Sinus Rig Repair

[Albert H]

Agreed again. Another possibility, with a three legged / fixed regulator, is to put a Zener diode in series with (what's normally) the ground leg. This will raise the output voltage by the value of the Zener diode, and raises the maximum input voltage by the same. It does however (obviously) mean that the device needs to be insulated from the case. So you could use a three legged (eg 78 series) regulator, treated in this way, as a pre-regulator; for subsequent three legged regulators wired in the normal way.

That would probably be my preferred solution, but bear in mind that the OP said that the supply was "about" 40V - I would expect that to rise a bit when the PA is off.

The last time I needed to do something like this, I ended up winding (by hand!) an extra low voltage secondary to the toroid. It didn't take too long, and gave me a pair of auxiliary supplies - one at 8V when rectified (to get regulated down to 5V for some logic) and about 24V AC which was rectified and regulated with a 7824 for the fan. This was probably the most efficient way of achieving the range of supplies I needed in the rig.

[sinus trouble]

Thanks again to you all!
Earlier today I ordered some LM317HVs, Silicon insulating pads and M3 screw insulating bushes!
The HVs are a bit extreme for what I need but should give me some comfortable headroom!

Mr shed, I did not realize a zener could be used with the fixed regulators?? I learn something new every day! lol

Mr Albert! Now that you've mentioned the loaded and unloaded supply voltage, it all makes sense! I always disconnect the PA before retuning the driver! In turn contributing to the problem further!

Anyways I will post some pics and results once I receive the parts!

[sinus trouble]

Hello Necks!
Just knocked up the new regulator circuit and initial tests are looking good! Each section gives me a smooth adjustment from 1.2v to 27v allowing me to optimize the voltage!

20170322_212843.png

All regulators are insulated and not showing any signs of thermal problems, the trimmers are good quality and sealed from Cr*p getting in!
I may check for any ripple on the scope and add caps if necessary? But no strange DC noise apparent so far!

[Albert H]

Looks like you've cracked it. I just hope that you don't exceed the input voltage for the ICs when the PA is off. I've made that (expensive and annoying) mistake!

[sinus trouble]

Cheers Mr Albert!
According to the Datasheet, the 'HV' types are rated 60v max? But we shall see?? lol
I just hope in the event of failure they don't send 40 odd volts through the exciter!!
That will be catastrophic!!

[MC Spanner]

@Sinus Trouble, a very tidy build! How much power out do you get with your 1W drive? I would usually expect about 100W from this device. But you're running at 40V, if I've understood it correctly. What's your bias current?

[radium98]

@MC Spanner surely he dont run exciter 1w board at 40v and he must expect a well tuned up to 600-800mw depending on what type of transistor used
4427
3866
@sinus trouble what about testing output with 1947 ?to see what do u have

[sinus trouble]

Hello Mr Spanner!
Thankyou for the compliment on my work! As you can see the rig is not designed for professional/pirate use!
It does 50watts nicely which is great for my needs!
However I wish I did not build this PA! the hairpin in the output was a nightmare to get right and there is something wrong with the bias? The Idq doesn't reach anywhere near the 100ma requirement? Not even 50ma!!
@ a gate threshold voltage of 2.5v it is only drawing around 3ma? Any higher and I risk damage!

Maybe the circuit works perfect with more RF drive? (8watt as designed) I have built the BLF177 test circuit a couple of times before and obtained 90watts @ 40v

@ Mr Radium
I do have a 5watt version of the driver board which I did contemplate fitting into this rig?
With simpler No tune PAs available I don't think this old rig/PA is worth spending anymore time and money on!

[sinus trouble]

Sinus BLF177 Old.png

[MC Spanner]

Hi Sinus, you may not care about any of this so feel free to ignore and I'm sure you know when biasing a FET that current is king so I won't insult you by going on about it. Your FET is clearly working so it should be possible to bias it to the correct Idq. So check the usual suspects... reference voltage source (Zener or regulator), check for DC loading at the gate, (RF path should be DC blocked from the bias) check your input match in particular those yellow trimmers.

Running with the recommended 50V and 100mA Idq it's quite easy to get 150W out of this device with no more than 3W. 8W would be overkill. I wouldn't recommend it even at 40V. Sort this bias problem out first.

Hairpins are usually a pain in the arse. It's probably just the nature of the beast but if you're using someone else's design for 150W at 50V your output match may have needed tweaking for your different operating conditions. You've dropped both the power and the voltage though, so perhaps not!

[sinus trouble]

No No! I really appreciate any advice! Although I have a grasp of the theory, its easy for something simple to be overlooked!
I have done a few basic tests and calculations regarding my bias circuit, it seems the 78l05 is providing a nice 5v across the preset pot but lacking a small amount of current to nudge the FET into conduction!
If I do get a response from the FET? There is an issue ive been meaning to sort out for a long time, the low pass filter! The coils are 1mm copper and could be contributing to losses, they should be at least 1.5mm
Also caps whilst fine at 50w need an upgrade!
Worth solving and shouldn't break the bank balance in the process!

[MC Spanner]

Yes I can see you designed the rig knowing that you wouldn't be getting 150W out of it. At 50W those coils look fine but as you say a bit skinny if you turn the power up. You may have to tweak thicker ones slightly to ensure the same filter characteristic. As for your bias circuit, if your regulator isn't overloaded and you haven't got an absurdly high resistance between the pot and the gate and it's controlling the gate voltage as you would expect, then I'd guess your low supply voltage means you need more on the gate to get the channel to conduct. I would proceed very carefully here, but since you won't get the full 150W at 40V it's academic. If you do want 150W out, then you're going to have to up the supply voltage so this should make things more predictable as far as the bias is concerned. You're also going to need more than 1W drive too.

[sinus trouble]

Cheers Mr Spanner!
I have seen others use a lm7805 for the bias which I think is completely unnecessary, the 78L05 is more than adequate and maintaining 5v nicely!
The potentiometer is 25k which shouldn't be an issue?? In the past ive used a 5k with no obvious side effects??
Also from pot to gate, I have only fitted one 100ohm resistor and it should be two in parallel! needs to be sorted!
Something that I have noticed is that I get a considerable audible 50hz hum on the carrier (no modulation) when I increase the bias?? This leads me to believe that my 10,000uf smoothing cap is underrated?? A slump in DC could be restricting anymore gain??

[MC Spanner]

Well, I'm not sure how badly it will affect the gain but it won't help. You say you're getting 50W out no problem, does that mean no hum when your rig achieves 50W? Assuming your PA is 50% efficient (this will change with various criteria) at 50W out that's 100W in, and at 40V that's 2.5A. I'd say you're doing quite well there as I'd allow 10,000uF smoothing per Amp on a linear power supply for this sort of PA.

Now. You said, you couldn't achieve more than 50mA of Idq (well under the recommendation for this device) when you turn the bias up. But, turning the bias up makes it hum. At that bias level you may have less gain but you won't be biasing it into excessive linearity for your application. Therefore, assuming constant drive, sounds like you should be getting more power out but still conservative on the bias. I'd get your multimeter and power meter and verify all this of course, but it could be that you can achieve more power with your current configuration, and you just need more smoothing to make it all a bit nicer, say another 10,000uF cap on the supply.

By the way, your bias components sound OK, don't forget it will see a high resistance looking into the gate of a FET. You could probably get away with a lower value of pot, lots of designs do, as you say. But don't draw too much current from your 78L05 obviously. And yes, dropping that resistance into the gate won't do any harm.

[thewisepranker]

It's probably because lots of people have the generic TO-220 flavour of "7805" to hand. I've probably got about 100 here.
The 78L05 in its TO-92 package is somewhat redundant nowadays. Who needs a regulator only capable of 100 mA at 5 V, with a maximum of 0.75 W power dissipation at 25 °C? You'll blow it up at 100 mA from 12.5 V, or 50 mA from 25 V.

[Albert H]

You probably won't "blow it up" because the thermal protection in the 78L05 is superb. However, if it's over-run, it'll go into thermal shutdown so you'll get nothing at the output terminal! It's also good practice to fit RF bypass capacitors right at the legs of the chip. Some brands of the device go bonkers if they get tickled by some RF!
I've used them many times to power PLL circuits often with a dropper resistor or higher voltage 78L regulator on the way in. It's far from "obsolete" - it's compact, cheap and reliable. I've used the SM version in my exciter boards.

[sinus trouble]

Yes I believe you are both correct in respect to regulators! to-220s I would imagine are far more common to general electronic enthusiasts! and the 78LXX are great for things like PICs Etc...

Mr Spanner, there is deffo something strange going on? I done some clear measurements!

Bias Vgth = 2.45V
DC Supply (PA) = 42.8v @ RF Power out 50watt
Idrain (PA) = 2.03A @ RF Power out 50watt
Idq (PA) = 3.11mA @ No RF drive

The suggestions you made earlier should improve the situation somewhat? The mosfet runs cool and stable at all times, the 50hz noise is quite minor and to be expected on a unregulated supply of this kind!

I will report back on any developments!

[radio-berlin]

Sinus. The hum you are hearing when you increase power is probably coming through the supply rail to your oscillator.
Cut the track the your oscillator stage (or remove resistor) and bang a 9v reg in the there with decoupling caps and some elec caps to smooth. A little 78l09 will work perfectly. If you have copied the NRG layout you will probably have a resistor dropping the voltage from the supply rail to the oscillator, remove this resistor and replace with regulator

[sinus trouble]

Nice 1 Mr Berlin!!!
There is indeed a 47ohm resistor feeding the oscillator and can easily be modified!
If my thinking is on the same wavelength as yours? A modest ripple on the oscillator DC supply could cause a small but significant frequency swing like a constant 50hz modulation??

[Albert H]

Yes it could. Remember - the VCO is very sensitive, and will be affected by crud on the supply. A regulator with close RF bypass capacitors, and an electrolytic smoother (220µ or so) on the output side of the regulator should quieten things down!