Post a schematic of how things are connected. Using resistors just isn't going to work. The voltage they drop will depend on current. At low current, they won't drop enough voltage to limit the input to the voltage regulator. A zener diode would be better. It's a waste of power, but at least it will give you a more constant voltage drop.

A 35VRMS secondary will give you about 50V DC when rectified and filtered.

 

The resistors aren't working well but I don't have anything else. I do have some "junk" diodes from the broken oscilloscope. I'll try them and see what happens there. They should be on the primary? I don't know how much current or voltage these were rated for but they were on a CRT oscilloscope from the 80's so I would assume they can handle a little current at moderate voltage.

So if the output of the lm317 is connect to the heat sink and my negative terminal is grounded to the case is that a problem or am I overthinking all that? I just don't want to blow something when I have a long drive or wait to get anything extra. That was why I was looking at the resistors as an option for reducing the voltage.

 

The resistors aren't working well but I don't have anything else. I do have some "junk" diodes from the broken oscilloscope. I'll try them and see what happens there. They should be on the primary? I don't know how much current or voltage these were rated for but they were on a CRT oscilloscope from the 80's so I would assume they can handle a little current at moderate voltage.

I'd be a lot more cautious if I were you. Working on 120VAC and not really understanding what you're doing can be dangerous.

Don't go sticking just any old diode or resistor on a 120VAC circuit; unless you like the look and smell of components burning up.

So if the output of the lm317 is connect to the heat sink and my negative terminal is grounded to the case is that a problem or am I overthinking all that?

This is a problem that needs to be addressed. You require a good thermal connection between the heatsink and regulator, while not shorting the output because the heatsink is probably grounded. The usual way to do this is to electrically isolate the regulator from the heatsink.

That was why I was looking at the resistors as an option for reducing the voltage.

Resistors and diodes are brute force approaches and generally simply waste power.

The switching, tracking preregulator I suggested is one of the better solutions. It drops the input voltage to the regulator in an efficient manner and, since it's output tracks the linear regulator, it minimizes power dissipation in the LM317. You get most of the efficiency of a switching regulator, but don't have to deal with filtering noise.

 

Testing some diodes in the input. So only one diode either direction my variac growls at me. With two in place both directions it is much happier. I get a reading of 17.85vdc and 20.35vac with a 300uf capacitor after the rectifier. I don't have a 1000uf cap. All I have is a old 300, 100 and 100uf electrolytic can. Going to put all three together for 500uf and test once more. This is just strange. It's too bad this transformer has to put out too much voltage. I may have another. I just remembered I have another transformer from that broken oscilloscope that may have a 30v winding!

 

Testing some diodes in the input. So only one diode either direction my variac growls at me. With two in place both directions it is much happier. I get a reading of 17.85vdc and 20.35vac with a 300uf capacitor after the rectifier.

You're going to hurt yourself or burn something down; I won't be a part of it...

 

Post a schematic of how things are connected. Using resistors just isn't going to work. The voltage they drop will depend on current. At low current, they won't drop enough voltage to limit the input to the voltage regulator. A zener diode would be better. It's a waste of power, but at least it will give you a more constant voltage drop.

A 35VRMS secondary will give you about 50V DC when rectified and filtered.

I'd be a lot more cautious if I were you. Working on 120VAC and not really understanding what you're doing can be dangerous.

Don't go sticking just any old diode or resistor on a 120VAC circuit; unless you like the look and smell of components burning up.
This is a problem that needs to be addressed. You require a good thermal connection between the heatsink and regulator, while not shorting the output because the heatsink is probably grounded. The usual way to do this is to electrically isolate the regulator from the heatsink.
Resistors and diodes are brute force approaches and generally simply waste power.

The switching, tracking preregulator I suggested is one of the better solutions. It drops the input voltage to the regulator in an efficient manner and, since it's output tracks the linear regulator, it minimizes power dissipation in the LM317. You get most of the efficiency of a switching regulator, but don't have to deal with filtering noise.

I'm being as safe as I can with the testing of 120vac. I'm using my isolated transformer then into my variac. From there I'm sitting in my chair not touching any ground or anything else with more then one hand(only temperature) with one hand on the variac switch ready to turn it off unless I'm touching something else. It's about as safe as it can ever get. From there I have my watt meter on amps and my multimeter leads giving power out. Then some alligator clips from routing a little more. And of course my little fire extinguisher next to me. The only exposed connector I'm ever touching out of this is the variac and the neutral power line side if I slipped off something. Since everything routes through the transformer before my testing of resistors or diodes(hot side into transformer then out into neutral line) it will limit any current(or so I hope).

I just solved the lm317 heat sink problem. I took off a kaypro computer powersupply some little what looks like rubber isolator and plastic bushing. I don't have a connection to ground anymore. I may also take some of the filter capacitors off of it and put the, together to see if I can get a 1000uf 50v rated. Or as close as possible. And fine. I'll stop testing resistors and diode. Feel better? I'm going to continue working on the circuit for now and address my voltage later if it an issue.

The insulators were from a regulator chip just like this(same size but different part number). I need to buy some of those. Why can't they just come in the package with the chip?

 

So that transformer from the oscilloscope has a 30v winding. Also a 181v winding. And I for the first time time ever measured volts in the ohms range. It's survived and works fine. Suprisingly never did that before. Well I clipped the 1v and 181v leads down short but left enough I ever wanted to use them and did heat shrink on the 181v then electrical tape and all them. I put the two primary leads together so I didn't have to do the same with a 120v set. Soldered it in. Making progress!

 

So I took all the parts, well mostly out of the case. I will reuse the diodes, potentiometer gang, meter and terminal strips. The rest rest is either considered bad or worthless(mostly just wires, the resistors if good will go to other stuff). As for the transformer, I tested it off my variac. Secondary winding puts out 73v. Center tap to either winding is 35ish volts. Would that be a problem? The rest of everything looks good. I'll try to hit a RadioShack soon and see what parts they may have before ordering online. Even if a little more it's better then a week from us or way more from China.

The 35-0-35V secondary is so you can have a full wave rectifier with only 2 diodes.

 

Got the circuit board mostly done. Just drilling some holes to mount it before soldering in my like 10 wires to everything. Well it was going good until my drill battery died. Of course! But I'm mostly there. I just need the 1000uf capacitor and a power cord. Can I make a substitution for it? I have some 2200uf 16v on the kaypro power supply. I know I can wires those in series but it's still not 50v. What can done for that last little stretch? Thanks

 

Hello,

As electrolytic capacitors can have a tollerances upto 20%, the voltage balance will be unpredictable.
Thats why it is not wise to put the capacitors in series.
Also putting a to high voltage on a capacitor can make the burst and cause fire.

Bertus

 

I took apart a old computer power supply. And behold. A 1000uf 200v capacitor. There are two in case I need a second. I got the circuit board kounted and will now proceed to wiring everything else.

 

Now that you're not putting parts willy nilly on the primary...

You need to address the maximum input voltage of the LM317. It's 40V and if your rectified, filtered DC is 50V, the regulator won't be happy and you may be making a trip to buy another.

The self oscillating, switching, tracking preregulator circuit I posted is one of the most efficient ways to deal with that. It uses commonly available parts and will relieve you of the need to drop the secondary voltage by brute force.

Reusing salvaged electrolytic capacitors is fine as long as you understand that when they wear out they can create difficult to debug problems if you don't have a scope.

 

So I forgot to test the rectified voltage during testing. It worked perfectly. Only thing wrong was the meter wired backwards and I had to cut on of the transformers primary leads. I guess it wasn't as happy as I thought(It smelled like epoxy/car detail?!?! That was weird but solved now). Meter was dead accurate with my digital multimeter and the potentiometer adjusted everything nicely. And both ranges adjusted perfectly. I'm happy. I'll measure that one voltage.

Edit: Quick test shows 35-37/38vdc after rectifier. I'm happy. Just need to solder two more leads on the potentiometer as I didn't know which way would adjust it up in volts clockwise. And then find and install a power cord. I also picked up 1/2a fuses at the RadioShack for it. That way I can only draw 1/2a so I don't overload like the transformer or something else. Thanks for all the help.

 

I also picked up 1/2a fuses at the RadioShack for it. That way I can only draw 1/2a so I don't overload like the transformer or something else.

Fuses don't work the way most people think. Here's an excerpt from littlefuse.com:
The fuse must open at 135% of rated current within one hour. It also must open at 200% of rated current within 2 minutes for 0-30 ampere ratings and 4 minutes for 35-60 ampere ratings.

 

Hello,

As electrolytic capacitors can have a tollerances upto 20%, the voltage balance will be unpredictable.
Thats why it is not wise to put the capacitors in series.
Also putting a to high voltage on a capacitor can make the burst and cause fire.

Bertus

Its lithium cell that can cause a fire - pretty rare with electrolytic capacitors.

They can overheat and build up a head of steam - in the early years, numerous techs were injured (some blinded) by the can bursting off like a bullet.

Most modern electrolytics have creases stamped in the top of the can - this deliberately weakens the can so the top can split and it vents safely.

Any electrolytics without this feature - wear eye protection!

 

Fuses don't work the way most people think. Here's an excerpt from littlefuse.com:
The fuse must open at 135% of rated current within one hour. It also must open at 200% of rated current within 2 minutes for 0-30 ampere ratings and 4 minutes for 35-60 ampere ratings.

Regardless it's still better to have it. Well I got it completed! Got a powercord installed. And finished solderi everything and got my led working. I picked up some green ones at RadioShack and had to do a new set of resistors for them.

 

Well I got it completed! Got a powercord installed.

Nice! Congrats! Now you can work on some fun stuff. Don't forget to put some strain relief on the power cord when you close it up.

 

Nice! Congrats! Now you can work on some fun stuff. Don't forget to put some strain relief on the power cord when you close it up.

That's the problem. The only strain relief I have is fixed. It won't let the cord move. That's fine until you need to remove the case. And the cord is too thick to knot like the original. I like need something fixed to the cord on the inside and not the case. That way I can remove the case. I'll have to see if there's anything I have or at like ace.

Also I popped my first fuse. I guess it's like knives, until you blow the fuse it's not really yours. I tested a 6v6 vacuum tube which uses 6v at 2.5ohms. This thing growled and now I know why. It was trying to draw 2.5amps. Then I blew the fuse when I hooked up a 12ohm resistor. It's all fun and games until the fuse pops. I still got 3 more! So more fun ahead!

 

I tested this with a 75ohm as it should be 30v 75ohms= 0.4a. A little low on wattage for my resistor but it was a quick test. It worked and I was able to get a max of 20v with the load. That seems that it will be minimum resistance that I can have on it. I then also tested it with a computer fan at. 12v and 0.2a. I let that run for a while. It preformed perfectly. If I ever find a high amperage 30v transformer I'll mod the circuit for higher amps output. But I'm happy. Thanks for all the help.

 

I like need something fixed to the cord on the inside and not the case.

A thick typwrap (wire tie) clinched around the cord makes a nice restraining ring.

Can you post a schematic of your final design?

ak