Something shocking: Dynaco ST35 with EFB mods

[kirk57]

After finding a cheap ST35 chassis on Ebay, I decided to use the Z565 OPTs I bought for my Magnavox 9304 and just build a new ST35. I bought a new PT from Dynakit, and the PC13 and EFB boards from our friend Dave.

Got it all wired up and getting ready to bias the tubes, and touched the chassis and got a good jolt.
Grabbed my DMM and sure enough, 47 VAC between the outlet ground and chassis. I tried a different outlet in a different room, another outlet on the variac, a different cord, swapped polarity of the cord, no difference: 47 VAC.

I'm using the original two-wire cord setup, all new parts of course. I seem to recall a similar issue with the Magnavox and in that case I went to a 3-prong plug oriented for least electrocution potential. Can I do that here? I hear different opinions on the subject.

*Everything* is new on this amp except the chassis and speaker connectors.

Any ideas how I figure out the problem? Where would you start?

 

[Retrovert]

Maybe a filter capacitor in the power supply is either leaky or backwards wired?

Check the polarity. The polarity of cans can often be difficult to read.

These normally explode when reverse wired, which makes them easy to find. Puked guts, foil and insulator, like a streamer from New Years...

 

[gadget73]

Do these have a safety cap from both legs of the power cord to chassis? If so you're always going to be at ~60vac relative to ground. Removing one cap and leaving it only on the neutral side will sort it. If no caps from power cord to chassis, you have leakage somewhere. Being AC and reading to line ground, I'd be inclined to suspect its on the primary side, so power cord, switch, fuse, transformer, and any wiring between. First step is a good visual inspection.

A grounded cord will technically get rid of the voltage to ground issue, but if you have a leakage path it might cause the smoke to come out of somewhere. Need to fix the leakage problem first. Add a ground if you'd like after that.

 

[BinaryMike]

There's nothing on the primary side of the power transformer except a fuse and line cord in this model. Have you checked DC resistance between primary and chassis? It should be infinite on a standard ohmmeter.

 

[Retrovert]

Do these have a safety cap from both legs of the power cord to chassis? If so you're always going to be at ~60vac relative to ground. Removing one cap and leaving it only on the neutral side will sort it.

The official ST-70 didn't use a death cap.

If one was used, and it is still desired, it must be one of the special X1/Y2 safety capacitors, so if/when it fails the failure is an open, not a short. Otherwise a guitarist dies.

If no caps from power cord to chassis, you have leakage somewhere.

Or a filter cap is backwards or failing. Leakage current is what cooks them. The ohmic heating can be considerable.

A grounded cord will technically get rid of the voltage to ground issue, but if you have a leakage path it might cause the smoke to come out of somewhere. Need to fix the leakage problem first. Add a ground if you'd like after that.

Ummm, one may not ground a live chassis and expect the voltage to magically vanish. What one might expect is that current is now passing through the ground wire, which is a bad idea. Ground is there in case of a short, not as a routine substitute for the proper neutral to return the current. That says something is wrong with the return via neutral.

Also, grounding the chassis can create ground loops and lots of noise.

 

[heyraz]

It looks like a pretty straightforward design. I'd look for a "dumb" mistake like a solder blob or perhaps a stray piece of clipped wire causing a short.

 

[gadget73]

Ummm, one may not ground a live chassis and expect the voltage to magically vanish

Sure you can. Tie it to ground, it will be zero volts relative to ground. I won't tell you that it won't potentially blow up or cause other issues if you do that, but that is the purpose of a safety ground connection. Also note that I did say it needs to be fixed before this is done.

I've added grounds to things that had 2 prong cords without issues. Some people have had issues though.

 

[kirk57]

Maybe a filter capacitor in the power supply is either leaky or backwards wired?

Check the polarity. The polarity of cans can often be difficult to read.

These normally explode when reverse wired, which makes them easy to find. Puked guts, foil and insulator, like a streamer from New Years...

This I can rule out, as all the PS caps are on DG's EFB board, and that is stuffed correctly. I followed the directions to connect the EFB board.

Do these have a safety cap from both legs of the power cord to chassis?

No safety caps

There's nothing on the primary side of the power transformer except a fuse and line cord in this model. Have you checked DC resistance between primary and chassis? It should be infinite on a standard ohmmeter.

This I will do today

It looks like a pretty straightforward design. I'd look for a "dumb" mistake like a solder blob or perhaps a stray piece of clipped wire causing a short.

I fully expect the word "dumb" will figure into it when I find the problem

 

[StuartB]

Have you pinged Dave? he has been very responsive to me when I had questions on using his boards.

 

[Retrovert]

Sure you can. Tie it to ground, it will be zero volts relative to ground. I won't tell you that it won't potentially blow up or cause other issues if you do that, but that is the purpose of a safety ground connection. Also note that I did say it needs to be fixed before this is done.

This is a misunderstanding of neutral and ground in an electrical distribution system, where one has the return path for circuits (signal ground) and an emergency path to prevent electrification of a chassis in case of fault (safety ground), which are entirely different concepts.

Current leaking to ground is a circuit fault arising from leakage current. It isn't cured by using a ground wire. It is cured by identifying the source of the leakage current and fixing it. It is a circuit defect.

Neutral in any current distribution system is a reference point and the return path for current. While neutral may be grounded at the distribution point, it need not be. The hot and neutral are intended to form the complete electric circuit. Such circuits may function without an external ground, as is the case in a two-wire distribution system.

Ground is an emergency conduction path intended to only carry fault current arising from a breakdown (defect) in the equipment. That breakdown may occur because of insulation failure, improper circuit construction, or accidental fault (short) externally induced. In many cases ground, unlike neutral, is an uninsulated wire because it is not intended to routinely carry current.

One should, in fact, be able to interrupt the ground wire and still have proper circuit functioning, assuming no fault has arisen. Neutral, to contrast, cannot be interrupted without disrupting the current path.

Electrical ground is a safety feature. It is not a second signal ground and must not be treated as such. Any equipment using safety ground as a second neutral violates the electric code and could not receive certification from UL or any other ratings body.

This is an important distinction

Ground is therefore not equivalent to neutral and should not be treated as equivalent. The fact that neutral is grounded at the distribution point does not make the two circuits equivalent.

I've added grounds to things that had 2 prong cords without issues. Some people have had issues though.

The safety ground is a long wire which may act like an antenna. It can, in consequence, induce noise into a high-impedance circuit.

The ground planes created by the third wire are never at quite the same potential and hum and noise can result.

I've experienced this with tube tuners, preamplifiers, and amplifiers.

 

[Retrovert]

This I can rule out, as all the PS caps are on DG's EFB board, and that is stuffed correctly. I followed the directions to connect the EFB board.

If you disconnect the output of the power supply and instead connect it to a load resistor do you still have a hot chassis?

If so, that suggests internal leakage, perhaps from a fault transformer winding.

The fact that it is a fraction of the mains potential suggests leakage current, particularly at a low-enough current that nothing has yet burned out.

 

[triode17]

47 VAC between the outlet ground and chassis. I tried a different outlet in a different room, another outlet on the variac, a different cord, swapped polarity of the cord, no difference: 47 VAC.

Try hooking a 10k, 2w resistor between the outlet Earth (aka ground) and the chassis. If you still read 47v, that's a 4.7ma leak which is dangerous. If it drops way down to a volt or two, it's just slight leakage and the 3 prong cable should alleviate it.
BTW, get a portable GFCI outlet for the bench. It will blow on an Earth fault.

 

[Retrovert]

UL or CE would never certify a unit with even low-current leakage. When leakage exists, something is wrong, maybe even very wrong, and a bigger failure is likely lurking in the wings.

+1 on the GFI. I have seen them protect against electrocution from unsafe equipment. I changed all of my outlets to GFIs for this reason and have installed them for others as well.

 

[gadget73]

UL allows 0.5 ma of leakage.

https://library.ul.com/wp-content/u...e-Dielectric-Voltage-Withstand-Test_v5_HR.pdf

 

[ducati_EL34]

+1 on the GFI. I have seen them protect against electrocution from unsafe equipment. I changed all of my outlets to GFIs for this reason and have installed them for others as well.

Worry much?

 

[heyraz]

UL allows 0.5 ma of leakage.

Yeah, but he "got a good jolt", so there's definitely something wrong going on.
I thought perhaps if the chassis served as a "star ground" and an unpolarized power cord was inserted so the plug's neutral was in the hot side of the socket...maybe. But that can't be the case.
47 volts sounds like a bias supply to me. I'm not familiar with that amp. Have you checked over the bias supply circuit?

 

[gadget73]

the trouble is none of that stuff references to earth ground. The only thing in that amp that should be capable of reading voltage to earth is leakage on the primary side. Might be worth poking a voltmeter into the power socket just to make sure the neutral does actually read basically zero volts to ground. A couple above is fine, a bunch above is not.

Also I suppose its worth verifying whatever was plugged in as a source isn't the actual problem since the RCA jacks should be tied to chassis somewhere along the way. Leakage from the source would show up on the chassis of the power amp.

 

[nj pheonix]

the trouble is none of that stuff references to earth ground. The only thing in that amp that should be capable of reading voltage to earth is leakage on the primary side. Might be worth poking a voltmeter into the power socket just to make sure the neutral does actually read basically zero volts to ground. A couple above is fine, a bunch above is not.

Also I suppose its worth verifying whatever was plugged in as a source isn't the actual problem since the RCA jacks should be tied to chassis somewhere along the way. Leakage from the source would show up on the chassis of the power amp.

That's an interesting point.
I seem to remember someone saying they did a upgrade (or perhaps a new build) a while back and had a functional issue due to stock jacks being physically tied to the chassis and upgraded not.
Making that connection solved the issue.
I don't remember who did the build.
I'm thinking it was Tom (wharfcreek?)
That inherited the amp for fixing.
Not sure.
Maybe I have the facts crossed

 

[Retrovert]

Responding to my comment about the wisdom of GFIs, "+1 on the GFI. I have seen them protect against electrocution from unsafe equipment. I changed all of my outlets to GFIs for this reason and have installed them for others as well."

Worry much?

That's a very dismissive comment. To begin, I actually don't much worry about my AC equipment, particularly the older gear, because my outlets are GFIs and I use a GFI on the bench.

I installed the GFIs to not worry. Every once in a while I hear stories from someone I know about how one saved them, friend, or family from a shock or death. Particularly from miswired lamps, tools (drills and saws), kitchen appliances (toasters, blenders, and mixers), and the odd hot-chassis TV or radio.

I don't think it is a superior attitude to proclaim that one doesn't use safety devices. GFIs are inexpensive and do their job 24-7-365. Humans make mistakes. That's life. Safety devices—GFIs and Interlocks—prevent occasional lapses, as well as poorly made product errors, from being fatal. Anyone who works on old equipment should have GFI outlets. Even those who like to ride motorcycles over 100 MPH. Because when it comes to vacuum tube gear with a hot chassis , even luck has its limits.

 

[Retrovert]

Also I suppose its worth verifying whatever was plugged in as a source isn't the actual problem since the RCA jacks should be tied to chassis somewhere along the way. Leakage from the source would show up on the chassis of the power amp.

Which is a good point.

As an extreme case, I once had a potentially hot-chassis tuner connected to a potentially hot-chassis amplifier. The tuner plug was reversed, the amplifier's was not. Or vice versa. It was forty years ago and I can't remember which one was swapped. But one had a plug inserted in the wrong direction.

I don't remember which one had a death cap, or maybe both did, but one had to be present and leaky for what happened to have happened.

Because the RCA jacks connected the two chassis the fuse blew. Yet either piece separately worked.]

After that I had a GFI in an outlet to prevent problems. In those days the GFI cost about $40 apiece, and that was in the days when this was about $120 in today's debased currency. When I redid my apartment outlets I paid under $10 apiece for the fancy Levitton GFIs. Quite a difference in price was made by volume production.