tracking down a short

[Bugman3183]

 hello all, I recently finished assembling an amp that I found the layout and schematic of on the net.   it's basically a cascaded 5e3  type amp using a full wave bridge rectifier.   upon completion I used my doom lightbulb tester on it, and  with the amp in standby position the lightbulb dimmed for about five or six seconds and then there was a loud pop and smoke everywhere.  I unplugged the amp and tried to find where the smoke was coming from.  I  visually checked all the filter caps and didn't see anything unusual.   so next I used my dmm to check the diodes and they seemed to be bad because I was getting the and reading regardless of how I oriented the leads.   this whole situation surprised me because the bulb definitely dimmed considerably, plus the amp was in standby where the ground was lifted between the bias resistor/ bypass cap and diode rectifier do it seem that diodes should've been affected.   so next I replaced the diodes and tested again but the bulb did not dim.   so of  course I feared the worst, that it  must be my pt.   next I disconnected the pt ht secondaries  from the diodes and tested again.   this time the lightbulb dimmed.   next I used dmm  to check voltage across ht leads and got 600v  so that looked good to me.  next I changed the bias resistor and bypass cap and checked the dpdt I used for the power/ standby switch for continuity acne that checks out but yet now the lightbulb only dims  when the ht  it's  disconnected.  I feel as though I'm out of things to check because all of the filter caps are after the standby switch.   sorry for the extremely long post and any help would be appreciated.   thanks.

[Bugman3183]

 sorry I meant to write it seems the diodes shouldn't have been affected

[HotBluePlates]

... I used my doom lightbulb tester on it, and  with the amp in standby position the lightbulb dimmed for about five or six seconds and then there was a loud pop and smoke everywhere. ...

Are you sure the lightbulb tester is wired correctly?

If it is, a dim bulb would indicate little current draw (and proper operation). If it were wired like any other lightbulb, a dim bulb would mean the amp is drawing too much current and diverting it away from the bulb.

In other words, "correct lightbulb tester" and "dim bulb with amp smoke" are incompatible situations.

... 5e3  type amp using a full wave bridge rectifier.   ... I used my dmm to check the diodes and they seemed to be bad because I was getting the and reading regardless of how I oriented the leads. ...

You are sure you have a bridge rectifier? With 4 solid-state diodes?

What power transformer did you use? Does it have a center-tap on the high voltage winding (generally red-yellow color code)? Did you connect that center-tap to ground?

HINT: If you have a bridge rectifier, you should not connect the center-tap to anything, if one is present.

[Bugman3183]

 thanks for the response.   to answer your first question, I'm sure the lightbulb  tester it's wired properly I've used it on two previous builds.  18  watt tmb and a 5e3  and the tester performed flawlessly. 

 yes I used a bridged rectifier with four diodes using  a mps 275 0 275  power transformer   and didn't have the ct ( made that mistake once before)  but the bulb definitely dimmed so I was feeling confident everything was cool, imagine my surprise when I heard that loud pop!   I've since used the tester on a couple other amps that I have that operate properly and it performed as expected.   I'm completely baffled by this one.   thanks.

[Bugman3183]

 also, I've since removed the bridged rectifier, used two new diodes instead and connected the ct  to the positive side of the bias resistor and bypass cap and the lightbulb still shines brightly.

[HotBluePlates]

... yes I used a bridged rectifier with four diodes using  a mps 275 0 275  power transformer   and didn't have the ct ( made that mistake once before)  ...

You wanted 275*2*1.414 = 777v as your B+?

The overwhelming majority of vintage guitar amps did not use a bridge rectifier. As a result, most of the available power transformers for replicating old amps would have a winding like what you seem to describe, 275-0-275.

The "0" indicates that a grounded center-tap was envisioned, along with a full-wave (non-bridge) rectifier. For a 275v-0-275v winding and a solid-state full-wave rectifier you would expect 275v *1.414 = ~388vdc B+

So if you truly used a bridge, you very likely exceeded the voltage rating of whatever filter caps you used. They're probably dead, and may continue to be good-quality short circuits as the over-voltage punched through the dielectric.

Let's assume you PT isn't shot. Disconnect the output of the rectifier from the first filter cap (to disconnect the PT/rectifier from the rest of the power supply). If you don't really want nearly 800vdc of B+, then use a regular full-wave rectifier (look at vintage Fender amps with solid-state rectifiers for an example). You will need to reconnect the center-tap to ground after conversion to the full-wave rectifier.

Now, try reconnecting the PT/rectifier to power and your lightbulb tester, and see if just this section performs correctly. You may need a single new filter cap (rated say 450vdc) on the output of the rectifier to develop a measurable d.c. voltage. Verify nothing from this new stand-alone filter cap back to the power plug is shorted (obviously, the PT secondary winding will show some resistance to ground because the CT is grounded and there is a relatively low resistance through the secondary winding).

[PRR]

My penny says electrolytic capacitor was reversed and blew-up.

However if you bridged a 2-diode CT PT, as HBP suspects, you had 777V on 450V caps and yes this will go BOOM. (But last time I did this right, it blew foil and tissue ALL over the guts, no wondering what happened.)

Having blown-up a cap, you may have further damage from repeated testing. Use new diodes, wired right, one 10u-20uFd 450V cap, check polarity, and lamp-test with voltmeter clipped to the cap. If you really wanted 350V, and it comes up to 350-360V with dim lamp, you are OK. If it rushes up to 700++V, unplug fast and figure out what's wired wrong. (This metered-fireup technique has avoided two BOOMs for me.)

[eleventeen]

I once worked at a factory where freshly "built" eg; "PC-board stuffed & flow soldered" things were being built. There were two older German women who were the first to apply power to the completed units. Above them, the ceiling was large wooden beams, filled in with foil-backed insulation. Above their benches, the foil was perforated with holes from exploding electroytic caps, whose chance at serviceable life had been tragically cut short by backwards insertion on the assembly line, one floor below. But when they popped, it was pretty funny (to everyone else) and the gals would let out a stream of excellent cursing.

My job was to fix the rejects. Sometimes I would spend all day flipping diodes around and replacing blown-out backwards electrocaps.

[firemedic]

I love German cursing. What a great language.

[Ed_Chambley]

Is it possible to connect the center tap to the negative of the first filter cap and use a bridge rectifier and not double supply voltage?

[HotBluePlates]

Is it possible to connect the center tap to the negative of the first filter cap and use a bridge rectifier ...

The negative of the first filter cap is ground, so making that connection shorts half the diodes of the bridge.

EXCEPTION: If you have series filter caps for a higher total voltage rating, you could connect the center-tap to the junction of the two filter caps (mid-way between B+ and ground).

Is it possible to connect the center tap to the negative of the first filter cap and use a bridge rectifier and not double supply voltage?

Not like you're thinking.

"Double compared to what?" This is the question you should be asking yourself.

A fullwave rectifier is normally used with a center-tapped power transformer. The CT gets grounded. When the transformer is energized, one end of the total secondary winding is positive compared to the CT, and the other end is negative compared to the CT. For our B+, the end with the positive voltage compared to the CT forward-biases the diode it is connected to and passes current. The other diode is reverse-biased and blocks current.

*At that moment, the diode which is conducting allows a voltage to appear at the first filter cap which is equal to the voltage at that end of the transformer winding through to the center-tap.

During the other half-cycle of the wall voltage, the voltages of the ends of the secondary winding reverse, and the opposite diode conducts again with the same voltage appearing at the filter cap.

*The important thing to see is that a full-wave rectifier and center-tapped transformer winding is connected such that the filter cap only sees the voltage of half of the total secondary winding at any moment. So a 300-0-300v PT only presents the positive half of 300vac to the first filter cap, and the cap charges to the peak voltage of 424v, minus diode voltage drop.

Bridge rectifiers don't use a center-tap (except, as PRR pointed out before, when you're building a bi-polar supply; but that doesn't apply to vacuum tubes generally). The arrangement of diodes is such that at all times there is a direct connection between both ends of the transformer secondary to the first filter cap.

For the same 300-0-300v winding described earlier, the filter cap sees 300+300= 600vac, rectified and tries to charge to the peak voltage of 848v minus diode drops.

*As a result, transformers intended for use with bridge rectifiers are wound for half the total voltage from end to end of the secondary.

So "doubling" of voltage when using a bridge with a transformer intended to be used with a full-wave rectifier is only a comparison of what voltage you get in relation to what voltage you expected. And the doubling is only due to utilizing the entire secondary all the time.

Is it possible to connect the center tap to the negative of the first filter cap and use a bridge rectifier and not double supply voltage?

To answer your question after presenting all that... The way to not double supply voltage is to use the full-wave rectifier as intended with the transformer in question.

[sluckey]

Is it possible to connect the center tap to the negative of the first filter cap and use a bridge rectifier and not double supply voltage?

Yes. BUT YOU MUST NOT CONNECT THE NEGATIVE SIDE OF THE BRIDGE TO THE SAME POINT AS THE CT IS CONNECTED! In fact, don't connect the negative side of the bridge to anything unless you need a bipolar power supply. IOW, only half the bridge is being used to produce B+. I'd rather just use two individual diodes and not have to wonder about the mystery bridge later down the road.

[HotBluePlates]

Good call Sluckey!

Note to on-lookers:
A bridge normal has two input terminals for a.c. from the power transformer. It also has + and - outputs for connection to either end of the first filter cap. If you use a bridge package where all 4 diodes are housed in one square block with 4 leads, you'll see the inputs and outputs marked according to this scheme.

What Sluckey is describing is essentially to not use 2 of the 4 diodes by not attaching the - output to anything. Now if you connect the CT to ground, use the a.c. inputs and only the + output, you have wired your bridge package for regular full-wave (non-bridge) operation.

I read Ed's question more narrowly as still using the entire bridge rectifier along with a CT. I also think it's important for folks to realize how these circuits are fundamentally different. After knowing that, you can toss it aside and just focus on how to rig a bridge for non-bridge operation.

[Ed_Chambley]

The reason I asked is the package diodes which can be center bolted are very clean and neat to use.  They can also be mounted anywhere, but if you do not ground the rectifier, then it is not bridged, it is simply 2 diodes for each HT, correct?

I understand the power phase shift and do actually have a build doing both.  One of course is a Marshall where the CT is connected to the center of 4 100uf (2 pairs of series) which is how a lot of the guys on the Metro Forum build 100 watt lead amps.

What I was referring to here is I have a 5E3 modified build whereas the CT is connected to the negative at the filter cap and uses a package diode which is not grounded.  Basically this is really like a Full Wave since there is no ground on the diode?  The problem is it appears to be a bridge, but since there no ground it is not.  Am I thinking correctly?

[HotBluePlates]

Look at the picture.

The tube rectifier takes the place of the two diodes in the bridge package which are connected between the "~" pins and the "+" pin.

So yes, if you do not connect the "-" pin to anything, then those other two diodes are connected to nothing, and with a grounded center-tap, the package functions as a normal full-wave rectifier.

[Ed_Chambley]

Pardon me, but I am asking all these questions because I am designing and am in the middle of making a layout.  In the attached is a cropped bassman full wave.  There are 6 total diodes, 2 sets of 3.  What is the purpose of the 2x 3 in series as opposed to a Marshall which traditionally uses 2x 2 in series.

I have rejected the idea of the package diodes since in using one, there is no benefit since 2 of the diodes are not used if the negative is not used in a CT Transformer.  I will put them on the first of the board or simply make a small board.

I intend to start a new thread in the next few days after all my parts have arrived.  Also, is there really any benefit in using FRED diodes.  I have read some say they have noticed an increase in harmonics, but I just cannot understand how this is possible.

Another thing is I got a 66 Bassman last night in for a tuneup.  When i puled the chassis I noticed someone has changed the rectifier arrangement.  Each diode has a .1uf cap and a 470 ohm resistor connected from the anode to cathode, so the setup is 6 diodes each with the cap and resistor.  I do not understand how this can work or what the intended purpose.  It is really just a tube job and checkup, but the amp seems to lack punch of a traditional bassman.

[sluckey]

There are 6 total diodes, 2 sets of 3.  What is the purpose of the 2x 3 in series as opposed to a Marshall which traditionally uses 2x 2 in series.

Using diodes in series like this will increase the PIV of the whole string. For example, you have a bucketful of diodes with a 400V PIV rating. But you need to operate them in a circuit such as a Bassman that will expose them to 900V. Using 3  of the 400V diodes in series will increase the PIV rating to 1200V. Now you can safely operate the string in the Bassman circuit.

Same principal applies to the Marshall amp. Why only 2 in series? Maybe Marshall B+ is not as high, or maybe Marshall didn't mind operation closer to the bleeding edge, or maybe they used 500v or 600v diodes. The idea is the same though. It allows you to use lower voltage diodes in a higher voltage circuit.

Oh yeah, back in the day when that Bassman rolled off the assembly line, there were no cheap consumer priced 1000V diodes. Silicon was still relatively new and 200V, 400V, and 600V ratings were common and affordable.

Each diode has a .1uf cap and a 470 ohm resistor connected from the anode to cathode, so the setup is 6 diodes each with the cap and resistor.

The cap and resistor will 'snub' diode switching noise. Probably wasted on a 66 Bassman.

[Willabe]

I thought the 470 R's were to force the diodes to share the voltage across them equally?


             Brad    

[sluckey]

Probably so???

[Ed_Chambley]

Thanks again Guys.  Makes perfect sense when you understand it.  I guess I have not been doing this long enough to know 1000v diodes were not readily available in 66, or at least not as affordable.  No benefit, just increased voltage.