100 /100 cap can in series for 50uF ?

[Colas LeGrippa]

I have to replace a multiple cap can in a Sunn (Solarus) amplifier. I have to get 50uF total for the reservoir cap  with one double cap can 100uF/450V. Here it is : from both ++ tied together to -, I get 200uF, from each + to -, 100 uF and from + to + , I get 50uF. you can easily understand that.  How can I hook up two + side to get 50uF ? One to the B+ and the other to the chassis ? What do I do then with the negative terminal ? I could use a 50uF alone, but I am not sure if 450V will be enough. By using two 100  bypassed with 220k resistors, I jump to 900V .

Colas

[Willabe]

I don't think you can do that with a cap can. You only have 1 - lead and you would need 2 - leads for 2 caps in series. Got to go; B+, +, -, +, -, ground.

Schemo shows 1'st filter cap rated at 600dcv all other caps 525dcv. Doesn't list any B+ node voltages.

Schemo also shows a GZ34 rect. tube which has slow turn on so no B+ voltage spike at turn on. That's good you don't need much extra head room dcv for B+ node cap values.

Get 2 single axle caps at 100uF/300dcv, there smaller than 450dcv cap and might fit?

              Brad    

[kagliostro]

You can only parallel it

K

[Colas LeGrippa]

ok but why do I get a reading of 50 ?

[Willabe]

ok but why do I get a reading of 50 ?

     

[Willabe]

from + to + , I get 50uF. One to the B+ and the other to the chassis ? What do I do then with the negative terminal ?

ok but why do I get a reading of 50?

Wait a minute. I think it's because it's now a non-polarised cap the way your hooking up your meter.

It's, in, +, -, -, +, out the way your meter is hooked up. I think since the caps are still in series that's why your getting 50uF. I don't know what the dcv raitings are this way.

                  Brad      

[HotBluePlates]

I think I know what you're describing, which is just different connections of your meter leads to the cap sections?

*If* you have a dual section 100+100uF can, you cannot place those cap sections in series. It's not physically possible because they share a the negative terminal.

You might be measuring from one positive terminal to the other. Forget any reading you get there; in use, you will have a high d.c. voltage trying to polarize the cap in one direction. I'll bet a day's pay the cap will blow up when you switch the amp on the first time with this hookup (it *is* backwards of how typical non-polar electrolytics are made from a pair of polarized electrolytics). Typical non-polar electrolytics are made by connecting the positive leads together, and having individual negative terminals, which then become the 2 leads used for the composite cap.

Also, your measurement doesn't include any significant d.c., and will lead you to a wrong conclusion.

In the old days, Hewlett Packard made test gear where they needed to place electrolytic cap cans in series to obtain the needed voltage rating. 2 separate cans were used (even though each had multiple sections). The one with its negative terminal at ground potential was bolted to the chassis to make the ground connection. The second cap was place in series by connecting the can's negative terminal to a positive terminal on the first can.

The trick was that the can with the elevated negative temrinal voltage had to be mounted on an insulating phenolic wafer to isolate the can from the chassis (ground), and further had a cardboard sleeve that fit over the can so you didn't get shocked by 400v (or whatever voltage that can's negative is at) if you touched it.

Hewlett Packard had very clever engineering for their tube test gear in the 40's-70's. If they couldn't cheat and get a can cap to do what you're contemplating, I promise it's because it can't be done.

[Colas LeGrippa]

I won't do it I swear to God

[alerich]

In the old days, Hewlett Packard made test gear where they needed to place electrolytic cap cans in series to obtain the needed voltage rating. 2 separate cans were used (even though each had multiple sections). The one with its negative terminal at ground potential was bolted to the chassis to make the ground connection. The second cap was place in series by connecting the can's negative terminal to a positive terminal on the first can.

The trick was that the can with the elevated negative temrinal voltage had to be mounted on an insulating phenolic wafer to isolate the can from the chassis (ground), and further had a cardboard sleeve that fit over the can so you didn't get shocked by 400v (or whatever voltage that can's negative is at) if you touched it.

That's what Sovtek did in their MIG 60 head. They used two cap cans that each contained 100uf and 30uf sections @350VDC. They paralleled the caps in the cans and then connected the positive lead of the first to the can (-) of the second. The second cap can is enclosed in a thick plastic sleeve. The junction between the two cap cans provides the screen voltage while the full potential across both cans provides the B+ for the plates.

[sluckey]

It's, in, +, -, -, +, out the way your meter is hooked up. I think since the caps are still in series that's why your getting 50uF.

Exactly. But it's not wise to use it the way Colas would like to.

[HotBluePlates]

It looks like I may have been wrong.

I found Cornell-Dubilier's Electrolytic Application Guide, which discusses non-polar electrolytics on page 13.

CD states that connecting - to - or + to + will provide a non-polar cap, however the capacitance rating is half what you'd expect from series caps, or 1/4 the individual cap capacitance. They also state there is a different production process they use for non-polar caps so they're a little different than just back-to-back connection of standard caps.

[PRR]

> the capacitance rating is half what you'd expect from series caps, or 1/4 the individual cap capacitance.

That's debatable, conservative.

But the hard-fact is the Voltage Rating is the same as the one forward cap. (The one in reverse breaks-down at about 1 Volt.) So two 450V caps back-to-back is still a 451V cap. But less capacitance than one cap would be.