Building the FG085 Arbitrary Function Generator (2/2)

[jp]

I ran into a similar situation yesterday. I was building a 3 amp 12 volt power supply using a 12 volt 3 amp wall wart. No load on wall wart is 14.7 volts rms. I thought excellent. Wired up a test circuit. I tested the circuit and loaded it to 2 amp draw. Low an behold the AC voltage dropped to 12.5 volts rms and the DC voltage then droopped below the drop out voltage range of the LM317 regulator I was using with a current boosting pass transistor. Lesson learned. Back to the drawing board. New approach needed to get what I want with the transformer I have. I guess if every project work 100% right off the proto board electronics would be boring. Enjoy your vids. Great job. Happy New Year

[MsMammeta]

the voltage drop that you have found is not caused by the signal generator: the cable has an impedance of 50ohm and you're placing a 50ohm load at the end of it. This creates a voltage divider of 1/2 where you are measuring the voltage with your oscilloscope. Try to measure the voltage at the output of the signal generator while you have the load connected and you will find that the output is 5 volt.

[DeeegerD]

The output impedance is 50 ohms and you are using a 50 ohm terminator thus creating a voltage divider (2.4v vice 5sh). This is common for cheap FGs that report their true output voltage instead of what the output voltage will be when properly coupled with something else. Nice build.

[MrSailbadthesinner]

If you need 3v headroom for the regulator then you need a minimum Vin of 18v, add on Vripple across the capacitor and Vfwd for the bridge and you will need considerably more than you 12v transformer can offer. (Check out Line and Load regulation to see why your transformers give higher voltage than you expected). Your 18v transformer (gave 20v OFF load) gives 25.4v peak at nominal input, allow for +/-10% for the mains supply and you will have 22.8 to 27.9v. Less say 1.8v for the bridge. Your 2200uF capacitor has a tolerance, typically +50—20% so worst case it's 1760uF. The FG085 draws 150mA + load (say 10V into 50ohms = another 200mA) So for 350mA a 1760uF cap will give 2v ripple (at 100Hz = rectified 50Hz). So you get a minimum voltage into your reg of 22.8 – 1.8v - 2v = 19v and a maximum voltage into the reg of 27.9 – 1.8 = 26.1v. Worst case Vdrop on the regulator is 26.1 – 15v = 11.1V at 350mA = 3.9W. So for a maximum temperature of say 100oC and an ambient of 30oC you should have a heatsink better than 17.9 oC/W on your regulator. Linear Power supply design isn't quite as simple as many expect! Hope this helps others.

[WaltonPete]

The finished build looks much better in the enclosure than the original back plate assembly - looks like a purpose built device, rather than a kit. Great job!
I know exactly what you mean about the slotted screws - they were used extensively at one time and now, when I come across one, I find it difficult to keep the screwdriver in the slot! Perhaps we've been spoilt by all the other types of screws that are routinely used these days?

[sebastian]

The specs of the transformers are given under normal operation (under load; so the 18V transformer may give 20V in open circuit but under nominal load it should give the promised 18V.
Another thing is that the mains voltage changed since the transformer was fabricated, so it will be out of specs no matter what. Many european countries changed 220V mains with 230/240V (that is around 5-10% higher) so the appliances that uses transformers will be stressed in theirs power supplies and also the transformers are more easy to fail or at least will work hotter. I knew a friend used to put a small value resistor in series with the primary just to bring the primary voltage a bit closer to that 220V that was originally meant for the transformer. That makes sense to have 20V from a nominal 18V transformer when the input voltage rise from nominal 220V to 240V.

If the output impedance of the generator is 50ohm and you use a 50ohm load, you should have half of the voltage measured on the load instead full voltage measured without the load. The internal impedance and the load makes a voltage divider and when both values are equal, the output is half the input. So, it works as intended. Maybe the 50ohm output impedance is useful when driving higher impedance inputs (preamps, amplifiers where the inputs are in kohms range)

[sebastian]

The old signal generator is homemade, a kit or industrial built? If you build it, have a video about it?

[cbcdesign]

This kit is easy to build but the finished item produces some rather noisy wave forms which get progressively worse as the amplitude goes down. Its very noisy if you need to produce wave forms with peak to peak values in the low 1 volt range. I found adding a small value 10nF 50V multilayer disk ceramic capacitor across the BNC connections helped clean up the signal so that its quieter on lower amplitude signals at lower audio frequencies which is what I need it for. Longer term I may build up a buffer circuit with an attenuation function.