HP 608C Repair.

[neo]

I have a 608C on my bench today and the problems include;
A) No measurable output.
B) Squirrelly meters.
C) Black beauty paper caps.

To start i am going to replace the paper caps with new ones, though i would appreciate advice on the first two problems. Meanwhile, here are some pictures to ogle at.

In order,
A picture of the transformer in this thing, note the quarter for reference.
The rear.

[neo]

More pictures.

Top view
Right side view.

Just a note, this thing weighs 100 pounds. The manual says 60 pounds but thats for the non rack-mount version, don't ask me where the extra 40 pounds is but i get the impression they beefed up the chassis.

[neo]

Right side paper caps replaced with capacitors up to modern snuff. At least the capacitors are, not sure about my soldering but neither the Heathkits nor the Hickoks have complained about it.

[Specmaster]

I have a 608C on my bench today and the problems include;
A) No measurable output.
B) Squirrelly meters.
C) Black beauty paper caps.

To start i am going to replace the paper caps with new ones, though i would appreciate advice on the first two problems. Meanwhile, here are some pictures to ogle at.

In order,
A picture of the transformer in this thing, note the quarter for reference.
The rear.

I expect that the transformer might be an oil filled one, was an ex military unit? I had an old oscilloscope that had oil filled transformers and capacitors as well. These photos show the unit, these are not my photos but some I found on the web.

[neo]

I have a 608C on my bench today and the problems include;
A) No measurable output.
B) Squirrelly meters.
C) Black beauty paper caps.

To start i am going to replace the paper caps with new ones, though i would appreciate advice on the first two problems. Meanwhile, here are some pictures to ogle at.

In order,
A picture of the transformer in this thing, note the quarter for reference.
The rear.

I expect that the transformer might be an oil filled one, was an ex military unit? I had an old oscilloscope that had oil filled transformers and capacitors as well. These photos show the unit, these are not my photos but some I found on the web.

Honestly, i don't know about either. It wouldn't surprise me if it was oil filled transformer and it definitely wouldn't surprise me if it was military. The only things i have ever seen that were completely devoid of stickers has been military, my bet is either the navy, the air force, or a very specialized lab.

It was made in 1973.

[amspire]

Looks like it is pretty good performance for a 1954 RF oscillator design. They were still making it in 1973? Drift less then 0.005% over 10 minutes. They really worked hard on the stability - low tempco Invar steel, split stator variable capacitor so there were no sliding electrical contacts. 1% accuracy and 0.1% resettability. The 608D model includes a crystal so you get 0.01% accurate reference points every 5MHz. I gather you used headphones to zero the beat frequency.

http://www.hpl.hp.com/hpjournal/pdfs/IssuePDFs/1954-08.pdf

Only cost $850 new in 1954. Probably half a family car.

[neo]

Looks like it is pretty good performance for a 1954 RF oscillator design. They were still making it in 1973? Drift less then 0.005% over 10 minutes. They really worked hard on the stability - low tempco Invar steel, split stator variable capacitor so there were no sliding electrical contacts. 1% accuracy and 0.1% resettability. The 608D model includes a crystal so you get 0.01% accurate reference points every 5MHz. I gather you used headphones to zero the beat frequency.

http://www.hpl.hp.com/hpjournal/pdfs/IssuePDFs/1954-08.pdf

Only cost $850 new in 1954. Probably half a family car.

$850 in 1954 WAS a family car. Accounting for inflation that is $18k

[amspire]

Output accurate to 1db across all frequencies from -127dBm to +7 dBm. Below 0dBm, the accuracy is typically double or more. 10Mhz to 480MHz. That is a good generator!

[neo]

I can't find the last 4 paper caps, they hid them well. 
Those 4 may only be in the 608D, thats the only manual i can get.


C#   uF        ESR    VAL
C52  15       3      10
C52  12.44    3.4    10  second half
C42  106      0.56   80
C55  30       1.8    10
C40  112      0.49   80
C75  1300     0.21   1000
C45  53.9     0.59   45
C46  15.62    2.5    10 
C7   26.36    2.3    10
C7   23.04    3      10

They did something odd with C75, i had to pull it to test it. If you say it's good i'll put it back in taking extra care to get it exactly as it was, without a manual i stand a snowballs chance of actually understanding it.

The rear with new caps and the left side.

[neo]

Left side without shield and transformer.

[amspire]

They did something odd with C75, i had to pull it to test it. If you say it's good i'll put it back in taking extra care to get it exactly as it was, without a manual i stand a snowballs chance of actually understanding it.

The rear with new caps and the left side.

Ah! You have had the modifcation for regulated DC filament operation:

http://bama.edebris.com/download/hp/608c/HP_608C_Service_Note.pdf

C75 is a 1000uF 25VW electro for the filament supply. Could be worth replacing, but for the job, your existing still looks adequate. It probably only powers the filaments of one or two valves so it is not anything special. The reason for the mod was to reduce ripple on the RF output. This mod means your valve VHF oscillator is infected with a transistor!

[neo]

They did something odd with C75, i had to pull it to test it. If you say it's good i'll put it back in taking extra care to get it exactly as it was, without a manual i stand a snowballs chance of actually understanding it.

The rear with new caps and the left side.

Ah! You have had the modifcation for regulated DC filament operation:

http://bama.edebris.com/download/hp/608c/HP_608C_Service_Note.pdf

C75 is a 1000uF 25VW electro for the filament supply. Could be worth replacing, but for the job, your existing still looks adequate. The reason for the mod was to reduce ripple on the RF output. This mod means your valve VHF oscillator is infected with a transistor!

Thanks! 

As i said it was made in 1973 (as far as i know), a transistor in the ointment was inevitable. 

[amspire]

As i said it was made in 1973 (as far as i know), a transistor in the ointment was inevitable. 

I can believe they were selling this after 19 years of manufacturing. It is a very good VHF oscillator today - except for the size and weight. There was still a lot of valve stuff available in 1973.

[neo]

I've still not found a culprit.

Replaced all paper caps, checked all eloctrolytics, put everything back as it was and i turned the thing on.
Still no output, percent modulation meter is still screwy. Everything glows, v15 which is a 5651 particularly brightly.

[neo]

Looking at the front panel for the first time in months i stand corrected there are stickers and it wasn't made in 1973, it was cataloged in 1973.

It was owned by the United States Navy, circa 1973.  So now it just jumped to the top of my priority list to get this thing working.

[amspire]

Luckily, it is fairly easy to debug with basically one valve per function. If I were debugging this with an oscilloscope, I would probably be using one of the cheap 100x high voltage probes that are available for about $20. You could make a simple capacitor isolated 1N4148 RF detector probe and just go searching for RF at the oscillator, buffer and power amplifier.

Also the resistors can drift high - probably worth checking. Many of them look like carbon composition resistors which are often not particularly stable.

The RF band switch is critical - if the contacts need attention, that will kill the oscillator.

[neo]

Luckily, it is fairly easy to debug with basically one valve per function. If I were debugging this with an oscilloscope, I would probably be using one of the cheap 100x high voltage probes that are available for about $20. You could make a simple capacitor isolated 1N4148 RF detector probe and just go searching for RF at the oscillator, buffer and power amplifier.

Also the resistors can drift high - probably worth checking. Many of them look like carbon composition resistors which are often not particularly stable.

The RF band switch is critical - if the contacts need attention, that will kill the oscillator.

The only thing of those i know how to do is check the resistors, though that will take some time. I'll get started on that.

[amspire]

Luckily, it is fairly easy to debug with basically one valve per function. If I were debugging this with an oscilloscope, I would probably be using one of the cheap 100x high voltage probes that are available for about $20. You could make a simple capacitor isolated 1N4148 RF detector probe and just go searching for RF at the oscillator, buffer and power amplifier.

Also the resistors can drift high - probably worth checking. Many of them look like carbon composition resistors which are often not particularly stable.

The RF band switch is critical - if the contacts need attention, that will kill the oscillator.

The only thing of those i know how to do is check the resistors, though that will take some time. I'll get started on that.

When you are checking them, remember most are +/- 20% tolerance. You are probably used to 5% or better. So if a 1K resistor is somewhere between 800 ohms and 1.2K, it is still within tolerance. If there are only 3 colour bands on the resistor with no tolerance band, they are 20%.

[neo]

Luckily, it is fairly easy to debug with basically one valve per function. If I were debugging this with an oscilloscope, I would probably be using one of the cheap 100x high voltage probes that are available for about $20. You could make a simple capacitor isolated 1N4148 RF detector probe and just go searching for RF at the oscillator, buffer and power amplifier.

Also the resistors can drift high - probably worth checking. Many of them look like carbon composition resistors which are often not particularly stable.

The RF band switch is critical - if the contacts need attention, that will kill the oscillator.

The only thing of those i know how to do is check the resistors, though that will take some time. I'll get started on that.

When you are checking them, remember most are +/- 20% tolerance. You are probably used to 5% or better. So if a 1K resistor is somewhere between 800 ohms and 1.2K, it is still within tolerance. If there are only 3 colour bands on the resistor with no tolerance band, they are 20%.

Thank you for reminding me and making sure i was aware, though i already knew that.

I ran the band switch through a few times with contact cleaner and it didn't make a difference, still no output.

[neo]

It might be of note that when i got this thing it worked, squirrelly meters not withstanding. Then it just stopped, i suppose it could of been on the edge and just said "enough"...

[Specmaster]

It might be of note that when i got this thing it worked, squirrelly meters not withstanding. Then it just stopped, i suppose it could of been on the edge and just said "enough"...

Have you had the tubes tested, it could be as simple as replacing a duff tube or two.

[neo]

It might be of note that when i got this thing it worked, squirrelly meters not withstanding. Then it just stopped, i suppose it could of been on the edge and just said "enough"...

Have you had the tubes tested, it could be as simple as replacing a duff tube or two.

Unfortunately I've no tube tester and the drug store variety has long since gone the way of the dodo.

[amspire]

Even without a tub tester, I think it will be a very interesting exercise in fault finding, because you can just work through it stage by stage. Are the supply voltages OK? Is the 225V series regulator OK. Is the 400Hz modulation oscillator OK? 1KHz? The Pulse generator? The transistor DC filament regulator OK? The Oscillator? The Buffer? The Output Amplifier? The Level meter?

You will get there.

[neo]

Even without a tub tester, I think it will be a very interesting exercise in fault finding, because you can just work through it stage by stage. Are the supply voltages OK? Is the 225V series regulator OK. Is the 400Hz modulation oscillator OK? 1KHz? The Pulse generator? The transistor DC filament regulator OK? The Oscillator? The Buffer? The Output Amplifier? The Level meter?

You will get there.

All of which would be easier with the correct manual, all i can find is the 608D.

It still will be a very interesting experiment and exercise. 

Edit to note: I'm not relying on the meters to tell me if i have an output, i put the output on my scope and got nothing.

[Specmaster]

It might be of note that when i got this thing it worked, squirrelly meters not withstanding. Then it just stopped, i suppose it could of been on the edge and just said "enough"...

Have you had the tubes tested, it could be as simple as replacing a duff tube or two.

Unfortunately I've no tube tester and the drug store variety has long since gone the way of the dodo.

Hmm, I know what you mean, but I've discovered that a local company to me that specialises in sourcing hard to find parts actually has a tester for their own use and will also test for customers. Try looking on line for companies that sell tubes in your area, they will also more than likely be able to test them for you.