Call Search
     

New to Ham Radio?
My Profile

Community
Articles
Forums
News
Reviews
Friends Remembered
Strays
Survey Question

Operating
Contesting
DX Cluster Spots
Propagation

Resources
Calendar
Classifieds
Ham Exams
Ham Links
List Archives
News Articles
Product Reviews
QSL Managers

Site Info
eHam Help (FAQ)
Support the site
The eHam Team
Advertising Info
Vision Statement
About eHam.net

   Home   Help Search  
Pages: [1] 2 3 Next   Go Down
  Print  
Author Topic: RCA 813 150 watt oscillator  (Read 10694 times)
K0WRX
Member

Posts: 8




Ignore
« on: November 02, 2011, 01:46:44 PM »

Just wondering if anyone has built the RCA 150 watt oscillator they advertised years and years ago?  Seems very interesting and I am thinking about assembling one.  Info on the circuit they offered would be appreciated.  Thank you...Pete...K0WRX
Logged
N2EY
Member

Posts: 3785




Ignore
« Reply #1 on: November 02, 2011, 03:03:43 PM »

I have seen the original "Ham Tips" RCA periodical with that rig. It was from 1938 IIRC.

I wouldn't build one except as a curiosity. Here's why:

1) The pre-WW2 crystals had huge slabs of quartz and could take a lot of crystal current. Modern crystals can't. Even an FT-243 would be working very hard. Good ham-band crystals aren't common today, why risk them?

2) You only get one band per xtal, and you can't use VFO without major modifications.

3) The keying system for such a rig must involve a relay capable of withstanding the high voltage.

4) The chances of getting a good T9X signal are pretty slim.

5) You need 1000-1500 volts B+, and parts that can stand it.

6) The tx will produce 100 watts output at most. There are easier ways to get that much RF from a simple tube tx.

For example, a 6AG7 oscillator driving a pair of 807s or 1625s will produce 100 watts output, be much easier on the crystal, can double in the oscillator (2 bands for some crystals) and requires only 750 volts for the final plates and 300-350 volts for the oscillator. And it can be driven by a VFO without major modifications (you need about 20 volts of RF from the VFO, but that's not difficult).

73 de Jim, N2EY

Logged
K0WRX
Member

Posts: 8




Ignore
« Reply #2 on: November 02, 2011, 03:47:11 PM »

Jim...Thank you, this is exactly the kind of info I am looking for.  Your input is well taken and respected.  Thank you for your time spent answering .....Pete...K0WRX
Logged
G3RZP
Member

Posts: 4099




Ignore
« Reply #3 on: November 03, 2011, 07:22:57 AM »

Jim said:

>The pre-WW2 crystals had huge slabs of quartz and could take a lot of crystal current<

And even then, with a 60 mA flashlight bulb in series, crystal fractures weren't unknown at those sort of power levels. They were generally X or Y cut, which is why they were so thick.

An 813 will do more power as an amplifier than a pair of 807s, but you need  1 kV on the plate for 150 watts input. At 2kV, you can still drive them with a 6AG7 to around 350 or 400 watts input, but with modern crystals (or even FT243s) you would be better off with a low power  osc driving the 6AG7.
Logged
VA3AEX
Member

Posts: 22




Ignore
« Reply #4 on: November 04, 2011, 10:24:42 AM »

Have a look at this website http://www.eclipseofthechurch.org/w7ekb/glowbugs/tx/813osc/813osc.htm
Interestingly, it shows that W7QQQ did build the same circuit, although I'd also be worried by the voltage across the key!

73
Logged
N2EY
Member

Posts: 3785




Ignore
« Reply #5 on: November 06, 2011, 07:22:03 AM »

Have a look at this website http://www.eclipseofthechurch.org/w7ekb/glowbugs/tx/813osc/813osc.htm
Interestingly, it shows that W7QQQ did build the same circuit, although I'd also be worried by the voltage across the key!

The original writeup specifically says to use a keying relay, and NOT to put the key in the circuit. That's good advice - but is a reason not to build the thing.

The 813 was RCA's New Thing in 1938, and they wanted to show off what it could do.

73 de Jim, N2EY
Logged
G3UUR
Member

Posts: 119




Ignore
« Reply #6 on: November 07, 2011, 01:54:06 AM »

The original writeup specifically says to use a keying relay, and NOT to put the key in the circuit. That's good advice - but is a reason not to build the thing.

You're obviously dead set on discouraging Pete from building this 813 150-watt oscillator, Jim, but are you suggesting here that he shouldn't build it because even with a keying relay it will be dangerous, or is it that you think keying relays will be hard to come by these days?

Personally, I think it's a fun project and safe enough if Pete's sensible and doesn't take any risks when testing or trouble shooting it. As far as keying the screen is concerned, that should be quite simple to do in a safe way these days with HV switching MOSFETs and opto-isolators.

On the subject of modern crystals and crystal current, you might be surprised what some of these little critters can stand. The original article states that the crystal current is 14.2mA on CW, though I suspect that it'll be a lot higher if the anode circuit is not tuned properly. This sort of current is well within the capability of modern HC-18/U, HC-25/U and HC-49/U crystals on 80 and 40m. 

73, Dave.
Logged
W8JI
Member

Posts: 9302


WWW

Ignore
« Reply #7 on: November 07, 2011, 04:41:41 PM »

I'm afraid I'm going along with N2EY in encouraging him to not build the combined oscillator-PA in that link.

An oscillator driving the antenna is bad enough, but this circuit is a disaster for several reasons:

1.) It keys the screen, but has no pull down resistor on the screen. Screens in high voltage high power tubes should never float, and that is what it does when the relay opens

2.)  It will be difficult to get a safe relay

3.) The only thing keeping 1500 volts plus the peak RF voltage off the link is the insulation in the link system. The link is a horrible design for safety

4.) C4 has the rotor and shaft hot with at least 1500 volts 

 5.) The filament supply is hot with RF. It is not RF bypassed to ground, and cannot be bypassed of the oscillator will not work as intended by feedback from L1/C3, so that means all that RF on the filament is running right back into the filament system

6.)  Beam forming plates are tied back to the floating cathode. That's bad news for stability

I can't recall ever seeing a worse circuit design.

73 Tom
Logged
G3UUR
Member

Posts: 119




Ignore
« Reply #8 on: November 08, 2011, 03:39:57 AM »

I'm afraid I'm going along with N2EY in encouraging him to not build the combined oscillator-PA in that link.

An oscillator driving the antenna is bad enough, but this circuit is a disaster for several reasons:

1.) It keys the screen, but has no pull down resistor on the screen. Screens in high voltage high power tubes should never float, and that is what it does when the relay opens

2.)  It will be difficult to get a safe relay

3.) The only thing keeping 1500 volts plus the peak RF voltage off the link is the insulation in the link system. The link is a horrible design for safety

4.) C4 has the rotor and shaft hot with at least 1500 volts 

 5.) The filament supply is hot with RF. It is not RF bypassed to ground, and cannot be bypassed of the oscillator will not work as intended by feedback from L1/C3, so that means all that RF on the filament is running right back into the filament system

6.)  Beam forming plates are tied back to the floating cathode. That's bad news for stability

I can't recall ever seeing a worse circuit design.

73 Tom

Everyone is entitled to his or her own opinion, Tom. However, I think calling the design a disaster is highly exaggerated. It's primitive, certainly, and rather poor by todays standards, yes, but not a disaster unless it blows up when you first key up. Since Pete has a K0 call, I would think he has a bit of experience and is not a newbie. He could get a lot of fun out of using an old transmitter like the RCA 813 oscillator design and there are several on the air already. Incidentally, they don't sound as bad as some Yaesu FT1000s I've heard on CW. Any harmonic problems can be sorted out with some additional low-pass filtering. Instability doesn't appear to be a problem despite the beam plates being connected to the filament and varying in voltage with RF.

Floating the screen in this design is not a problem and a pull-down resistor would probably kill the weak oscillation that occurs while the screen is open and produce chirp. The charge in the screen decoupling capacitor supplies the screen current for a short period after the key is opened and the capacitor voltage decays until the screen current drops to zero. There's always a finite but small screen voltage there. Keying the screen with a HV switching MOSFET has already been done, I believe.

With insulation on both the coil and the link, I see no more of a safety hazard than using a single blocking capacitor in a parallel-fed PA. That's frequently done with no choke or resistor to ground at the pi-network output. They rely on there being a DC path to ground in the ASTU or antenna, which isn't always there. So, modern practice is very often no better, though I certainly wouldn't condone it because it's poor practice.

C4 can easily be arranged to have the rotor earthed. That's a minor issue.

The filament transformer is usually mounted near to the tube and the primary can be decoupled and filtered to keep RF out of the mains electricity supply.

I'd agree that an improved version ought to be described on the web somewhere to help those who might want to make one, but I wouldn't discourage them from building one. They have every right to their bit of fun just the same as us.

Were you a health and safety man in a previous life, Tom?

73, Dave.     
Logged
G3RZP
Member

Posts: 4099




Ignore
« Reply #9 on: November 08, 2011, 11:27:28 AM »

What issue off Ham Tips is this in?
Logged
G3UUR
Member

Posts: 119




Ignore
« Reply #10 on: November 08, 2011, 11:41:40 AM »

What issue off Ham Tips is this in?

Vol.-1, No. 4, December, 1938. If you've got your own copy, the circuit is in the lower left-hand corner (bottom of first column) and continues in the third column of the first page, then finishes in the third column on the back.
Logged
W8JI
Member

Posts: 9302


WWW

Ignore
« Reply #11 on: November 08, 2011, 11:50:20 AM »

Everyone is entitled to his or her own opinion, Tom. However, I think calling the design a disaster is highly exaggerated.


It is a terrible circuit for many reasons.

Quote
It's primitive, certainly, and rather poor by todays standards, yes, but not a disaster unless it blows up when you first key up.

It's a terrible RF system. They do not even isolate the RF into the filament transformer!

It's also terrible for safety for many reasons.

Quote
Floating the screen in this design is not a problem and a pull-down resistor would probably kill the weak oscillation that occurs while the screen is open and produce chirp. The charge in the screen decoupling capacitor supplies the screen current for a short period after the key is opened and the capacitor voltage decays until the screen current drops to zero. There's always a finite but small screen voltage there. Keying the screen with a HV switching MOSFET has already been done, I believe.

So we have a MOSFET isolating a key line from 1500 volts? Seems like cathode keying would be better.

So we have the screen floating positive long enough to keep the crystal running as a capacitor slowly ramps down, but ramping down fast enough to not output RF?

Who are we kidding here??? It would have terrible backwave if the crystal stayed on.

Quote
With insulation on both the coil and the link, I see no more of a safety hazard than using a single blocking capacitor in a parallel-fed PA. That's frequently done with no choke or resistor to ground at the pi-network output.


That would be pretty silly, to omit the safety choke or use a resistor. But the link is a more likely problem.

Quote
They rely on there being a DC path to ground in the ASTU or antenna, which isn't always there. So, modern practice is very often no better, though I certainly wouldn't condone it because it's poor practice.

That's right. It is an unsafe practice.

Quote
C4 can easily be arranged to have the rotor earthed. That's a minor issue.

Sure, by adding a blocking cap and a choke the rotor can be earthed. Might as well change to a pi network so coupling can be adjusted. It would only take one or two more parts. :-)

Quote
The filament transformer is usually mounted near to the tube and the primary can be decoupled and filtered to keep RF out of the mains electricity supply.

They don't show that.

Quote
I'd agree that an improved version ought to be described on the web somewhere to help those who might want to make one, but I wouldn't discourage them from building one. They have every right to their bit of fun just the same as us.

I'm not trying to stop anyone's fun. I just don't think it is a good idea to learn on a death trap.

By the way, in the USA a callsign is no indication of license term.  A K0 or W0 can be a new no code or someone licensed for 75 years. Not that that matters one way or another. Unsafe poor designs are unsafe poor designs, and this circuit is terrible when, for just a few dollars more, it can be made safe.

Quote
Were you a health and safety man in a previous life, Tom?

No, I just don't like to kill people or see people kill themselves for no reason. It would be just as easy to build something safe, instead of a death trap.

If someone is trying to learn and have fun, it would make more sense to use a few hundred volts and cathode key. 1500 volts (plus peak RF voltage), and no coupling adjustment on an oscillator directly coupled to an antenna, makes little sense as a fun learning project.   

73 Tom

« Last Edit: November 08, 2011, 12:05:26 PM by W8JI » Logged
W8JI
Member

Posts: 9302


WWW

Ignore
« Reply #12 on: November 08, 2011, 12:06:37 PM »

What issue off Ham Tips is this in?

Here is a copy:

http://www.eclipseofthechurch.org/w7ekb/glowbugs/tx/813osc/813osc.htm
Logged
G3RZP
Member

Posts: 4099




Ignore
« Reply #13 on: November 09, 2011, 01:19:28 AM »

It looks like a tri-tet, but the cathode is tuned to around 1.5MHz - rather low. The numbers on RF crystal current sound suspiciously low to me, but still represent around 9 mW in a typical AT cut crystal, which is at the top end for modern crystals. The spacer is another matter - but in those days, you very rarely both worked on the same frequency. Like VHF in the 60s, where you called CQ and tuned the band for a reply.

With the much lower mass of modern crystals, I doubt that letting it oscillate weakly in key up would avoid drift, as it would heat up rapidly in key down.

I think that there are better and safer ways to use an 813.....
Logged
G3UUR
Member

Posts: 119




Ignore
« Reply #14 on: November 09, 2011, 10:56:30 AM »

[quote author=W8JI link=topic=78685.msg553060#msg553060 date=1320781820
It is a terrible circuit for many reasons.

It's a terrible RF system. They do not even isolate the RF into the filament transformer!

It's also terrible for safety for many reasons.

Quote
Floating the screen in this design is not a problem and a pull-down resistor would probably kill the weak oscillation that occurs while the screen is open and produce chirp. The charge in the screen decoupling capacitor supplies the screen current for a short period after the key is opened and the capacitor voltage decays until the screen current drops to zero. There's always a finite but small screen voltage there. Keying the screen with a HV switching MOSFET has already been done, I believe.

So we have a MOSFET isolating a key line from 1500 volts? Seems like cathode keying would be better.

So we have the screen floating positive long enough to keep the crystal running as a capacitor slowly ramps down, but ramping down fast enough to not output RF?

Who are we kidding here??? It would have terrible backwave if the crystal stayed on.

Quote
C4 can easily be arranged to have the rotor earthed. That's a minor issue.

Sure, by adding a blocking cap and a choke the rotor can be earthed. Might as well change to a pi network so coupling can be adjusted. It would only take one or two more parts. :-)

Quote
The filament transformer is usually mounted near to the tube and the primary can be decoupled and filtered to keep RF out of the mains electricity supply.

They don't show that.

73 Tom[/quote]

Why would you need a blocking capacitor and choke if the output coil was tuned by a variable capacitor with the fixed vanes connected to the anode side of the coil and the rotor grounded? The decoupling capacitor provides the return path from the HV side of the tuned circuit.

They don't show any of the constructional details, Tom. It's left to the individual to work that out. Amateurs generally have more common sense than you seem to credit them with. As kids many of us played with circuits that were just as dangerous as this and we survived.

The MOSFET switching circuit I've heard about for keying the screen works from a 400V supply as far as I'm aware. Why would you make life harder for yourself by doing anything else?

Peter, the tests I did on HC49/u crystals some years back, when the question of how much current these little crystals could handle first came up, show that they are pretty difficult to destroy. Even pumping 100mA of crystal current through the 40m ones could not kill them. By measuring the maximum downward shift and then what they got up to on a long key down I was able to estimate that their internal temperature was reaching in excess of 200C but not as much as 300C. They eventually went unstable and the crystal current dropped back if the excessive current was applied too long. Remarkably, they appeared to recover and continued to work okay at a more reasonable crystal current. This is not to say that there was no change in the crystal performance or frequency. However, they were still usable for amateur purposes. The only modern crystals I could kill with my power Pierce oscillator were the squat HC-49/S and /US ones.

Dave.
Logged
Pages: [1] 2 3 Next   Go Up
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.11 | SMF © 2006-2009, Simple Machines LLC Valid XHTML 1.0! Valid CSS!