Magnetron- 4J50
King_of_Junk, Fri May 04 2007, 02:02AMI'm trying to power up a 4J50 magnetron that I purchased at Nortex Electronics, a surplus store in Ft. Worth.
I was wondering if anyone had any suggestions for the build (I'm rather lost on what to do for diodes)
This schematic was sent to me courtesy of Patrick Jankowiak of
It requires 23kV negative DC input.
I'm thinking about rectifying a 15-30 NST for the input. It should give 21 or 22kV DC (approximately 22.5 I believe)
But according to the mysterious yet helpful Patrick, a bridge rectifier may not be my only option...
I usually stick to AC, so I'm not particularly good with the behavior of DC and DC circuits
So far, I have every element of the circuit, or something that should work, except for the diodes.
I was wondering if there is any easy way to make high voltage diodes? I mean, of course, there are ways to make your own high voltage capacitors, and winding your own transformers is not unheard of...
I suppose that fact may be irrelevent though.
Any suggestions are requested.
It just pains me to be so very close, yet so far...
Re: Magnetron- 4J50
Shaun, Wed Jun 20 2007, 11:18PM
You can just series-parallel existing diodes to get what you need. For example, digikey's 1N4007-E3 is 1000V and 1A, so 3 parallel strings of 50 in series would be equivalent to a 3A 50kV diode. Normally I would say make the diode have 1.5x rating for the applied voltage, eg. a 15kV diode for a 10kV power supply, but the schematic seems to have already done this.
As for rectifying the NST, you could use the same 1N4007-E3 diode to make a rectifyer, just 32 in series for each leg of the rectifier.
Now that adds up to 278 diodes, which on digikey will run you about $20 US. Not a whole lot.
If you meant making your own diodes from scratch, im pretty sure thats impossible.
Shaun, Wed Jun 20 2007, 11:18PM
You can just series-parallel existing diodes to get what you need. For example, digikey's 1N4007-E3 is 1000V and 1A, so 3 parallel strings of 50 in series would be equivalent to a 3A 50kV diode. Normally I would say make the diode have 1.5x rating for the applied voltage, eg. a 15kV diode for a 10kV power supply, but the schematic seems to have already done this.
As for rectifying the NST, you could use the same 1N4007-E3 diode to make a rectifyer, just 32 in series for each leg of the rectifier.
Now that adds up to 278 diodes, which on digikey will run you about $20 US. Not a whole lot.
If you meant making your own diodes from scratch, im pretty sure thats impossible.
Re: Magnetron- 4J50
King_of_Junk, Thu Jun 21 2007, 02:11AM
Well, you know, there are things like mecanical rectifyers and capacitive rectification, so I thought I might see if there wasn't some sort of a recipie for this.
I'd heard if you get a piece of one pure metal and put it against a piece of another pure metal, you sometimes get a diode... but that's only what I'd heard, far from a recipie for a high voltage diode.
And I've subsequently obtained a 4j52a magnetron and a 15kVDC power supply (what a combination!) which might be repaired with little difficulty. I think I might try that as opposed to something which is going to be a huge pain to build.
King_of_Junk, Thu Jun 21 2007, 02:11AM
Shaun wrote ...
If you meant making your own diodes from scratch, im pretty sure thats impossible.
If you meant making your own diodes from scratch, im pretty sure thats impossible.
Well, you know, there are things like mecanical rectifyers and capacitive rectification, so I thought I might see if there wasn't some sort of a recipie for this.
I'd heard if you get a piece of one pure metal and put it against a piece of another pure metal, you sometimes get a diode... but that's only what I'd heard, far from a recipie for a high voltage diode.
And I've subsequently obtained a 4j52a magnetron and a 15kVDC power supply (what a combination!) which might be repaired with little difficulty. I think I might try that as opposed to something which is going to be a huge pain to build.
Re: Magnetron- 4J50
Shaun, Thu Jun 21 2007, 04:48AM
Well what I suggested is kind of a huge pain to build, soldering hundreds of tiny diodes together (It's good soldering practice for me tho, lol). It is proven tho, which around the hv community is always a huge plus.
And of course there are other ways to convert AC to DC, like capacitive+mechanical like you metioned. As for putting two metals together, I seriously doubt metal of semiconductor-grade purity can be found (99.999999% pure!).
I guess it doesn't matter at this point tho, since apparently you already have a supply.
Goodluck using the magnetron, I would probably be scared half to death of the thing.
Shaun, Thu Jun 21 2007, 04:48AM
Well what I suggested is kind of a huge pain to build, soldering hundreds of tiny diodes together (It's good soldering practice for me tho, lol). It is proven tho, which around the hv community is always a huge plus.
And of course there are other ways to convert AC to DC, like capacitive+mechanical like you metioned. As for putting two metals together, I seriously doubt metal of semiconductor-grade purity can be found (99.999999% pure!).
I guess it doesn't matter at this point tho, since apparently you already have a supply.
Goodluck using the magnetron, I would probably be scared half to death of the thing.
Re: Magnetron- 4J50
Steve Conner, Thu Jun 21 2007, 07:02PM
A magnetron only conducts in one direction, rather like that other well known electronic component that only conducts in one direction.
Steve Conner, Thu Jun 21 2007, 07:02PM
A magnetron only conducts in one direction, rather like that other well known electronic component that only conducts in one direction.
Re: Magnetron- 4J50
GreySoul, Thu Jun 21 2007, 07:34PM
If you've got access to microwave parts you should have access to the diodes from microwaves.
Also, you can get 1N4007's from ebay for about $10/1000 pretty reliably.
Of course you can also get 30KV diodes on ebay for about $10 each, but they're usually shipped from China and that takes about 3 weeks.
On the subject I hope, how would one go about building a diode?
-Doug
GreySoul, Thu Jun 21 2007, 07:34PM
If you've got access to microwave parts you should have access to the diodes from microwaves.
Also, you can get 1N4007's from ebay for about $10/1000 pretty reliably.
Of course you can also get 30KV diodes on ebay for about $10 each, but they're usually shipped from China and that takes about 3 weeks.
On the subject I hope, how would one go about building a diode?
-Doug
Re: Magnetron- 4J50
King_of_Junk, Sat Jun 23 2007, 07:01PM
Yes, but it doesn't really help to waste half the cycle with applying the positive to the filament, now does it?
The whole purpouse is to power the magnetron as efficiently as possible. And I wouldn't have anywhere near the voltage I need for the 4j50 without the use of diodes in some kind of doubler circuit.
Access to magnetrons? They are surplus from this warehouse that few people know about.. I really got it for free just about (I trade work for parts).
King_of_Junk, Sat Jun 23 2007, 07:01PM
Steve Conner wrote ...
A magnetron only conducts in one direction, rather like that other well known electronic component that only conducts in one direction.
A magnetron only conducts in one direction, rather like that other well known electronic component that only conducts in one direction.
Yes, but it doesn't really help to waste half the cycle with applying the positive to the filament, now does it?
The whole purpouse is to power the magnetron as efficiently as possible. And I wouldn't have anywhere near the voltage I need for the 4j50 without the use of diodes in some kind of doubler circuit.
Access to magnetrons? They are surplus from this warehouse that few people know about.. I really got it for free just about (I trade work for parts).
Re: Magnetron- 4J50
Marko, Sat Jun 23 2007, 08:07PM
Steve: the featured magnetron is apparently of pulsed type and so requires some kind of capacitor charging supply. I don't know how would that one work in CW.
It requires a large 100 joule cap (and apparently it can go to 250 joules if 250kW/1ms figure is correct) and a pulse forming inductor in order to acheive required pulse length.
Regarding diodes, simply series a big bunch of 1N4007's. If there's enough of them you probably won't need anything more although some people use equalizing resistors in paralell (expensive) to reduce effect of unbalanced junction capacitances.
Marko, Sat Jun 23 2007, 08:07PM
Steve Conner wrote ...
A magnetron only conducts in one direction, rather like that other well known electronic component that only conducts in one direction.
A magnetron only conducts in one direction, rather like that other well known electronic component that only conducts in one direction.
Steve: the featured magnetron is apparently of pulsed type and so requires some kind of capacitor charging supply. I don't know how would that one work in CW.
It requires a large 100 joule cap (and apparently it can go to 250 joules if 250kW/1ms figure is correct) and a pulse forming inductor in order to acheive required pulse length.
Regarding diodes, simply series a big bunch of 1N4007's. If there's enough of them you probably won't need anything more although some people use equalizing resistors in paralell (expensive) to reduce effect of unbalanced junction capacitances.
Re: Magnetron- 4J50
Steve Conner, Sun Jun 24 2007, 08:11AM
I reckon Firkragg is right, it probably won't do much in CW. The current will be too low to get it going.
That other half cycle isn't "wasted"- the tube just doesn't do anything. I've seen people run tube RF generators this way to avoid the expense of a HV rectifier. It's not optimal because it passes a lot of DC through the transformer which can saturate it.
Steve Conner, Sun Jun 24 2007, 08:11AM
I reckon Firkragg is right, it probably won't do much in CW. The current will be too low to get it going.
That other half cycle isn't "wasted"- the tube just doesn't do anything. I've seen people run tube RF generators this way to avoid the expense of a HV rectifier. It's not optimal because it passes a lot of DC through the transformer which can saturate it.
Print this page