Recently one of our function generators (UTG9005C) started smoking and got slightly damaged (a couple burned transistors and resistors).
The function generator has 50 Ohms output impedance and was generating a square waveform with voltage 20 Vpp (this is pretty much the maximal output voltage) to a 50 m long coaxial cable with characteristic impedance 75 Ohms. The coaxial cable was terminated with an open circuit so there was a reflected wave.
The coaxial cable was attenuating the reflected wave to less than 50% of the generated one (according to the oscilloscope), so the amplitude of the reflected wave must have been about 7-10 Vpp.
Could this reflected wave be the cause of damage of the function generator? If yes then how would you recommend to protect it?
I suppose one option would be to connect a resistor in series between the coaxial and the generator but i would like to hear other suggestions.
Two suggestions: Firstly not just a resistor but a well designed attenuator between the function generator and the co-axial line would help. See https://en.wikipedia.org/wiki/Attenuator_%28electronics%29.
The second option would be to use a vacuum tube amplifier with the appropriate output impedance between the function generator and the co-axial line. Amplifiers with tube output elements tend to be much more robust than those with semiconductor output elements.
Dear Philip, thank you for the suggestions. I suppose an attenuator would be best indeed however common ones would not fit because the output probe of the function generator is with two crocodile-type plugs. Perhaps i can find some sort of converter or use a different probe. Will have to check the market.
The vacuum tube amplifier sounds like too much of a headache.
A reflected wave coming from the generator just applies voltage to the output connector. But without current flow I guess this is not the reason of your smoking device.
The UTG9005 is not short-circuit protected and so you have to take care not to short your output connector e.g. with the end of your coaxial cable. A 50Ohm resistor in series with the connector should protect your generator.
Addendum: Crocodile clamps often tend to short your circuit. :-)
Dear Marc, an open circuit reflects the wave with the same direction as the incident one, while for a short circuit the reflected wave is inverted. Considering the reflected wave lags the incident one in both cases the voltage at the source becomes higher than the applied one (square wave) for certain intervals of time. So if for example the applied voltage is 20 Vpp, because of the reflected wave it becomes near 30 Vpp (not 40 because of the attenuation). So the instantaneous dissipated power at the function generator's output might be higher than it was designed for.
A short circuit of the crocodile clamps might cause such problem but i'm sure this is not the case this time, because i was observing the waveform at the oscilloscope. If there is a waveform, there is no short circuit :)
I was not sure if you observed the reflected wave while your generator was smoking. So in that case the generator can not handle voltages in the range of 30Vpp.
Anyway. A 50Ohm resistor in series should work as well as the attenuator which normally is designed for 50 Ohm termination. Of course your output voltage will decrease appx. to 50% of the generator voltage.
The reflected wave only alters the impedance seen by the signal generator. If the cable is an odd number of quarter waves it would reflect a very low impedance that a non short circuit protected signal source could not handle. That could damage the output circuit of the signal generator.
Thank you Alan, this is another explanation for what could have happened. At certain frequencies the input impedance of a lossless line is zero (each half wavelength). Of course 50 m of coaxial cable isn't a lossless line but it's input resistance would still be close to zero.
Why not add 25 ohms output impedance in order to match the 75 ohm line?
Another hint: just place an 75ohm resistor at the end of your cable to avoid reflections. [ r=(Ze-Z)/(Ze+Z) ] // Z(cable) = 75ohm
'
Surely the answer to this question is to look inside
the generator and find the source of the smoke.
If it is in the RF output amplifier, you will have one
answer; if it is discovered in the power supply you
have another answer.
.
Barrie
Henrik, Marc, impedance matching is of no importance here because the idea is to demonstrate the reflected wave and the distortions caused by it. The plan is to use the coaxial cable with a potentiometer load.
Berrie, the generator's electronics was working properly. The only thing that didn't work was the output circuit (no output signal). Anyway it is fixed now. As i said in the first post there were a couple of burned transistors and resistors.
Yes, I suppose so. Unless there are other interesting ideas for protecting the function generator from reflected waves.
Vasile, no DC offset voltage was used and the transistor wasn't killed instantly. I was observing the waveforms for at least 10-15 mins.
But you are right. Unless i use some well designed hardware repeater/attenuator after the generator, incidents will happen from time to time.
- Badges
- Science topic
- Similar topics
- Engineering
- Electrical Engineering