Protection circuit design

Everything technical about radio can be discussed here, whether it's transmitting or receiving. Guides, charts, diagrams, etc. are all welcome.
Post Reply
nrgkits.nz
Neckmin
Neckmin
Posts: 318
Joined: Fri Oct 17, 2014 10:35 am

Protection circuit design

Post by nrgkits.nz » Wed Apr 22, 2020 12:29 pm

Here is my protection circuit, this is still a work in progress and i'm current adding to it.

Any commends or suggestions would be greatly appreciated.

I've got as far as etching a prototype board and testing on a spare MRF186 PA, it worked very well and switched immediately on detecting a high amount of reverse power - from me unscrewing the connector into load. The output device survived and still worked!

The temperature and reverse power inputs take upto 5v. VR1 and VR2 set the tripping point. An ordinary LM35DT can be used to get a temperature reading - I used one and bolted it to the PA heatsink. The reverse power will require a coupler, or you can use the coupler on the output of the Chinese PA board. C9 47uF acts as a "power on reset" and ensures both flip flops are cleared/reset on startup.

The PA bias output goes straight to the wiper of the variable resistor in the bias network on PA, when either the SWR or temperature trips, the PA bias gets taken to ground through the 2N7002. Another add-on circuit i'm still working on uses PIN diodes to switch the PA drive into a load also - although grounding the bias on its own also worked quite well with the MRF186, output power went down to about 3 Watts.

If either the reverse power or temperature goes above the tripping point, you'll get the relevant LED light up. You'll then need to hit the reset button or power cycle the board.

For the SWR tripping point I set it at around 1.8:1 to 2.0:1 (10% power being reflected). The easiest way to set the SWR tripping point is to reduce the output power of the PA right down to 10% of the full power output (10W for the 100W MRF186 PA) and then measure the voltage on the forward side of the coupler, then set VR2 on the protection circuit so that you get the same voltage reference into the LM393. This will work with other couplers, providing the forward side of coupler is identical to the reflcted side, if not then you'll need switch the connectors around (PA output to coupler output, coupler input to the load) then repeat again with output power set to 10% and measure the voltage on reflected side of coupler.
You do not have the required permissions to view the files attached to this post.

radium98
proppa neck!
proppa neck!
Posts: 514
Joined: Fri Aug 26, 2016 7:01 pm

Re: Protection circuit design

Post by radium98 » Wed Apr 22, 2020 8:12 pm

Nice good work :) any picture of the assembled pcb .just few comparator and and a negative response will do .keep it simple .good work good luck

User avatar
thewisepranker
proppa neck!
proppa neck!
Posts: 522
Joined: Fri Aug 22, 2014 12:53 pm
Location: Brighton

Re: Protection circuit design

Post by thewisepranker » Wed Apr 22, 2020 9:16 pm

Q1 and Q2 will never be turned on fully because Vgs needs to be a positive value and exceed the threshold by a nice margin. If it is working as drawn it's probably because you got lucky with choosing a FET with a high Rds(on) and low threshold voltage. Either change them for P-channel devices or put them underneath R15 and R16 respectively.

You should put current limiting resistors on the gates of Q1, Q2 and Q3. Figuring out what value the resistors should be is a bit more difficult than it needs to be because it's often buried in the CD4013 datasheet as a power dissipation in the usual absolute maximum ratings. The limit is 100 mW per transistor inside the CD4013. Some datasheets state 10 mA for inputs and some say the same but for inputs and outputs, with the 100 mW per transistor dissipation still being true. So, why not just go with the 10 mA - you don't need sharp edge transitions anyway.
You've got two FETs per D-type output so the peak current to each gate needs to not exceed 5 mA. Choose the value of your gate resistors such that whatever next nearest value you choose is the higher resistance rather than the lower resistance option.

Put a resistor to ground on the gate of Q3 but not such a low value that you form a significant divider. Q1 and Q2 don't need them.

Is Q3 man enough to short the RF transistor gate capacitance to ground? I don't know what RF transistor you're using but I wouldn't have thought so. You might be lucky because of the high Rds(on) of the 2N7002. If you choose a bigger FET (which I would), perhaps add a few tens of Ohms or so between the gate of the RF transistor and the source of your protection transistor. It'll still get strongly shunted to ground.
I also appreciate you may have just put a 2N7002 in there because it's all you had conveniently to hand in LTSpice.

Omit R10 and R14. It's not a good idea to introduce DC impedance on the rail because it can cause havoc with your thresholds as the comparators are loaded and delivering current, and could worst case cause them to chatter. You want a low impedance connection to the power rail and if you've got noise problems, add inductance not resistance. I assume you've added them to improve the effectiveness of the shunt C, but it could cause you bigger headaches. The inductive reactance of your trace will be more than 10 R at 100 MHz anyway.
You've drawn two sets of decoupling filters, C4, R10 and C8, R14. You only need one per LM393 because it's a dual comparator, so what you could do with that spare capacitor (C8) is fit it in parallel with C4 and choose a value one or two decades away from C4. So 100 pF and 10 nF, for example. Or 100 pF and 1 nF. Choose a value other than 1, 10, 100, etc. like 0.82 nF or something. They're cheaper and easier to get hold of. There's nothing special about 1 nF or 100 pF, but they're used everywhere for this purpose, for no reason other than being a round number.
Make sure the decoupling capacitors are as close to the supply rail of the comparator as possible, SMD preferred, and don't forget about the path length between the other end of the capacitor and the GND terminal of the comparator.

nrgkits.nz
Neckmin
Neckmin
Posts: 318
Joined: Fri Oct 17, 2014 10:35 am

Re: Protection circuit design

Post by nrgkits.nz » Thu Apr 23, 2020 1:50 pm

Thanks for that, some great advice there. The circuit was quickly put together with existing parts I have available and seemed to work fine despite the issues as you’ve listed - but it would be good to correct all of that anyway - you’ve made some very good points, so I now need to update my schematic and do another prototype. I’ll post a photo tomorrow of the existing prototype I have.

radium98
proppa neck!
proppa neck!
Posts: 514
Joined: Fri Aug 26, 2016 7:01 pm

Re: Protection circuit design

Post by radium98 » Fri Oct 16, 2020 8:25 pm

I’ll post a photo tomorrow of the existing prototype I have ????????

rigmo
no manz can test innit
no manz can test innit
Posts: 158
Joined: Wed Jan 30, 2019 9:35 pm

Re: Protection circuit design

Post by rigmo » Tue Oct 20, 2020 10:55 pm


rigmo
no manz can test innit
no manz can test innit
Posts: 158
Joined: Wed Jan 30, 2019 9:35 pm

Re: Protection circuit design

Post by rigmo » Tue Oct 20, 2020 11:20 pm

DRAW :)
You do not have the required permissions to view the files attached to this post.

rigmo
no manz can test innit
no manz can test innit
Posts: 158
Joined: Wed Jan 30, 2019 9:35 pm

Re: Protection circuit design

Post by rigmo » Tue Oct 20, 2020 11:26 pm

USE S50085A - lm358
You do not have the required permissions to view the files attached to this post.

rigmo
no manz can test innit
no manz can test innit
Posts: 158
Joined: Wed Jan 30, 2019 9:35 pm

Re: Protection circuit design

Post by rigmo » Tue Oct 20, 2020 11:31 pm


rigmo
no manz can test innit
no manz can test innit
Posts: 158
Joined: Wed Jan 30, 2019 9:35 pm

Re: Protection circuit design

Post by rigmo » Tue Oct 20, 2020 11:38 pm

hi current
You do not have the required permissions to view the files attached to this post.

User avatar
sinus trouble
proppa neck!
proppa neck!
Posts: 573
Joined: Fri Aug 22, 2014 11:34 pm

Re: Protection circuit design

Post by sinus trouble » Wed Oct 21, 2020 1:09 am

Some nice designs! :)

The FET switching versions are superior in terms of speed and reliability over the relay type.

Ive always found it hard to get the reset conditions right? A hard reset via push button or power down to reset is the safe option, However it is manual!

On the other hand? The 'Bouncing effect' can occur with an automatic reset! Its a tough problem i have not found a solution to so far.
I am as stupid as I look! :|

User avatar
sinus trouble
proppa neck!
proppa neck!
Posts: 573
Joined: Fri Aug 22, 2014 11:34 pm

Re: Protection circuit design

Post by sinus trouble » Wed Oct 21, 2020 1:43 am

Just To Add!

Device temperature is an important factor to consider? Maybe a cooling fan may fail? SWR would be irrelevant!

Short circuit will certainly be overlooked in a matter of Micro seconds!

Could using a device well within its parameters be the simple alternative?
I am as stupid as I look! :|

User avatar
thewisepranker
proppa neck!
proppa neck!
Posts: 522
Joined: Fri Aug 22, 2014 12:53 pm
Location: Brighton

Re: Protection circuit design

Post by thewisepranker » Wed Oct 21, 2020 2:31 am

sinus trouble wrote:
Wed Oct 21, 2020 1:09 am
On the other hand? The 'Bouncing effect' can occur with an automatic reset! Its a tough problem i have not found a solution to so far.
An automatic reset that hasn't been thought through properly will oscillate because that's most likely what it has (inadvertently) been designed to do. When the reflected power exceeds the threshold, the protection circuit cuts the power to the amplifier. Great, that's the end of the matter, right? Well, as a result of cutting the power, the reflected power no longer exceeds the threshold because there cannot be any, so the protection circuit does the only other thing it can do - turn the power back on.

A common approach is to add positive feedback around the comparator to introduce hysteresis, but doing so isn't ideal because the feedback system can be temperature sensitive, so when your transmitter cools down after being "on" whilst running without the PA doing anything for a while, it can re-trigger. You could increase the amount of hysteresis but there are also other ways to achieve the same goal.
I've built circuits like this before and will post up the info when I get round to it.

rigmo
no manz can test innit
no manz can test innit
Posts: 158
Joined: Wed Jan 30, 2019 9:35 pm

Re: Protection circuit design

Post by rigmo » Wed Oct 21, 2020 3:42 pm

I do not recommended automatic rese for named reasons and few anthers basically no... only if you have remote monitoring and can see all important values.. voltage swt temperature fan speed etc....
the best ragged temperature protection is temperature switch .... KSD3-9 NC
You do not have the required permissions to view the files attached to this post.

User avatar
sinus trouble
proppa neck!
proppa neck!
Posts: 573
Joined: Fri Aug 22, 2014 11:34 pm

Re: Protection circuit design

Post by sinus trouble » Thu Oct 22, 2020 1:53 am

Its a tricky situation, But not impossible to automate?

I have a theory? As we all know a comparator compares two inputs then outputs accordingly! Yet i dont see any designs using the forward power as a reference respective to the reflected power? They all seem to use a DC reference threshold to trip the protection?

Say hypothetically the PA went into protect mode but still gave a safe level of power to the comparator for it to do its thing until its safe to power up again?

Ofcourse a serious antenna fault would need to be addressed but it may prevent false alarms.

Again just a theory, I could be wrong?
I am as stupid as I look! :|

Albert H
proppa neck!
proppa neck!
Posts: 1710
Joined: Tue Apr 05, 2016 1:23 am

Re: Protection circuit design

Post by Albert H » Thu Oct 22, 2020 2:36 am

I've spent months designing protection systems for transmitters - often for pairs of transmitters for 100% redundancy. It's truly one of the "dark arts" and I'm unwilling to share much of my work - after all, I have to have some commercial advantage!

I have a telemetry system that introduces an extra subcarrier into the FM signal, and sends various pieces of data along with the broadcast signal. PA temperature, VSWR, Supply Voltages, ambient temperature and humidity are all sent. Our link system includes facilities for remote switching of equipment - main and reserve aerials, PAs, drivers, exciters, even power supplies! We can even change power output in some cases. There is a microcontroller at the output site that will "decide" what the system is going to do, and this works well when the source site is unattended. However, the various controls can be manually operated from the studio end, putting a human in charge of the decision - making.

It's a truly difficult thing to get right!
"Why is my rig humming?"
"Because it doesn't know the words!"
;)

Post Reply