开云体育

Just built a U3S with relay board and QLG2. 7805 equivalent in TO220 case for 5V power. Was working perfectly for a few day's testing.?

then tx started cutting out a few seconds to half a minute after Tx. Turned out voltage was dropping to 4.90 (measured at input to main board) and processor was resetting. Current draw at this point was 280 mA. ?Non transmit current 140 mA. Power output with single BS170 was 200 mW. Processor never resets when in non transmit mode.?

I have touched up a number of the power solder joints (altho they looked fine before) with no change in behavior. Turning down the drive to BS170 to where output was only 100 mW allowed completion of transmissions.?

powering with known good 5V p/s directly did not solve the problem.?

Any idea where to go from here?

--

73, Dan? NM3A

Try it into a dummy load to see if it’s RFI

This is into a dummy load. All RF connections are well shielded with RG-316/SMA connectors or RG-8X/BNC connectors.?

--

73, Dan? NM3A

Further info:

External power to 7805 equivalent p/s board was mostly from 9V LFP. But it also did it with bench p/s (known good) and 12V LFP. Also tried bench supply directly to U3S and bypassing the 7805 equivalent p/s board. Behavior similar with all of the above.?

--

73, Dan? NM3A

Sounds like the current is limited some how.

Since it appears you have tried different sources, I have to believe the issue is inside the box.

Since it happens on TX, I suggest the PA portion.

? ? Bias setting would be the first check

? ? FET(s) would be next check

Does the power supply you used show current?? Monitoring the current draw in TX would tell you a lot.

-------
Robert Giuliano

KB8RCO

Robert,

Current draw is about 140-160 mA above idle current (140 mA), which is unchanged from when it was first working well. Turning bias all the way down reduces power out to zero and reduces current draw to about 40 mA above idle.?

I agree that it appears to be somewhere 'in the box'. It also appears to be an RFI problem, but why now, when nothing else appears to have changed? I'm still working on it. Next I'll remove the relay board and see if that makes any difference.?

--

73, Dan? NM3A

Hi Dan,

I'm a bit confused about your current draw. The bias current should only just be greater than when the unit is not transmitting, my usual method of setting it is to use a psu with an amp display, note the idle current draw and then increase the bias via the onboard pot until the amps displayed increases by 10mA's or less. Tx current will obviously be greater.

Power out will vary by band and how many bs170's are used, my u3s with 5v pa voltage and bias set as above to 2 bs170's gives 1w on 80m but only 0.1w on 10m.

As Rob says, check the bias setting first.

?

The universe is made up of Protons, Neutrons, Electrons but contains only one M0RON.

Andy,

?I guess I wasn't clear. Let me try again.

?

Total idle current with bias set to minimum is about 138 mA. If I transmit at this setting, there is no power out, but total current goes to about 180 mA. (Maybe due to relay draw?)

?

?I followed the procedure you mentioned to set bias - turned bias pot until total current just starts to rise - about 140 mA total. Transmitting at this setting gives about 200 mW out on all bands (single BS-170, 40-10 m). Total current during transmit at this setting is about 280 mA. At this level, rig only tx for a few seconds; varies from ~5 seconds to rarely a full cycle. Most commonly, the tx cuts out at about 8-15 seconds.?

If I turn down the bias til I only get about 100 mW out or less, Tx usually doesn't cut out. If the bias is at minimum, tx never cuts out.?

When the Tx cuts out, it's because the processor reset at about 4.90 V. It wasn't obvious when the QLG2 is connected, as the time is immediately corrected, but when I took out the QLG2, the time reset to 00:00 whenever the tx stopped.?

--

73, Dan? NM3A

What voltage are you running into the PA?

Same as the processor (5V)?

?

But during my initial setup, whenever I used voltage regulator inside the case, the U3S would reset on TX.??

I tried buck converters and 7805 (TO220).? With the 7805, I kept adding larger CAPs at the output to keep the voltage from dropping on TX, but that didn't help.??

I went to an external 5V supply and the issue stopped.

Later, I was told to add smaller CAPs 0.1 uF on the input and output, in addition to the 100uF (I used) when using the 7805 for radio.

I never tried this as I already was using an external supply that was working.

?

FYI:

But from memory and 1 BS170 and 5V PA, I never got anywhere close to 200mW on 10m.?

I was shooting for 100 mW and after a lot of fiddling with it, I got about 100 mW on 10m.

-------
Robert Giuliano

KB8RCO

Robert,

After a lot of testing, I have come to the same conclusion that you just gave. Turns out the 7805 is overheating and that is what is causing the voltage drop. And the reason I didn't see it in earlier testing was because I was only using 7 or 9 V input. Going to 12V or 13.8 and it shows up. At first I thought it was a bad 7805, but the replacement does nearly the same thing. I am going to try a big heat sink and see if that limits the temperature rise.?

--

73, Dan? NM3A

Measure the volt drop across the 7805 , Work out the power and consult data sheet?

Do you have any capacitors attached to the 7805? Reason I ask is that I had a problem similar to yours with unstable power causing a microcontroller to reset haphazardly. Turned out to be a faulty ceramic it would short when the output current increased beyond a certain level. I only found it by removing it from the circuit, after that no problems, I never replaced it.

Good luck, u3s is quite robust, bound to be a supply issue.?

Andy

--

The universe is made up of Protons, Neutrons, Electrons but contains only one M0RON.

RE: 7805 regulators (all 78xx for that fact) have circuits that will?

cause it to shut down or loose regulation those being over current

and the other being over temperature (to HOT).

?

IT will over heat is the current time the voltage across it exceeds

a very few watts. A heat sink of adequate size will help greatly.

IF it runs ok at 7V (you need more input than output) but it will

heat greatly at 12V.? So as an example at .15A and 2V (7805 at 7V)

you have to deal with .300W (300 milliwatts).? Jack that up to 12V

its now 7V at .150ma or 1.05W, that a lot more heat.

?

The internal circuit will oscillate if the input and output are

not sufficiently bypassed.? I use 10uf at input and 10uf at

output plus .1uf as well.? ?I?e seen poorly bypassed 78xx parts

oscillate vigorously.

?

--

Allison
------------------
Post online only,?
direct email will go to a bit bucket.

Andy, Bruce,

?

Fixed!

?

Yup, it's a supply problem. 7805 was overheating and cutting out. I put a big heatsink on it and it is working FB now. Peak temperature (IR gun reading) is 160 F (~70 C) after long Tx. Idle temp is 130 F. This was using 14.5 V input. Generally going to be less input V, but tolerable for now. (See pic for one sequence of 40-10 m. Search for AC3IE if you want to check it out.) I probably will put some dropping diodes in line to distribute the heat among more devices. There are the recommended I & O caps on the 7805, but they seem to be fine.

Thanks to everyone who helped with ideas!

--

73, Dan? NM3A

Follow up on my current solution.?

I played around with multiple series diodes to drop the voltage input to the 7805. Compromise was 3 1N4001s in series. This gives about a 2.4 V drop at 300 mA.?

At 13.8 V input (nominal '12V'), 7805 stabilized at 140 F, which seems a reasonably safe temperature. At 15 V, it is about 165 F, which is high, but tolerable.?

More diodes does reduce the temperature (4 diodes- 130 F, but at the expense of raising the minimum voltage for 7805 regulation. With 3 diodes, minimum input voltage is 9.4 V.?

I admit that this is a rather inelegant solution, but it is cheap, as it all came from my junk box. A better solution would be to replace the whole 5V p/s with a buck converter, as it should be much more efficient and allow for a wider range of input voltage. Drawback from that is time and expense of obtaining one and possible RFI, although that is not a big concern in a dedicated transmit only rig.?

--

73, Dan? NM3A

Dan,

?

In addition to the heatsink, you need to add the proper capacitors as clearly stated on Page-1 of the datasheet. Self-oscillation is a very common problem.

?

?

I have single and dual 7805's in three of my U3S transmitters as well as in numerous other projects so that everything plugs safely into my shack 13.8 supply or a car.

?

I also saw you mention buck-boost converters for future projects. While these can be useful for some projects, caution needs to be taken especially on radio projects, as they oscillate and can/do cause interference of their own.

?

73,
BrianB
N6CVO

May be worth considering.

?

https://www.mouser.co.uk/ProductDetail/MaxLinear/SPX29300T-L-5-0-TR?qs=S%2FCBhQS5rCqJBma%2FXVtMMA%3D%3D