, triboelectric effects and more) 🤣
That module looks more like a handmade hobbyist project than the innards of a professional scientific instrument.
Yeah, like he said, whoever was picking the cap's was not thinking about production issues.
We used to make stuff like that. ( I have that same Wavetek and have used it since the 1970's. Bulletproof. ) It's like sausage... you really don't want to know how it's made!
Every nano Henry and pico Farad counts with a complex design like this. It probably was hand built a few at a time and took at least a day factory tech time to calibrate. 60+yo technology!
ONE hundred dollar FPGA with some extras (step attenuator etc) would do the job today...
It is odd that in electronics the expectation of low volume production products be the same as high volume. That is to say, it seems to me to be unlike any other subject I'm familiar with. For example, in mechanics a low volume product would be mostly machined, while a high volume product might include stampings and maybe castings which are only viable with large product counts. In electronics it seems the lack of these high volume necessities is seen as a detriment, though I don't really see how.
A fella from Indy told me that the modules were flying all over the place in the repair center. So having all the same rev and matching diagrams was pretty much impossible in a given generator. To boot, some modules don't like others. So, simultaneously, the engineers were working out problems, the purchasing agents were shopping and changing parts, and the repair centers were trying to get something out the door. All, with Uncle Sam screaming for contract deliverables. Well, anyway, it makes a GREAT repair video saga! 73's DE W3IHM
Good times
Nice job of troubleshooting. I admire your persistence. (I have a nice Tek 2754 that probably needs 1000 tantalums replaced and it looks like the Space Shuttle inside. 21 GHz lab spectrum analyzer with ungodly dynamic range. I've been postponing doing what you are doing for 15 years. When it works, it's amazing. Was over $75,000 new. A work of art. )
An epic repair video 🙂👍
When you know it it sounds very reasonable, covering a wide range of frequencies on a higher band and then mixing it down to where you need it.
imagine the time it took to build one of these things,... no doubt very expensive.
There is about a week of 'wire girl' time and a day of tech time to manufacture something like this back in the late 70s. $6.50/hr and $12/hr was premium wage back then for the best... BOM parts was about the same cost = 1/3 each plus brick & mortar expenses to wholesale. Add another 40% on top for initial engineering, profits and sales/advertising etc. Even the Chinese couldn't do it today...
Another very busy can.
So many levels and angles.
Get a sponsored GHz scope? Vintage Tektronix are great at 1Ghz but are there any non-Windoze Tek scopes at >2G? Maybe the tds794d? You don't have the space😂
While you have the modules apart - maybe replace the TTL with HCT cmos equivalents, and the op amps with lower noise modern equivalents.
One would want to tread carefully! Straight replacement of old TTL with HCT could cause all sorts of problems.
@@jrkorman AD797 or OPA1612 are 10dB quieter, lower bias, lower offsets etc. than 4558's (just have to watch the supply doesn't exceed +-17v). ACmos for the 40MHz comb generator would get 6dB more output at 120MHz. Lower power dissipation (current loops, too) would cool the machine down significantly - probably running at 50c inside the case - and decrease the spurious freqs ( -30dBc ain't so great). Probably could reduce phase noise by half on the UHF varactors.
Gotta know what your doin... and check everything before/after. 👍
edit: They are using upper/lower sideband mixes for various freq bands - snapshot FFT at extremes with 350MHz Rigol scope to verify spurs. He has the tools! You could probably replace most of the discrete RF amplifiers with Minicircuit's microdot X packages and drop supply currents quite a bit... VHF/UHF analog artistry and a little experience 👌
@@jrkorman That is a rather unique machine, it can do 500KHz peak deviation - well worth modifying and upgrading !
VHF goldmine for hams... or older data streaming repairs etc.
Hi, I saw your video and I fell in love with it. On Sunday I bought a trunk with thousands of components at a flea market, I understood that there was something good in the middle. in fact I managed to find different versions of the 8041, 8085, 8088, 8086, ram, rom, eeprom, plus hundreds of TTLs as well as thousands of transistors, capacitors, .... etc. I wanted to know if you could help me choose the best versions of components and build a computer from scratch. It has always been a passion of mine. I hope you can help me, thanks...
the 8085 was my favorite. do a search on my channel for 8085 and you will find many projects
The "no-see-um" caps look like tiny early polystyrene caps but I didn't have 4K turned on to see detail. These module were not built for service ability.They have the air of "git-r-done'. Don't beat yourself up about the mixing. TBH who would think that a device of that vintage would be using the low side of a mixer to get to 500mhz? I'd be in that boat with you. See if you can find a 7000 series scope from Tek with a sampling head.
I was thinking about how to do modulation: AM can be done with an unbalanced diode mixer ring. FM over that range you would have to have a fixed LO that you could FM and mix down to the desired frequency. Just my thought.
Lot's of companies used "high side" mixdown to reduce in band spurs from a double balanced mixer. 3X higher than max desired freq = NO inband spurs (walking down in FFT or spectrum). Plus it was much easier to build a 1.3 : 1 ratio than octave wide oscillators. The limit was UHF bipolar cheap transistors and varactors. 5GHz Ft's and FR4 board. A generation earlier was much more complicated mix divide or double mix divide to make fine freq resolution synthesizers with YIG oscillators (hard to pull those).
@@jim9930 The point was a device prior to the late 80s using GHZ generators to get to a lower frequency would have rare and so not worthy of beating yourself up for.
@@mnoxman I briefly looked at the owners manual; the guys who designed it struggled to make that work. Anybody outside the narrow field of broadband RF test equipment would be lost... He is a physics major with specialty in optics - fantastic work to catch on to old tricks and dig in to repair that beast ! Back in the 80s, there were only a few hundred engineers capable of building it. And the manual is hard to read... my hats off. Fixing stuff is the best way to learn.
But a 1uF tantalum will have a resistance of milliohms. Ceramics might not be a good substitute.
no not a 1uF. I'll do a video on that in a few weeks.
How could they build it like that, it looks worse than my prototypes.
ceramics are microphonic, isn't it?
... a little story for ya? We hired a technician to calibrate and test broadband RF test equipment in the 80's. John was the lead tech responsible for the telemetry link from 2ncd stage on the Saturn V's. He told stories of launch day - "Lock , baby, lock" was their montra because of microphonics in the PLL synthesizers. (not only caps, board flexing, triboelectric effects and more) 🤣
Lucky for me, by RF generator is not flying and will live a boring vibration free life. Unless we get an earthquake.