Author Topic: General Radio 1396B Tone Burst Generator - Understanding power supply  (Read 780 times)

0 Members and 1 Guest are viewing this topic.

Offline barbarojaTopic starter

  • Contributor
  • Posts: 22
  • Country: us
Hi. In the process of recapping a GenRad 1396B. The old caps (around50+ years old) seem to be ok but for one or two that don't.

Leaky caps are

(1) Mallory marked 400 400 25, 4450-5621 85C. Present in main PSU. Has two 400uF caps really.
(1) Sprague 4450-6020, 500MFD-20VDC D73710. Present in scaler PSU.

Those caps are not available anymore, and if found, replacements are very old as to guarantee trouble free operation. Analyzing the schematics, there are some caps before dropping resistors and some more caps of different value.

So here come the questions:

1. The schematics suggest that the circuit sees the whole value of the dual Mallory PSU caps (parallel connection). Wondering why they took that design decision (using dual caps. Maybe was avaibable at the time?), and if it would change performance of the device to use different (bigger) values. Standarized caps today go 100uF, 150uF, 220uF, 330uF, 470uF, 680uF.

2. Could they be using different size caps as to reduce inductive reactance? (Dave has a video on that). I find particularly interesting that both the main PSU side and the scaler PSU side use different values after the dropping resistor.

Circuit diagrams here: https://manuals.repeater-builder.com/te-files/GENRAD/GENRAD%201396B%20Instruction.pdf

There's so much to learn from vintage tech.

 

Online amyk

  • Super Contributor
  • ***
  • Posts: 8525
Re: General Radio 1396B Tone Burst Generator - Understanding power supply
« Reply #1 on: November 29, 2024, 02:33:37 am »
Those are just regular power supply decoupling. 680 and 820 would be fine for replacement. Keep in mind that the originals would've had 20% tolerance anyway.
 

Offline barbarojaTopic starter

  • Contributor
  • Posts: 22
  • Country: us
Re: General Radio 1396B Tone Burst Generator - Understanding power supply
« Reply #2 on: November 29, 2024, 04:44:10 am »
Those are just regular power supply decoupling. 680 and 820 would be fine for replacement. Keep in mind that the originals would've had 20% tolerance anyway.

Yeh I can imagine that. Any reason to use different values in the original design? (I mean why not use 820 for everything for example)
Just trying to understand it a lil better.
 

Offline coppercone2

  • Super Contributor
  • ***
  • Posts: 11333
  • Country: us
  • $
Re: General Radio 1396B Tone Burst Generator - Understanding power supply
« Reply #3 on: November 29, 2024, 04:48:09 am »
usually its

1) price
2) size
3) hearsay

The #1 thing you will notice is the change in capacitor film technology, aka improvement. They can make thinner films more reliably that do the job so the caps are smaller, and better electrolyte.

With old machinery you can guarantee the film is durable when its above a certain thickness. With newer processes, the film can be made thinner better. Possibly it has more dielectric resistance and stuff like that too. Or it might be the same stuff and just tested better to give you the 'reliable spec'.

Lets say you have a small instrument that uses a particular cap. If the big ones can use it, use it, then you can get more volume discount. They have to balance between quality and cost. It still came out really good. Its just like having a internal E-series they tell the designers to use, based on their manufacturing stockpile.

When they are too big it might stress things too, during turn on, absurd values require special management circuitry.


To me that one is kinda a loose spec. Like a ships anchor. It kind of depends on what the equipment is plugged into. But there is a minimum size that you can determine looking at ripple, theoretically, if you consider just a clean 60Hz sine wave with no distortion with some stable load plugged into the output of the equipment with some particular settings, at a standardized line impedance for the generic spec (automotive or residential/office like). But when you start having generators an heavy equipment nearby things can get very complicated I think. 


And parallel caps.. that has to do with
1) better form factor, its how you get a 'rectangle shape' instead of a 'tank shape'
2) lower ESR, its parallel after all
3) possibly more economical if you can pick from your companies "series". They could also pick from values they noticed statistically are in stock more or easier to source from a alternative manufacturer
4) because of #2, it probobly also filters RF better, since the caps are more ideal. That means maybe they can get rid of some other component, like a lowish value capacitor.. so really its possible that you end up with the same amount of capacitors because it turns out you need another one if you go with a big one that has higher parasitic values




Find the best lowest ESR capacitor you can find at some value, and compare it to the ESR you can get with parallel capacitors of equal quality but smaller size.  I am thinking the first one can be alot bigger, the second one, after the resistor, you want that one similar
« Last Edit: November 29, 2024, 05:15:38 am by coppercone2 »
 
The following users thanked this post: barbaroja

Offline coppercone2

  • Super Contributor
  • ***
  • Posts: 11333
  • Country: us
  • $
Re: General Radio 1396B Tone Burst Generator - Understanding power supply
« Reply #4 on: November 29, 2024, 05:07:37 am »
and in this case, it does look like a RC filter, that is normally not seen so much, because its wasteful. I think with the RC filter design (it looks like a CRC, Pi filter) they are definately thinking about bode plots when they build it that way, aka they want the frequency response to go out further, meaning capacitor parasitics might be being considered

I am kind of thinking if you combine the 600 and the 800, you might still want to put a 50uF or something in place of the 600, to get in on whatever they were trying to do with what looks like a RC filter


And that resistor, if you just have it right after a big cap, it might choke out the circuit? And if you put the resistor first, it might surge the circuit (the capacitor is too big)

I think, espeically if you put new diodes, the first cap can actually be made quite a bit bigger, and the second cap after the resistor.. you might wanna stick to the value they have. Like a 680uF or 820uF one would be the common choice
« Last Edit: November 29, 2024, 05:18:08 am by coppercone2 »
 
The following users thanked this post: barbaroja

Offline jonpaul

  • Super Contributor
  • ***
  • Posts: 3656
  • Country: fr
  • Analog, magnetics, Power, HV, Audio, Cinema
    • IEEE Spectrum
Re: General Radio 1396B Tone Burst Generator - Understanding power supply
« Reply #5 on: November 29, 2024, 05:17:40 am »
A wonderful and unique instrument, I first encountered at Dolby Labs in San Fran 1976. Needed to test audio noise reduction systems like DBX, Dolby.


The R-C filters are common 1950 design practice. See 1935 RCA rcv tube manual.

We used  similar R-C  filters in our IEEE Spectrum 1942 SIGSALY  ADC reconstruction.

Multisection lytics are  seen  in epoch radios, TVs, instruments to save space and fewer chassis punches.

Vintage Rebuilders   remove and gut the alum lytics,    rebuild the cans caps with small modern axial/radil hidden in the old can.

See Antique Radio forums.

Jon

An Internet Dinosaur...
 

Offline coppercone2

  • Super Contributor
  • ***
  • Posts: 11333
  • Country: us
  • $
Re: General Radio 1396B Tone Burst Generator - Understanding power supply
« Reply #6 on: November 29, 2024, 05:20:14 am »
The interesting choice you have now is getting rid of the 1uF electrolytic and putting a foil or ceramic there.. I almost feel like these instruments tend to have problem with the modern inverter noise laiden power grid, and need some improvements. They did not have good choices for 1uF back then. The electrolytic caps between 0.5 to 5uF now.. are kind of going obsolete as a design choice. Surely it would be designed with a tantalum capacitor if it was designed some where past.. 1985?
« Last Edit: November 29, 2024, 05:22:44 am by coppercone2 »
 

Online fzabkar

  • Super Contributor
  • ***
  • Posts: 2805
  • Country: au
Re: General Radio 1396B Tone Burst Generator - Understanding power supply
« Reply #7 on: November 29, 2024, 05:31:46 am »
CR431 and CR430 are sitting at 1.2V below earth, yet they are connected to the -15V supply rail. Something is wrong here ...
 

Offline coppercone2

  • Super Contributor
  • ***
  • Posts: 11333
  • Country: us
  • $
Re: General Radio 1396B Tone Burst Generator - Understanding power supply
« Reply #8 on: November 29, 2024, 05:37:29 am »
what might be interesting too is a picture of the transformer(s)

the thought about not having a giant capacitor connected directly to the transformer is valid, though it generally seems they can take whatever you manage to throw at them
« Last Edit: November 29, 2024, 05:39:22 am by coppercone2 »
 

Offline Harry_22

  • Frequent Contributor
  • **
  • Posts: 603
  • Country: 00
Re: General Radio 1396B Tone Burst Generator - Understanding power supply
« Reply #9 on: November 29, 2024, 09:46:28 pm »
CR431 and CR430 are sitting at 1.2V below earth, yet they are connected to the -15V supply rail. Something is wrong here ...

Yes, you are right.
An error has crept into the diagram. They meant 1.5V
 

Offline barbarojaTopic starter

  • Contributor
  • Posts: 22
  • Country: us
Re: General Radio 1396B Tone Burst Generator - Understanding power supply
« Reply #10 on: November 29, 2024, 10:30:46 pm »
and in this case, it does look like a RC filter, that is normally not seen so much, because its wasteful. I think with the RC filter design (it looks like a CRC, Pi filter) they are definately thinking about bode plots when they build it that way, aka they want the frequency response to go out further, meaning capacitor parasitics might be being considered

I am kind of thinking if you combine the 600 and the 800, you might still want to put a 50uF or something in place of the 600, to get in on whatever they were trying to do with what looks like a RC filter


And that resistor, if you just have it right after a big cap, it might choke out the circuit? And if you put the resistor first, it might surge the circuit (the capacitor is too big)

I think, espeically if you put new diodes, the first cap can actually be made quite a bit bigger, and the second cap after the resistor.. you might wanna stick to the value they have. Like a 680uF or 820uF one would be the common choice

Awesome insights, thank you. Makes a lot of sense. I think I will try to go for the most similar values but now at least I understand why. Replacing the 1uF for film seems like a nice idea as well. Done that before in other equipment I recapped.

Edit: Just tested the "good" caps with a DER EE DE-5000 LCR meter @ 120Hz.

Main supply
The dual 300uF, 35V
1: Square 352uF 0.2ohm, Triangle 361uF 0.2ohm
2: Sq 338uF 0.2ohm, Tr 351uF 0.2ohm

Dual 400uF, 25V
1: Sq 548uF, 0.1ohm, tr 541uF, 0.1ohm
2: Sq 63nF, 130Kohm, tr 534uF, 0.1ohm, leaky.

Scaler supply
Sprague 500, 20V: 530uF, 0.2ohm although it leaked all over the mainboard.
Mallory 680, 15V: 720uF, 0.1uF.



A wonderful and unique instrument, I first encountered at Dolby Labs in San Fran 1976. Needed to test audio noise reduction systems like DBX, Dolby.


The R-C filters are common 1950 design practice. See 1935 RCA rcv tube manual.

We used  similar R-C  filters in our IEEE Spectrum 1942 SIGSALY  ADC reconstruction.

Multisection lytics are  seen  in epoch radios, TVs, instruments to save space and fewer chassis punches.

Vintage Rebuilders   remove and gut the alum lytics,    rebuild the cans caps with small modern axial/radil hidden in the old can.

See Antique Radio forums.

Jon



Thanks jonpaul. Very interesting, had no idea about these considerations. I will definitely have a look to that manual.

what might be interesting too is a picture of the transformer(s)

the thought about not having a giant capacitor connected directly to the transformer is valid, though it generally seems they can take whatever you manage to throw at them

Here you go. Picture attached.

Thanks for pinpointing the diagram error, fzabkar and Harry_22.
« Last Edit: November 29, 2024, 11:12:59 pm by barbaroja »
 

Offline barbarojaTopic starter

  • Contributor
  • Posts: 22
  • Country: us
Re: General Radio 1396B Tone Burst Generator - Understanding power supply
« Reply #11 on: November 29, 2024, 11:11:46 pm »
Measuring some good caps I have lying around:

330uF 25V Elna Silmic II: 300uF, 0.2ohm. Spot on with the old dual caps. Got no 680uF at the moment.
470uF 35V Panasonic FR: 506uF, 0.0ohm.


 

Offline coppercone2

  • Super Contributor
  • ***
  • Posts: 11333
  • Country: us
  • $
Re: General Radio 1396B Tone Burst Generator - Understanding power supply
« Reply #12 on: November 30, 2024, 12:07:35 am »
That transformer looks substantial for like a tone generator, I think it can handle pretty much any reasonable capacitor. I thought it might be some micro sized thing

Those capacitor all have good ESR other then the leaky one, it depends on what you believe. I cut alot of those apart and sometimes they are brown inside and it gave me the creeps so I typically replace them, but 99% of people would leave them alone. I actually saw lower ESR in the old ones then the premium ones I replace them with that are usually much smaller but have equally nice temperature/life specs.

The one that measures as leaky, that is a sign of some kind of flaw, you should definitely replace that one

There is extended tests you can do like charge them up to max voltage and see how they hold charge if you are interested


You can also smell them. They all might smell a little but if one smells alot more that is another good reason to replace it.

However,


Dual capacitors in 1 enclosure are NICE parts. I would not replace that unless you find something wrong with it. This is a special case IMO. So it looks like you should replace two, the physically leaked one and the one that measured bad. I think they have a performance benefit of being in the same package that helps the circuit they are powering, and changing them would make it worse.


And since its a gen-rad equipment.. you might wanna just take the time and wait / search for a true dual cap replacement, unless you need it right now. That is one of those things where 'they don't build them like they used to'.

Dual electrolytics, and tapped foil/paper/oil caps are NICE as hell, from a EM and a manufacturing quality standpoint. They got rid of that for cost reasons. Same with axial electrolytics, you can still get them, but less popular, however from a EM standpoint for certain circuits... its just 'correct' compared to a can. Like for coupling.
« Last Edit: November 30, 2024, 12:18:35 am by coppercone2 »
 
The following users thanked this post: barbaroja

Offline coppercone2

  • Super Contributor
  • ***
  • Posts: 11333
  • Country: us
  • $
Re: General Radio 1396B Tone Burst Generator - Understanding power supply
« Reply #13 on: November 30, 2024, 12:23:52 am »
I always hope that eventually some business will come up that can make custom foil and electrolytic capacitors that can come in the form you want. Axial, dual, back to back, isolated foil, tapped foil (dividers), single layer ceramic networks, etc. There is a vast amount of 'premium engineering' parts that are just not available or ever actually sold that can be made IMO. And some other fancy shit like electrolytics with feedthrough capacitors built in  :-DD

Also, case positive caps (reverse polarity) for axial, imagine you can get both polarity axial, with whatever ceramic 'feedthrough' built in, designed to be grounded by a braid for chassis work. So it would combine your ceramic and electrolytic cap into a low impedance ground.
« Last Edit: November 30, 2024, 12:29:09 am by coppercone2 »
 

Offline barbarojaTopic starter

  • Contributor
  • Posts: 22
  • Country: us
Re: General Radio 1396B Tone Burst Generator - Understanding power supply
« Reply #14 on: November 30, 2024, 09:33:12 am »
That transformer looks substantial for like a tone generator, I think it can handle pretty much any reasonable capacitor. I thought it might be some micro sized thing

Those capacitor all have good ESR other then the leaky one, it depends on what you believe. I cut alot of those apart and sometimes they are brown inside and it gave me the creeps so I typically replace them, but 99% of people would leave them alone. I actually saw lower ESR in the old ones then the premium ones I replace them with that are usually much smaller but have equally nice temperature/life specs.

The one that measures as leaky, that is a sign of some kind of flaw, you should definitely replace that one

There is extended tests you can do like charge them up to max voltage and see how they hold charge if you are interested


You can also smell them. They all might smell a little but if one smells alot more that is another good reason to replace it.

However,


Dual capacitors in 1 enclosure are NICE parts. I would not replace that unless you find something wrong with it. This is a special case IMO. So it looks like you should replace two, the physically leaked one and the one that measured bad. I think they have a performance benefit of being in the same package that helps the circuit they are powering, and changing them would make it worse.


And since its a gen-rad equipment.. you might wanna just take the time and wait / search for a true dual cap replacement, unless you need it right now. That is one of those things where 'they don't build them like they used to'.

Dual electrolytics, and tapped foil/paper/oil caps are NICE as hell, from a EM and a manufacturing quality standpoint. They got rid of that for cost reasons. Same with axial electrolytics, you can still get them, but less popular, however from a EM standpoint for certain circuits... its just 'correct' compared to a can. Like for coupling.

I did want to use it to calibrate the current limiters on some Westrex solid state amps for Vinyl lathe. The manual specifically recommends this machine. It states one cycle on, sixteen off with a certain Peak to Peak voltage, into a 1ohm load, with bigger fuses and then you trim some pots that control FETs wired as current limiters. However, I found a suitable replacement with the software REW. Can create whatever test signal you need.

I will try to get the originals, the 1396B is worth it. However, googling for those pieces has proven them to be hard to find.

As to test the caps, sure, I can do it. Any ideal procedure or should I just charge them and measure the discharge time?

I'm attaching a pic of this beauty.
« Last Edit: November 30, 2024, 09:35:09 am by barbaroja »
 

Offline coppercone2

  • Super Contributor
  • ***
  • Posts: 11333
  • Country: us
  • $
Re: General Radio 1396B Tone Burst Generator - Understanding power supply
« Reply #15 on: November 30, 2024, 09:40:56 am »
I don't know, I am saying if you really need convincing you can figurre it out, but I would keep everything that measures ok.
 
The following users thanked this post: barbaroja

Offline barbarojaTopic starter

  • Contributor
  • Posts: 22
  • Country: us
Re: General Radio 1396B Tone Burst Generator - Understanding power supply
« Reply #16 on: November 30, 2024, 10:36:33 am »
coppercone2 big thanks for your help, learning a lot.

Alright, just measured all of them. Just a made-up procedure where I charged them all to 10v with a controlled current and let them sit to see how they discharge. I recall reading in a manual of old vintage gear (cannot remember which one) that one could measure the discharge time with no load as a indicative of cap health. All of them but the big dual 400uF on the square terminal took more than 6 or 7 minutes to discharge to like 25%. The bad one was just a few seconds, so I guess the test does work. Interesting, the leaky 500uF Sprague on the scaler supply did just like the good ones. That one should not be as hard to get, tho.

 

Offline coppercone2

  • Super Contributor
  • ***
  • Posts: 11333
  • Country: us
  • $
Re: General Radio 1396B Tone Burst Generator - Understanding power supply
« Reply #17 on: November 30, 2024, 10:13:24 pm »
it is possible that it leaked and someone replaced it without cleaning pcb and that it is a good cap, hard to tell unless there is visible damage near the leads or whatever

I notice the better repair people, they mark the replacement parts with something like a scribe or paint marker.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf