1
$\begingroup$

I made a simple linear power supply for my transceiver using 12V 30A transformer, four diodes RHRP3060 forming a rectifier and three 30_000 µF capacitors connected in parallel. I also used a 200 Ω 50W resistor to discharge capacitors in a few seconds after powering the device off.

The power supply works reasonably well and I already made a few QSOs using it. However if I transmit in FM using 100W the oscilloscope shows +/- 6 V ripple. It's not that bad in SSB though. In this mode I see 9-15V which is in safe range for my FT-891, which is 13.8 +/- 15%.

What upsets me is that the cheapest switching 12V 29A LED driver I could find gives only +/- 1.3V ripple in FM mode. If anyone is interested, it's GDLI-350-IP20-12, the price is ~15$.

How can I reduce the ripple of my power supply to approximately the same level? Should I use even more capacitance, a high-pass filter, or maybe something else?

$\endgroup$
  • 3
    $\begingroup$ Terminology: that's ripple, not jitter. Jitter is when a periodic signal isn't quite as exactly periodic, but off by a random amount of +- µs (or whatever timescale applies. $\endgroup$ – Marcus Müller Feb 14 at 17:32
  • 2
    $\begingroup$ Your 200 ohm, 50 watt resistors are totally unnecessary in a regulated supply design. Your design is not regulated in any sense of the word. An LED supply is not a reasonable comparison as these are typically based on current, not voltage, regulation. Before you damage your radio, you should study linear voltage regulator designs. Your voltage ripple goal should be less than 0.25 volts. $\endgroup$ – Glenn W9IQ Feb 14 at 21:32
  • $\begingroup$ albeit 0.25 V is a harsh requirement, you definitely should not see 6V variations at all, in any reasonably designed power supply. My guess: a (used) PC power supply's 12V rails, if tied together, will do much better here, for basically free, given you can find a Pentium 4-era or later used PC on a corner somewhere. $\endgroup$ – Marcus Müller Feb 15 at 8:30
7
$\begingroup$

Unless you haven't mentioned all the parts, you have built an unregulated power supply. The output voltage of a transformer varies depending on load, and the other components will just follow that voltage.

For electronic equipment you should use a regulated power supply. This requires the rectifier to produce an output voltage somewhat higher than the target 12 V, and a regulator circuit (or IC) which controls power transistors to drop it down to the exact target voltage, allowing the final output voltage to stay constant even as the transformer's output voltage varies. These transistors must have a heat sink (for all but the smallest linear supplies) as it will produce heat equal to the excess voltage times the load current.

(Arguably, your current power supply should not even be called “linear”, as it does not have the above-described “linear regulator” circuit in it. The common terminology usage is not entirely consistent, as sometimes “linear” is used to describe any non-switching power supply.)

I cannot advise you on the detailed design of linear power supplies; this answer is just describing what is obviously missing.

$\endgroup$
  • 2
    $\begingroup$ If you want to design a power supply yourself, the "Webench Designer" from www.ti.com does an actually pretty nice job at taking Vin, Vout, Iout specs and coming up with viable designs. $\endgroup$ – Marcus Müller Feb 14 at 17:34
  • 1
    $\begingroup$ Given the current involved, the description of "a" pass transistor is misleading to the OP. It will most certainly require multiple pass transistors. $\endgroup$ – Glenn W9IQ Feb 14 at 21:41
  • 1
    $\begingroup$ @GlennW9IQ Thanks, edited! $\endgroup$ – Kevin Reid AG6YO Feb 15 at 1:20
  • $\begingroup$ @MarcusMüller I edited your comment with what might be an appropriate link to Webench Designer. $\endgroup$ – Mike Waters Feb 15 at 1:40
  • $\begingroup$ @MikeWaters not the designer itself, but the instructional video and a link to it on that site, but I think that's the optimal place to send people, so thank you!! $\endgroup$ – Marcus Müller Feb 15 at 8:08
1
$\begingroup$

I'm surprised you haven't damaged your radio! You need a voltage regulated linear power supply to provide a stable DC voltage to the radio. Your 100 watt transmitter likely draws about 20 amperes of current in FM mode. A popular supply is the Astron RS-35. If you examine the output of your current supply with an oscilloscope, you will see a lot of AC ripple or hum on the DC output. This will vary with load current being drawn. Your simple unregulated power supply output voltage will rise much higher during light load (receive) and drop to unacceptably low voltage during heavy load (transmit).

$\endgroup$
  • $\begingroup$ The RS-20 is too small for a 100 watt transceiver. $\endgroup$ – Mike Waters Feb 14 at 17:54
  • 1
    $\begingroup$ I agree if the efficiency of the transmitter is poor. 20 amperes at 13.6 volts is 272 watts of power. Hopefully the efficiency is over 40% even with the digital circuits and display running in the radio. I see the manufacturers specifications for the radio is 23 amperes. My mistake, thanks! $\endgroup$ – user14247 Feb 14 at 19:30
  • $\begingroup$ If the OP's specs are correct, the 12 volt 30 amp transformer is not likely to "drop to [an] unacceptably low voltage during heavy load (transmit)". It is more likely to come close to the target voltage - thus the misunderstanding. $\endgroup$ – Glenn W9IQ Feb 14 at 21:38
  • $\begingroup$ @BobMaverick most 100W rigs spec somewhere between 21A and 24A max current. $\endgroup$ – hobbs - KC2G Feb 19 at 17:57
1
$\begingroup$

You do have an unregulated supply .This issue has been raised by others .You can reduce hum by increasing Your total filter Capacitance .The ripple volts is inversly proportional to Total Capacitance .If you have Caps on hand then why not add these to what you already have .Large numbers of small caps will work fine .You could also resolve the Hum issue by using LC filtering which means a choke that will handle the current .

$\endgroup$
  • $\begingroup$ There is a limit to how much capacitance can be used. It's not a matter of "if some is good, then more is better, and a whole lot is just right". As the value of the filter capacitance is raised, the peak current through the rectifier diodes increases as the conduction time every cycle decreases. That can cause not only the diodes to fail, but the transformer to overheat. What is really needed is a regulator circuit, and not more capacitance. $\endgroup$ – Mike Waters Feb 23 at 14:45
  • 1
    $\begingroup$ @ mike waters .Big C does cause higher transformer peak currents like you say .When I did this I had greatly over rated diodes and was using less than 50% of the transformers rating .I checked the peak currents with a scope probe across a low resistance shunt .The transformer DCR and leakage inductance kept things managable using a 30VA transformer and 1 farad of total C but expecting DC current of less than 1 Amp continious .This was 18 years ago and it still runs 24/7 . $\endgroup$ – Autistic Feb 23 at 21:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.