Keyboard Shortcuts
ctrl + shift + ? :
Show all keyboard shortcuts
ctrl + g :
Navigate to a group
ctrl + shift + f :
Find
ctrl + / :
Quick actions
esc to dismiss
Likes
Search
QDX - Very low RF output after replacing burned finals
|
After I discovered that QDX can also do CW via Fldigi and MixWin I was trilled to work two QSOes on around 11.0V and PA was giving not more than 4.5W, and 1:01 SWR acording to my SWR/Powermeter. I used to do more qsoes before on different modes with this rig, also wspr with 2w over nights and all with same setup and with no problems at all. However, this time it seamed it was to much for my QDX and I heard a nasty "slap" sound and no CW any more.. Visual check showed one BS170 was broken apart, probably from taking to much heating. RX was ok, also terminal approach works but power output is less than 1w, probably 0.2-0.5 but the QDX board (Rev.5 and f/w. 1.10) is heating to much on TX and also on RX (!) so I left it rest for the time.
So I ordered pack of 10 and replace all four BS170 as replacing only one that was broken didn't repair anything, but also after replacing all 4 finals it is still the same:
- RX works fine, and QDX is ok on COM port recognized and CAT works and PTT all ok. PC is Windws7
- On TX I have very low power, probably less than 1W as it can not wake up my Chinese swr/rf meter that has limit around 1.5 W to show it. I know there is power out as it can be heard on other RX and also get some WSPR responce.
- The board is taking 300 mA on RX, and as I found here, it is twice as it should be..?
- The board is HEATING UP around the legs of two BS170 that are side by side, but not on the opposite pair.
- I also measured 0.33V on the middle leg of the BS170 pair that is not heated, but on the pair that is like heating it is 0V.
- I suspect by subjective finger touching that the IC 9AC3G9K (ACT 38) is also very warm, but it could be that is heated by this heating pair of finals as it is very close to it.
- Done Transmit testing from terminal but got only info that PTT and TX works, and I knew that already..
?
Where to go from here? Help is appreciated. I have only standard DMM, soldering iron, DC from 5 to 20V and cheap frequency counter, also 4-5 more BS170 in stock with some experience of building and troubleshooting couple of QRP kits and some big "guys" also but with older technology... Hopefully will resolve this. Thanks!
?
Srdjan, E78CB?
?
?
? |
|
It looks like the driver IC may be broken. The final of QDX is similar to the one of the QMX. So read please this document from QRP Labs WIKI.
73 Ludwig |
|
Having suspect to IC5 (74ACT08), following advises on the group's archive I have:
?
1. Removed all BS170 - RX current did fell from 300mA to 230mA (should be arround 180mA?).
2. Using only classic DMM I checked Gain voltages from IC5 to the finals to be for RX/TX: 0.0V / 2.5V for pair 1 and 0.6V / 2.2V for pair 2 (should be 0.0/2.5 for all?)?
3. On IC5 I find voltages on pins 3,5,10,13 to be on RX 0, and on TX 3.3V (seams to be ok?)
4. Also removed Q7 and shorted drain and source but nothing changed.
5. I still have IC2 5V regulator to be highly hot but working, and IC5 warm..
6. RX and CAT/PTT/Terminal all works fine
?
Is all this saying that I should replace IC5, as IC2 is still very hot and voltages on finals' gates are not as they should be? Also RX current looks to high..? I did't replace this size of SMT components before, and unit is bought assembled, but with my 20w iron I can try if that is the last resort... any other ideas or comfirmation before going in this direction..?
?
Srdjan, E78CB
?
?
?
?
? |
|
On Thu, Oct 17, 2024 at 08:30 PM, e78cb@... wrote:
2. Using only classic DMM I checked Gain voltages from IC5 to the finals to be for RX/TX: 0.0V / 2.5V for pair 1 and 0.6V / 2.2V for pair 2 (should be 0.0/2.5 for all?)?Yes, 0.6V / 2.2V looks not good. Maybe a broken IC5. ?
Pin 3? I guess you mean pin 2,...
The values are ok.
?
Which voltages at pin 1, 4, 9, 12?
?
Bad, see above.
?
I think it's to weak. You may damage the board.
?
73 Ludwig
? |
|
Pin 3? I guess you mean pin 2,...
The values are ok.
Which voltages at pin 1, 4, 9, 12?
?
// Yes, I meant pin 2, you are right! All voltages are:
RX/TX
1-1,61/1,61V
2-0/3,3 3-0,6/2,2 (to gate 1 of the finals) 4-1,61/1,61 5-0/3,3 6-0,6/2,2 (to gate 2) 7-0/0 8-0,01/2,45 (to gate 3) 9-1,61/1,61 10-0/3,3 11-0,01/2,45 (to gate 4) 12-1,61/1,16 13-0/3,3 14-5,01/5,0 ?
RX current on 12V is 230mA, IC2 regulator very hot
..................
I decide to unpower IC5 by disconnecting it's pin 14 from the board, that is 5V powering pin, and things did changed:
RX current fell from 230mA to 180 mA (finaly!), although IC2 is still hot, may be just a little less.. Measured voltages of IC5 in this state with pin 14 dsiconnected from the board is:
RX/TX
1-1,12/1,17
2-0/2,3 3-0,03/1,02 4-1,12/1,17 5-0/2,3 6-0,03/1,02 7-0/0 8-0/1,43 9-1,12/1,18 10-0/2,3 11-0/1,43 12-1,11/1,18 13-0/2,3 14-1,07/1,43 disconnected from the board This is probably interesting as IC5 has no power on state, but I wouldn't know..
?
However, looks like IC5 gives this higher current use, as powering it off brought the current draw to normal, so I think I will order it for a change. I just hope there is nothing more gone wrong that will burn it again, well we will see..
?
Thank you, Ludwig, for the guidance, if I repair this I will not use cw with more than 2-3 Watts, as 5 is dangerous (!) Probably will put some small fan and bigger metal shell on the BS170 gang's screw...
?
Srdjan, E78CB? ?
?
?
?
|
|
On Fri, Oct 18, 2024 at 06:05 PM, e78cb@... wrote:
if I repair this I will not use cw with more than 2-3 WattsI suggest NEVER do CW with the QDX. I guess the BS170 was damaged by overvoltage. There is no envelope shaping for DIT or DAH. If the PA voltage is switched off hard at the end of DIT or DAH an overvoltage is induced. It's only a question how high. And the question is which BS170 couldn't resist first. Ok, significant lowering the supply voltage reduce the danger. But IMHO it's a time bomb. ?
AFAIK the mentioned spikes should be no problem for digi modes.
?
73 Ludwig |
|
Hi Srdjan,
?
IC5 needs to be replaced.? There are multiple ways to remove the bad IC, with the best being a Hot Air station 9 a good investment IF you plan to service more SMD kits in the future).? Since you are not in the US, here is one I found on AliExpress:
?
I have NOT purchased or used this one.? It is the lowest cost and has the functions I like: both a hot air and soldering iron with digital control.? Buy at your own risk.? There are many out there.
?
The alternatives are to try to cut off the IC's leads with an Exacto knife, carefully remove each lead, and then clean up the pads CARFULLY (I damaged my first board after multiple repairs using this method).
?
The third choice is to purchase a ChipQuik removal kit:
?
I have not used ChipQuik either.
?
You should look on YouTube for directions on the above to decide which is best for you.? The alternative is to ask someone in the group to do the repair and pay the shipping and repair costs.
?
73
Evan
AC9TU
? |
|
To build on what others have said, There is a confirmed spike at the end of a transmission that has been confirmed on the bench.? The best way to protect against this failure is to install a commutating diode across L14, with the banded (cathode) side connected to the +12 volt side of L14.? A 1n4148 is a good choice for the diode.
?
When running digital modes, this spike occurs once in an FT cycle.? In CW, it would be every dit or dah, significantly increasing the chance of failure.
?
73
Evan
AC9TU |
|
On Sat, Oct 19, 2024 at 06:02 AM, Evan Hand wrote:
try to cut off the IC's leads with an Exacto knife, I have found that a #15 (small, curved) scalpel blade is the ideal tool for this.? I put the blade tip on the PCB surface right next to the IC body, then pivot the handle down, like using a paper cutter, to sever the lead.? This method minimizes the chance of slipping and damaging something.
?
I have also used and can recommend ChipQuik and clones.
73, Don N2VGU |
|
Thanks Evan and Don for advice...
?
I will think about hot air station, but not just jet as this is my first time to have to do this. I will try carefully to remove it with classic sharp hot iron and some mini sharp tools, and if I see it goes wrong I can ask my friend to do it as he has more experience with it...?
?
And I will keep out of cw on this one for now, although the idea of adding a diode looks very interesting to think about and it would probably be if keeping it at up to 2w run about it would give some longer use. However, as I have other rigs to play with CW I think I will pay respect to the original idea and design of the QDX unit, and just leave it be as it gave me enough of mind troubling for now :-)
Will order two pieces of IC5 and hope for the best! Thanks again. ?
Srdjan, E78CB |
|
////The best way to protect against this failure is to install a commutation diode across L14, with the banded (cathode) side connected to the +12 volt side of L14.? A 1n4148 is a good choice for the diode.///
?
Evan, what do you think, just how safe or efficient this protection would be? If you saw these spikes on the branch and has anyone tried this mode before? I can hardly put my QDX to run less than 2 Watts by lovering voltage to 7,5V so hard to go lower.. Would that power be safe enough for CW with this modification. It DID survive 5 minutes on 4 Watts on CW first time. May be some additional cooling or bringing these final robust mode that I saw on the group would help? Deffinitly would, but is it only the finals that are endangered with didts and dhats..? Just a thought for a mind...
?
Srdjan E78CB |
|
On 19/10/2024 22:34, e78cb@... via groups.io wrote:
I will keep out of cw on this one for now,Srdjan, Best keep off it all the time with a QDX. On the homepage "QDX is not suitable for on/off keyed modes such as CW because it does not have click-reducing RF envelope shaping" Many (thousands?) of QDX do not have the extra diode across L14 although not actually recommended it will do no harm. 73 Alan G4ZFQ |
|
When running digital modes, this spike occurs once in an FT cycle.? In CW, it would be every dit or dah, significantly increasing the chance of failure.
?
Interesting comment…surely every spike must then do a tiny amount of damage…over a long period of time using digital modes, this must be the same as running CW over a short time?
?
My long running QDX stopped working recently, indicating 2 flashes (suggesting insufficient audio level), but this was misleading as i eventually found it was the finals had failed. I don’t have the diode fitted across L14. Maybe I should, does Hans recommend this?
73 Ken G4apb? |
|
On Sat, Oct 19, 2024 at 06:05 PM, e78cb@... wrote:
?
?
Srdjan,
?
I did this mod or added two Zener diodes on the drains of the BS170s to all of my QDX.? I highly recommend the diode addition.? Both simulation and bench measurements prove the mod removes the spike.? A group of us did a lot of research into the issue when it was discovered by JZ, KJ4A.? You can search the group for JZ's call sign to see the discussion on the issue.? The simulation models should be in the threads.? If not, let me know, and I will post them again.
?
The QDX does not have key shaping capability, so it is NOT designed for CW.? In the digital modes, there is still a "click" at the start and end of transmission but a "slide" between frequencies that does not splatter during the FT transmission.? I would not use the QDX for CW.? Get a QMX or QMX+ that has key shaping capability and much more.
?
The above are my opinions, and they may be in error.? Feedback is always welcome.
73
Evan
AC9TU
? |
|
On 20/10/2024 10:22, KEN G4APB via groups.io wrote:
I don’t have the diode fitted across L14. Maybe I should, does Hans recommend this?Ken, I do not think Hans has agreed, otherwise it will be in the manual. I am assuming that most QDXs sold do not have this mod. I reckon Hans has said it is not likely to do any harm. It has been said that BS170s do not "wear out". When they fail they just fail. Whether it is due to an unusually high spike, who can tell? 73 Alan G4ZFQ |
|
All, Hans has observed this spike in his own bench testing, but his position has been that he doubts there is enough energy present to do harm. He has not posted a mod recommendation as a result of that belief. This is a point on which he and I have had much discussion. I am in serious disagreement with his conclusion and strongly recommend the use of a diode across L14. We each understand the other's position and respectfully agree to disagree. Damage can occur instantly if the spike voltage is high enough. Lots of energy is not required to puncture the FET gate oxide. Lower spike voltages may result in cumulative damage over time as the oxide is hammered again and again, resulting in a pinhole failure ( a 'walking wounded' condition, which grows to a serious gate-drain short, transistor failure, and often damage to the driver chip. JZ KJ4A On Sun, Oct 20, 2024 at 9:39?AM Alan G4ZFQ via <alan4alan=[email protected]> wrote: On 20/10/2024 10:22, KEN G4APB via wrote: |
|
Hi all Zener diode: I'm wary, because I think it may increase capacitances and I'm not sure what?effect that would have on harmonic content, power and efficiency. If it impairs efficiency then it could do more harm than good, if (as I suspect) the most likely failure mode is overheating of the BS170s.? Commutating diode across L14: I wasn't convinced of the necessity?for it, as JZ mentioned. I have done many hours of operation with various bands, QDX, QMX, antennas etc without failures. I did try the commutating diode (1N4148) across L14 and I noted a slight decrease in power output and efficiency at higher bands, but nothing so serious I would discourage it; so it appeared harmless,?I would say. So by all means use the commutating diode if you find it an interesting idea.? On a QMX/QMX+: I don't recommend?EITHER of these things on a QMX-series transceiver because anyhow they have envelope shaping so does not show?that dI/dT spike, even if one can debate it's danger on a QDX; furthermore QMX/QMX+ has SWR protection which should be able to protect against high voltage SWR scenarios. Another important point is that if a diode alters the PA characteristics (which it probably does) then it will likely also alter the phase distortion in?the PA, which won't matter for Digi FSK modes and CW, but it WILL matter for SSB because it will degrade my phase pre-distortion plan.? 73 Hans G0UPL |
|
Thank you for this information. Mike Krieger On Sun, Oct 20, 2024 at 10:47?AM Hans Summers via <hans.summers=[email protected]> wrote:
|
Donald S Brant Jr
to navigate to use esc to dismiss