Hi Everyone!
About 25 or more years ago, I built a rectifier to use my AC Buzz-Box welder for DC welding. Now that I am a full on grown up, I own a DC SMAW unit & don't need this for welding.
I decided I'd like to use this at my project bench as I learn electronics. Input will be through my isolation transformer, then a variable transformer. Output will serve Project DuJour.

I think I wired it like I did to do this:
Before a welding arc is struck, each capacitor will allow 120hz to be the output. Once a load is impressed, the capacitors could not keep up with the demand and the diodes would take over, giving me DC at zero Hz.

The varistor & resistor were probably included to maintain a small flow so the AC would be available for the next arc strike.
But, i've forgotten more than I ever knew in the first place.

My main question pertains to frequency:
At this time, if I connect it to the transformer, my frequency meter shows 120 Hz with or without a load (resistive test loads of 800 & 1200 watts). I don't have an oscilloscope, just a simple Fluke 87.

The voltage reading on DC (122vac input) is 108 no load. I understand the drop through the diodes. With the meter on 'AC Volts', it shows 53 volts.

I wondering why I get 120hz instead of zero when it is under load. And why, with the meter set to AC, I get 53 volts instead of zero.

Is the 120hz because of the way the capacitors are wired across the diodes?
If I removed the capacitors across each diode & simply put one smoothing capacitor across DC+ & DC- do you all think this would give zero Hz?

There is a drawing attached. A couple of photos are also attached. (The digital readout on the front is just the temperature inside the case, not volts.)

Thank you very much for any advice you can give. It will be appreciated & remembered. (Because now I write everything down! Hey! Where's my pencil?)
Paul

 

What you are measuing is the AC ripple "riding" on the DC voltage.

The DC voltage is lower due to that ripple, the voltage is varying between 100 and 173 or so volts 120 times per second (due to full wave rectification, Frequency is doubled).

This image is for 3 phase, but Full Rectified from Wikipedia. Single phase will look similar (pay attention to only one line color for 1 phase) if you pay attention ONLY to the top "humps". The capacitors you are using aren't large enough to supply current at the full voltage between each "hump", leaving you with 53V RMS of Ripple on the output.

 

Thank you for taking the time to reply, Thatoneguy.

I'm still a bit confused, even after studying your explanation. Guess I'm a bit thick skulled these days.

Is the 120hz ripple I'm reading something to be ignored if I will use this as a bench DC power supply, or should I address it?

If the ripple should be addressed-
After reading your linked article on decoupling & bypass capacitors, I think the 22uf electrolytic and a 0.1uf ceramic paralleled across DC+ & DC- would be ideal.
Would I also remove the existing capacitors across each diode?
And the varistor/resistor combination?

Did I understand correctly or did I mess it all up?

Thanks!
Paul
PS: I think your linked formula writing article will come in very handy as I learn more & drive you all crazy with posts.

 

You would need much larger capacitors. The bypass caps article is about3.3V-5V Logic level, but does apply to power supplies in general.

The size of the output capacitor to "smooth" the DC depends on the amount of current being drawn.

 

Thank you for your reply & patience.

I'll figure out what load I plan to use at the bench & size the capacitors accordingly. In its new life as a bench supply, the load on the rectifier will be limited by the size of my isolation transformer. Simple enough.

At times my brain refuses to understand things. A little left-right-up-down trickery. One of life's entertaining challenges.

As I re-drew the circuit from memory, your explanations from Reply 1 all became crystal clear.
Kind of a brain trick I use for other things.

Thanks again for helping. I learned a lot today & it's still early.
Paul

 

Right now, it's an unregulated supply. The output will be roughly where you set it, but won't stay there based on the load.

You'll want to use a regulator, such as a 7805 for 5V (max input is 36V), or a switching type regulator for higher current.

Regulators often have great rejection for 120Hz and below, meaning you get a clean DC output even if there is some ripple on the input of the regulator.

The downside of a regulator, especially linear types, is you are limited to around 1 Amp before they are overloaded.

I'm unsure of what you'd like to power, but if you give specs for voltage range and current range, something could be found that would be much better than essentially a rectified transformer output with a bit of filtering.

 

What you say makes good sense.
It seems I'm trying to make a crude rectifier into a proper device for careful bench work. It was great for brute force welding, but if it won't regulate properly, bad results will abound. Kind of like trying to sink a wire brad with a sledge hammer.

I don't have any particular output voltage nor amperage in mind. The input will be 120vac 60hz.

It was intended to be used for general experimenting & testing circuits I build as I learn about electronics & circuit functions. It'll be kind of like a school room unit.

Do you have any suggestions for a good starting point on what specifications to look for?

Thanks Again,
Paul

 

There are literally hundreds of designs for bench power supplies (Google "Bench Power Supply Schematic"). Be warned, not all the schematics available are "good designs", some may not even work. You would want to post it here to be sure it's a practical circuit before building it.

A power supply is typically one of the first DIY builds for many getting started in electronics.

Decide what you will want to power on your bench to get an idea of what you'd like for specs.

A 1.25-12V @ 1A regulated is a decent "all around" supply, but you may want current limiting, higher output, or a wider output voltage range.

I use a fixed +5V supply and a fixed 13.8V@30A switching supply for larger loads, most everything else is covered by a 0-30V@5A lab grade ($$) bench supply. If combining supplies, you'll want to ensure all the outputs are floating (no common point between them, such as ground).

What you decide to build depends on what you think you'll be working on the most. If doing a lot with op amps a split +/- supply is needed, so having one around is handy. If working with logic and microcontrollers (a very fun area right now, affordable single chip solutions to complex circuits), only a 5V or 3.3V supply at 1A or less is needed. For audio or RF power amplifiers, you would want a wider voltage range and current ability, etc.

A common bench supply has become a hacked PC Power supply. These provide 3.3V, 5V, and 12V at tremendous current (20-100Amps). The downside is the voltage cannot be changed much, there isn't any current limiting, but the amount of well regulated power is excellent for the price and small size.

 

Thank you again!
I'll spend some good time & thought deciding exactly what I would like to accomplish with this power supply. Your notes about various projects' needs is helpful.

I certainly want to build it myself, rather than buy, mostly for the learning experience. Cost very much comes into play as well.

Thank you for the invitation to post my schematics here for review. I'll definitely do that before committing to a design.

Now, on to studying...

Thanks Again,
Paul