Safety concern

[turtle]

Hallo. It is decades now since I did anything more complex than wiring up an Arduino, yet I find that I am now contemplating building a 1000w, 300v RF generator. One of my hobbies is making things from coated fabrics, which I join with an ultrasonic welder - but there are certain issues which I believe would be better solved with an RF welder.  It needs to run at 27.12MHz, and I have found https://docplayer.net/36509181-Application-note-low-cost-1000-watt-300-volt-rf-power-amplifier-for-27-12mhz-apt9701.html this. It runs with a 300v input, so I imagine the output at the electrodes is also 300V at 6A. In use my hands would be very near the electrodes and inevitably would touch one or both together. Until I know better my working assumption is that this would be lethal.

Could someone with actual knowledge tell me just how risky this might be?

Against that assumption I have two data points:
1/ while calibrating my ultrasonic welder I manage to grip the transducer where the two wires connect into it. As far as I can find, eg from https://www.bbt-medical.com/what-is-the-working-voltage-of-the-ultrasonic-transducer.html, the input voltage for a 1500W transducer (mine is 3,500w) is 1000v to 2000v at 2amps. This is AC at the sonic frequency, ie 20KHz. It gave me a hell of a start and my burnt finger took weeks to heal, but  failed to kill me.

2/ "wood welders" are commonly used in joinery shops to set glue, and are RF generators at 27.12MHz and 3000 or 4000w. Brands are 'workrite', 'tregarne', 'lamont', and all seem to be salvaged from the scrapheap left behind when Noah set sail. These have handheld guns with both electrodes underneath which you press down across the glue line, much as you press a domestic iron down. It seems impossible not to have incidents of people touching both electrodes, yet there seem to be no reports of injury.

While I'm happy to learn a shed load of new stuff, I would be most unhappy to get to the end and find I've built something too dangerous to use. Hopefully someone can clarify this point before I start.

Cheers.

[plasma]

Hi Turtle

Good on you for doing due diligent. I'm probably not the one to ask about leathly, but I think the RF burns would be wicked.

A different tack, how would you make it safe or safer to operate?

[Twospoons]

• To kill, the electricity needs to run through your heart - either arm to arm or arm to foot
• At 27MHz the current will flow mostly over the outside of your body (skin effect, and thats science, not a pun), and therefore will miss your heart
• The bigger risk is a device failure, resulting in 300VDC on the output, but there are blocking capacitors there that will stop that - assuming they have sufficient voltage rating
• The blocking capacitors will also break the loop to mains earth - little current will flow at mains frequencies
• For extra safety use an isolated 300VDC source, rather than rely on the blocking caps
• RF burns are still going to hurt - but they wont kill you

[johnnyzoo]

As far as I know, RF current through your body parts shouldn't cause heart malfunction or cramps like low-frequency AC or DC does. Some people say it can be almost painless (I haven't tried). You can find Youtube videos where people draw high frequency arcs using their fingers, which is obviously not recommended but it illustrates the properties of higher frequencies.

However, at higher power levels it will cause nasty burns, and I believe nerve damage is also possible in the long run, either due to burns or other effects.

Another danger may arise if the RF source is not galvanically isolated from the mains AC or DC supply.

Avoiding any contact with the energized parts is a good idea of course.


edit: Too slow, Twospoons was quicker with his reply. Mostly same things as what I wrote.

[Kevindk9]

I've been bit by a ~500 watt class e tesla coil flame at 15mhz and it hurt, but there was no real damage.

it only touched my finger for about 1/10 of a second and it seared the outer layer of my skin, turning it white, but not enough heat was transferred for a blister.

[davekni]

I've been bit by a ~500 watt class e tesla coil flame at 15mhz and it hurt, but there was no real damage.

Yes, burns are the big concern.  RF welder may be worse for given power than class-E Tesla coil.  Welder voltage will be lower than a class-E coil, but current will be higher, probably making burn worse.

Microwaves are particularly bad for eyes, where there is relatively little blood flow to carry away generated heat.  Anyone know if 27MHz near-field is a significant issue for eye heating?  I've wondered about that running my 13.56MHz class-E Tesla coil.

[turtle]

Many thanks to all. The burn off my ultrasonic transducer was quite bad so I'll definitely want to avoid that from RF. The electrodes will be small and will have guards, but of necessity will still be exposed. But if it's unlikely to be fatal, barring faults in the system, then I'll go ahead and start gathering the components. This https://www.microsemi.com/document-portal/doc_view/123626-drf1200-27-12mhz-reference-design-kit includes an oscillator circuit which I plan to borrow, and the 2000W, 300VDC power supply may be simplest to buy off Aliexpress. If I need more advice - or when, as I plainly have much to learn - I suppose the "Radio Frequency" forum would be the one to post to.

Again, thanks and cheers.

[Da_Stier]

Hi turtle,

I guess most things have already been said, but I have a few more thoughts to add.

- did you manage to find the ARF446 and ARF447 transistors anywhere, they seem to be obsolete and I couldn't find them
  (I only took a very brief look however)
- At 1000W, even frequencies in the HF band (which your 27MHz are) are quite harmful if radiated
  It heats up internal organs and as davekni said especally your eyes
  (which can blind you pretty fast in an extreme case)
- keep in mind that the amplifier you posted is matched for a 50 Ohm output
  If you don't match the rest of your system, the high reflected power will destroy your amp in seconds.
  it will also radiate like crazy and get all the HAMs and the FCC pretty angry at you 
- 1000W in a 50 Ohm load will generate around 224V, which will still hurt but this has already been discussed
- the CB frequency range is pretty close to what you want so you might be able to take some inspiration from a CB amp
  (or even use one to be honest)

Over all I don't want to discourage you, keep going, this is a very interesting topic. 


Greetings,
Michael

[Twospoons]

Ideally you'd want to put a circulator/isolator between your power amp and your work head, so that the inevitable reflected power can be dumped into a dummy load.  Not sure where you'd get one that can handle 1kW, and it probably wont be cheap.

[Da_Stier]

Twospoons is right however keep in mind that a circulators size is directly dependent on the frequency of operation.

For 27MHz it will be around the size of a pizza box or something.

[turtle]

For those two mosfets, I found https://www.transistormosfet.com/transistor?search=arf446 and https://www.jotrin.com/product/list?keyword=arf446, but as you say these are obsolete and so these sources are either old stock or fake. Maybe both. That possibility slows me down a bit.

My real problem here is that, while I know the end result I want, I really don't know how to get there and am hunting for the few items that are public. A small diy RF welder is probably one of those things that are never made public. Your idea about CB amps is interesting - there seem to be some available of around 100W to 300W, though these are probably peak ratings. Perhaps I could make a decent benchmark by getting one of those with a signal source, and sizing the electrodes so as to maintain a reasonable energy density. I'm only guessing at the 1000W figure, based on the fact that my small ultrasonic welder at 800W is too small, while the 3500W one runs at 30 to 50%. All the large commercial RF welders out there have large electrodes and correspondingly large power ratings. Mine will only ever be small so it's possible that the power could be well under 1000W.

[turtle]

I have now found several suitable RF amps at 1KW and more, in particular there is a chap in Australia who makes them to order and sells them on ebay. Seems rather more trustworthy than other sources. But what I will do first is get something at around 200W, learn how to manage issues with it, and set a marker on the energy density and power that will be required.

It will take a while, possibly weeks even, but I will report back to the 'Radio Frequency' subforum for those interested.

[Da_Stier]

Hi turtle,

glad to hear that you found other options.
It is always handy to have a backup plan or two.
To start somewhat small is always a good plan.

Something like this is a good starting point for trying different things.
https://de.aliexpress.com/item/32846358133.html
These amps use 900MHz GSM transistors, that turn out to be pretty usable in the HF band.
These are however pretty easy to kill with mismatched loads.

One more thing to give you on your way is that you should conisder building a SWR bridge.
Like this you can always monitor the forward and reflected power and shut the amp down (or limit the power) if the SWR gets to bad.

The CB amplifier guys are a "special kind of people".
As far as I know there is no place where it is actually legal to use a CB radio with a 10kW amplifier. 

MR BBI is an example of a youtube channel that covers a lot of high power CB amplifier builds.
https://www.youtube.com/user/wickeddreams911/videos


Good luck and looking forward to see what you come up with. 

EDIT:
What I also forgot to mention.
There are a few transistors that are used in basically every HAM radio amp.
One of them is the BLF188XR.
You get over a kW per transistor and they are pretty rugged and hard to kill.
There are also a lot of designs available for it.
This would be a good 3kW unit:
https://www.ebay.de/itm/222707564979
EDIT end

Greetings,
Michael

[turtle]

Many thanks for that Da_Stier. That 3000w job looks a real beast! I thought about getting something very small and "disposable", but in the end got this https://www.ebay.com/itm/164051111298?hash=item2632354d82:g:naoAAOSwuvBgQPEp at 600w with 2 MRF300s - still cheap enough to not be a loss if I destroy it, but with enough power that it has a chance of actually doing the job, and certainly of getting benchmarks. Now I need to figure out how to setup the output load to electrodes rather than an antenna, and then something for a controllable RF input. It's going to be close to Christmas before all the bits arrive and I start getting them setup.

Best regards

[Twospoons]

Do you have VNA? Something like this: http://miniradiosolutions.com/minivna-pro/
I have one - very useful for measuring the impedance of your work electrodes so you can build and test a suitable matching network without blowing up your RF amp.
Works as an RF generator too.

[Da_Stier]

That is an excellent point Twospoons. 

@turtle
If you don't have one and have never used one before, a nano VNA might be a good choice. Mainly because they are super cheap and have a gigantic community with very good tutorials for the specific device.

The minivna that Twospoons suggested is a very good device as well, if you prefer a slightly more expensive USB based device.

A VNA will enable you to look at all your impedances at different points and visually see the effect of changes in real time.
It also is VERY helpful if you need to tune something or design a matching network.


Greetings,
Michael