Turkish RDVV low power problem

MiXiN · Post by **MiXiN** » Sat Nov 05, 2016 12:05 am

Alright necks,

I have one of those Turkish RDVV transmitter's here, and according to the seller the final MOSFET used can be a RD06HVF1 or the RD15HVF1.

I temporarily fitted a legit' Mitsubishi RD15HFV1 (15W), and at the lower end of the band see only 3W, while at the higher end of the band I see just under 2W. These readings are into a Power meter terminated with a dummy load.

After about 5 to 10 seconds of lock, the 2 green resistor looking inductors start smoking and have now discoloured as a result. I've put a nut & a screw next to the inductors in the images to point them out.

The supply is 13.8V and the Transmitter draws around 2A.

Is there insufficient drive for the RD15HVF1, and please can anyone tell me why the inductors are smoking?

Sent from my SM-G900F using Tapatalk

shuffy · Post by **shuffy** » Sat Nov 05, 2016 12:46 am

The inductors are smoking because they are sourcing more current than they are rated for - I'm not familiar with that design, so don't know exactly where the inductors figure in the circuit but they must be in the PA drain circuit somewhere.

For an RD15HVF1 at 13.8V I'd expect to draw a couple of Amps, or slightly more, for the full 15W out. Because you're only getting a couple of Watts out and drawing 2A, it sounds to me like the gate bias is too high. If there's a way of reducing it easily then set the quiescent drain current to the recommended amount off-drive (top of my head, Idq = 400mA) and then apply the drive and re-test. If you're still getting less power than expected then I'd say you have insufficient drive from the preceding stage or a matching problem.

Since your inductors are smoking with 2A through them and I'd expect to draw at least that for my 15W out, I'd say less power out / RD06HVF1 might be more realistic?

teckniqs · Post by **teckniqs** » Sat Nov 05, 2016 5:37 am

If it was from that same seller then some of the components are most likely wrong around the transistor. RF-Head should have the answers.

Analyser · Post by **Analyser** » Sat Nov 05, 2016 10:38 am

Even if the circuit was working correctly, those inductors look hugely underrated for that power level.

You may just get away with it at 6W but 15W would almost certainly need hand-would inductors.
Look at pictures of the RDVV stuff online and I'm sure most of the inductors are not like those axial leaded ones fitted to the board.

Analyser · Post by **Analyser** » Sat Nov 05, 2016 10:43 am

Hold on, isn't that board designed for a 2SC1971 anyway?

Yep.

http://demo-moutoulos.com/eShop.php?vie ... &product=9

shuffy · Post by **shuffy** » Sat Nov 05, 2016 11:14 am

I was wondering that earlier - something rang a bell. Explains the matching. So, for a FET he will have had to provide DC bias - please tell me he hasn't wired the BD139 into the gate of the FET?

teckniqs · Post by **teckniqs** » Sat Nov 05, 2016 11:37 am

Analyser wrote:Hold on, isn't that board designed for a 2SC1971 anyway?

Yep.

http://demo-moutoulos.com/eShop.php?vie ... &product=9

That's exactly what I was going to say earlier but saw the board said 'FET' for the transistor.

....Can't help noticing how he likes to sell them without the transistors so it's out of his hands once transistor is fitted.

MiXiN · Post by **MiXiN** » Sat Nov 05, 2016 7:14 pm

Sorry I'm late in replying, and thanks for all the replies.

I've ordered some VK200 Ferrites to replace the 2 green inductors that start smoking after a few seconds.

To set the bias, should that be done with the MOSFET out of circuit - and if so, how is it done?

I'm going to substitute the RD15HVF1 for a RD06HVF1 per chance the drive is insufficient for the 15W MOSFET.

Here's an underside view of the PCB, and I'm pointing out the BD139 with a trimmer tool. The Collector is the middle pin. Not sure if this is controlling the gate bias of the final as this is a bit too technical for me.

My friend has the Turkish driver board using the Bipolar 2SC1971 and I recall it looking similar to this circuit, so when I see him tomorrow I'll take an image of that particular board.

Sent from my SM-G900F using Tapatalk

teckniqs · Post by **teckniqs** » Sat Nov 05, 2016 8:03 pm

Is that how the transistor was soldered on when tested because the joints look very dry, have you ran out of solder?

MiXiN · Post by **MiXiN** » Sat Nov 05, 2016 8:11 pm

teckniqs wrote:Is that how the transistor was soldered on when tested because the joints look very dry, have you ran out of solder?

No mate, I removed the RD15HFV1 MOSFET & remaining solder off the pads with some wick ready for when I put a RD06HVF1 in.

The MOSFET isn't actually connected in circuit there, but the image isn't the best. I just put it there to show where it normally goes.

teckniqs · Post by **teckniqs** » Sat Nov 05, 2016 8:15 pm

Fair enough, I couldn't help but notice it.

Analyser · Post by **Analyser** » Sat Nov 05, 2016 9:18 pm

I really have a feeling this board was originally designed for the 2SC1971 and it's been bodged to try and make it work with a FET(s).
Shuffy and others are correct in saying that the matching for those two families of devices are significantly different, especially on the input, plus you have the issue of providing a proper gate bias for the FET which won't appear on a '1971 circuit.
With regards to input power for the final, if it's enough for the '1971 it will be enough for the FET. The issue really is proper input matching and bias. And probably output matching.

Looking at photos of the top and bottom of the board can be useful but personally I find it quite tricky to decipher more complex problems such as this one. The best bet is to email the Turkish fella and see if he's got a circuit diagram. Would make things a lot simpler, IMO.

The Dutch designs that I've seen about seem to fall in to two camps, RDVV and PKBee. The RDVV stuff is a lot older and PKBee seems to update his designs more often. I've seem PKBee boards using the RD06 and the RD15.
RFHead is the man with all the info on this so I'll keep my trap shut until he sees this thread and comments.

Analyser · Post by **Analyser** » Sat Nov 05, 2016 9:21 pm

There is also one slight issue that there seems to be a capacitor missing to the right of the output microstrip. This could possibly be a cause of the fault but makes me wonder why any (2-3W) of power is coming out.

Edit; If the capacitor is going to ground at that point, it might not be needed.

shuffy · Post by **shuffy** » Sun Nov 06, 2016 2:33 am

MiXiN wrote:To set the bias, should that be done with the MOSFET out of circuit - and if so, how is it done?

On FET PAs I usually set the bias voltage (Vgs) to minimum before I install the FET, then install it and ramp the voltage up until I get the minimum recommended quiescent drain current (Idq). This should be done with no RF drive. I can't really tell what's going on from your pictures but it looks like he might be trying to derive a fixed Vgs from somewhere, which isn't the best way of doing it, and isn't consistent with his statement that it will work with an RD06 or an RD15 (despite the fact that the two devices are similar in some respects). I did find a version of this board adapted for a FET which has a small pot controlling the Vgs and it looks like the voltage into that comes from somewhere close to the emitter of the BD139 but I wouldn't swear to it.

https://www.moutoulos.com/eshop/img/p/1 ... efault.jpg

MiXiN wrote:I'm going to substitute the RD15HVF1 for a RD06HVF1 per chance the drive is insufficient for the 15W MOSFET.

I wasn't trying to say that the drive was insufficient for the RD15 - I was trying to say that the drain current resulting from driving an RD15 to 15W would be too much for your chokes. Your VK200 solution is probably the way to go here, but it will depend on you getting the rest of the circuit (bias, matching, drive) right first - you might find you can't actually develop enough drive from the preceding stages to achieve 15W with the RD15 although I'm pretty sure that the original Dutch circuit wouldn't have been far off. By the way, it looks like you have three of those green chokes sourcing the current to the FET, not just the two which are burning.

MiXiN wrote:Here's an underside view of the PCB, and I'm pointing out the BD139 with a trimmer tool. The Collector is the middle pin. Not sure if this is controlling the gate bias of the final as this is a bit too technical for me.

What he used to do with the bipolar PA was feed the collector from the BD139 emitter via a couple of chokes with suitable decoupling. It looks like they're doing this to the driver stage on your board - better than feeding the PA FET (especially with an RD15). What I can't see is how he's biasing the FET.

Try to work out the circuit or get one from the guy. If you can't do this for whatever reason and can't see what he's doing to bias the FET - I think I'd probably disable the VCO or remove a coupling capacitor from one of the early stages and remove the FET. Then check the DC voltage on the gate pad (Vgs) and see if that power level pot changes this voltage. If yes, minimise it and ramp up when you've reinstalled the FET, until you get the correct Idq. If not, what is the Vgs? The RD15 datasheet implies that for the optimum minimum Idq (about 0.5A) the Vgs will be about 2.4V. You certainly don't need much more than this here to begin with. You can measure Idq by lifting one of those melting chokes and putting an ammeter in series in the gap, or get a rough idea by measuring the current drawn from the supply. Make sure the FET is on a good heatsink. Once happy with the bias, I'd change the chokes, reapply the drive and see how much welly I can get out. Then the real fun starts - or it may just be possible that your smokin' chokes are the problem.

Selassie-i, Selassie-i, Galatasaray...

Maximus · Post by **Maximus** » Sun Nov 06, 2016 3:30 am

I think you're on the verge of smoking the original coils. I've seen people use the same parts before with no problems, then for some reason they crank up the bias or something takes off and draws excess current lol

Sent from my iPhone using Tapatalk

Maximus · Post by **Maximus** » Sun Nov 06, 2016 3:34 am

Maybe put a lower voltage into the tx and check the resistors around the final with a multimeter etc. You'll figure it out

Sent from my iPhone using Tapatalk

radium98 · Post by **radium98** » Sun Nov 06, 2016 7:20 am

more problem could be that to drive that fet rd i see some board have 2xbfr96 even to drive the final stage btw old time i did a pcb with original c1972 and i only get 2-3 watts so it was a low drive
and i could assume a fake rd15 for better diagnos i suggest you to put rd06 if it goes up so u have then 2 chances rd15 fake or as i said low input drive thoses are my suggestions
good day

MiXiN · Post by **MiXiN** » Sun Nov 06, 2016 1:07 pm

OK necks, thanks again for all your help.

I've ordered a RD06HVF1 & also some VK200 Ferrites to replace the underrated ones fitted, so I will fit these when they turn up after setting the bias as kindly mentioned by shuffy.

I've been in touch with Polestar (Turkish guy), and he says there's been no complaints and the VK200 isn't critical. He also suggested I might have used a fake RD15HFV1, which I definitely haven't.
As we know, the installed green resistor looking chokes are underrated because they smoke in a matter of seconds - so the VK200 is definitely needed in this instance.

Upon request of a circuit diagram - he told me he copied the circuit from here - http://www.hlektronika.gr/forum/showthr ... 59&page=14

I've directed him to this Forum thread, so hopefully he'll chime in.

polestar · Post by **polestar** » Sun Nov 06, 2016 6:07 pm

hi guys
what is the problem?
mosfet and c1971 design
next project 300mw design

Analyser · Post by **Analyser** » Sun Nov 06, 2016 6:28 pm

polestar wrote:hi guys
what is the problem?
mosfet and c1971 design
next project 300mw design

Where is the bias control for the Mosfet?