Hello,
I have a power supply, made about 30 years ago. It is very stable, 0-35V, 1A
It works perfect for most circuits, however on RF circuits (transmitters 400-500 MHz) I have a problem: my RF circuits draw about 50 mA. They use 3V.
A soon as I connect any transmitter, the voltage drops to 2.4 - 2.8V. I can limit the current, but I am not limiting it in this case.
I added an external capacitor, 25,000uF / 75V to the output, but it has no effect.
Any idea why it happens, and any suggestion about how can I fix it?
Thank you in advance!

 

Maybe try smaller HF capacitor(s) on the output - 100nF, 10nF, 1nF.
It is also possible that the transmitted RF is directly affecting the power supply - is it in a metal or plastic case? - or being picked up by the mains power lead.

 

Are we to assume that you are setting the power supply output to 3V?
Do you have an oscilloscope in order to examine the voltage output of the power supply?
Can you measure the current taken by the RF circuits?

 

Capacitors made for bulk bypass, like electrolytics, in an RF circuit
look like they are open, their ESR completely shot by parasitic inductive
and loss effects.

RF Capacitors - http://www.kemet.com/RF

Regards, Dana.

 

Thank you for your replies,
The tantalum capacitors on the board are new. 100uF. (+ 100nF + 1nF).
I have an oscilloscope, but I use a Fluke for measuring the events in real time.
There is no case, just PCBs. However, I've tried to use a copper strip and covered the entire PCB (without the antenna, 1/4 wave).
I have several identical PCBs, double sided (the bottom is GND plane). Each one shows me a different voltage reading, from 3V to 2.2V...

 

Connect the oscilloscope to the power supply output and look for possible oscillations.

Connect an ammeter in series with the RF circuit power input and measure the supply current.

 

Power supplies react to demand for current. It is more likely that you have an uncontrolled situation where you are not sure about the current requirements of PCBs under development. A power supply will reduce it's voltage output if it cannot supply the required current. You are at the low end of the adjustment range and the regulation there may not be what you think it is. I would try various fixed loads like 100' of AWG#16 wire to see if you get the same behavior.

 

As an aside, polymer tantalums better than legacy parts see attached.



Regards, Dana.

 

Thank you again for suggestions!
To MrChips, Papabravo and Dana:
I used the "scope" and all I could see at maximum amplification is the "natural" noise, random small junk - it looks exactly the same if connected to a resistor load or anything else, non RF. I cannot measure them - both frequency and amplitude are not stable at all. I would say that it is within 10-100 uV. I'll test it with a piece of "Solderwick" but it will become a "short" circuit, and I believe that I won't be able to see any additional data on my Tek... I also believe that it'll become an "antenna" to various junk signals in the room - but it worth a try!

The power supply may work within 0 to 1 Ampere, and I can adjust (limit) it as I wish. In this case, I don't limit it at all. The transmitters draw about 50 mA, tested with 2 X AA batteries (new ones...), or the power supply. I use a second Fluke for this task. It reminds me some SWR problems in the past... but they were by using 35-50W transmitters...

To Dana: I partially know (just very limited knowledge!!!) the influence of Tantalum caps in different circuits. I had to re-cap my scope, as it is an old Tek 2465. Due to my ignorance, I bought a lot of capacitors which have the same value as the ones I've replaced. It was a real fiasco! Then, I've been in touch with a forum who deals with Tek, and one guy from the forum sold me a kit of APPARENTLY the same values I had, but totally different. As soon as I've replaced them, my problems with this scope disappeared. However, this is a small PCB, and the only places I can insert the Tantalum caps are near the PS source, and at near the output. However, I tried several caps I have, and even tried to replace them by Electrolyte caps, just for testing... there were no positive results.
Thank you very much (so far...)
Regards,

 

The power supply, does it have a min load requirement ? Can you
tell how it was implemented, eg. 3 term reg vs discrete OpAmp
pass transistor design ? Use a dummy load on it to draw the
applied load equivalent, does it behave as a regulator ?

Regards, Dana.

 

Hello,

You might want to have a look at page 3 of the attached PDF of a powersupply designed by late Tony van Roon.
There you will see some inductors and a capacitor on the output to make it insensitive for RF.

Bertus

 

The must to be there is ferrite bed taken off the computer cabling, any hi-loss core of out diam about 1-2 cm and length 2-5 cm. Otherhow capacitors dont work well. The electrolytes are too slow. What You need it is cascade of 1 mF; 0,1; 0,01; 1 nano, 100 piko, 10 piko and that ferrite, in worst case ferrites may be more than one. Very good results often are using a coaxial as the power cord, try it, as well the 75 as well the 50 Ohm.

 

Can you slip some ferrite beads on your fluke meter leads?

 

The power supply, does it have a min load requirement ? Can you
tell how it was implemented, eg. 3 term reg vs discrete OpAmp
pass transistor design ? Use a dummy load on it to draw the
applied load equivalent, does it behave as a regulator ?

Regards, Dana.

Thanks for replying,
My power supply is "more or less" as the power supply designed by late Tony van Roon, less the ferrite beads and as I am using a double trafo, I don't need the two voltages option (0-6V and up to 30V. I have 0-30V continuous). I did the test using a regulator IC (5V to 3,3V DC), actually it is embedded on the PCB. Using 3.3V directly connected to the board , the voltage drops to 2.3V - 2.6V. Using the same PS, adjusted to 5V, connected to the regulator, the PS's voltage drops to about 4.2V, but the transmitter gets 3.3V from the regulator.

 

Hello,

You might want to have a look at page 3 of the attached PDF of a powersupply designed by late Tony van Roon.
There you will see some inductors and a capacitor on the output to make it insensitive for RF.

Bertus

Thank you,
As far as I remember, I used a similar IC, but the design is just a little bit different, as I haven't used ferrite beads...
I used large ferrite rings on the cables.
Sincerely,
Tedy

 

Can you slip some ferrite beads on your fluke meter leads?

Yes, done already
I put several ferrite beads at several points on the PS cable and the two flukes. I have ferrite beads (maybe they have another name...), anyway they look as a tube cut in two parts, and a plastic cover that pres them together so they form a complete ferrite tube. Sometimes, they are around power cables, USB cables or even small power supply cables, but mine have two halves...
They had no influence at all.
I even tried using a copper strip, and covered the leads, then grounding them. Again, unfortunately, no change, and it looks very strange to me...

 

Can you photo your setup?

 

Can you slip some ferrite beads on your fluke meter leads?

I have burned too much good digital instruments to allow me experimenting, but its worth to check. The classics is 2kV probe switch to 50 V RF signal, and 2 kV warranted 1:100 probe is dead after milisecond, and tester/oscilloscope is too.

For high intensity RF fields ought be used old stable analog heads of sixties. They never lie and never die just because RF.

 

Sounds like regulator you are using is not LDO, low dropout. And the power
supply is very low current if its output is dropping like it is.

Regards, Dana.

 

I suggest more filtering on the power output leads than just ferrite beads. Chokes with several turns of wire in both the positive and negative leads. Also, some RF bypass filtering in the mains leads to the supply. Then making certain that the entire case is providing adequate shielding. Of course, I just now looked at this thread and so these may have been suggested already.